Complete professional prompt handbook · Edition 01
AI Automation Specialist: 120 Copy-Ready Prompts & Shortcodes
A practical library for remote AI automation work, from client discovery and n8n/Make/Zapier architecture to APIs, AI agents, QA, privacy, delivery and portfolio proof. No external prompt directory needed.
12 workflows · 120 unique prompts · search, filter and copy · free to use
How to use
Search by task, category or shortcode. Replace every [BRACKETED INPUT] with your real, authorized project details. Press Copy prompt, paste into your AI assistant and verify outputs before production.
Shortcode format
/AUTO-001 through /AUTO-120 are navigation labels used in this library, not built-in ChatGPT commands. Copy the full prompt text; a shortcode alone does not execute the workflow.
Professional safeguards
Never paste credentials or confidential personal data. Confirm vendor API features in official documentation; dry-run with synthetic records; require explicit approval for irreversible actions.
Quick workflow: Discover → Map → Design → Integrate → Build → Test → Approve → Deploy → Monitor → Document. Suggested portfolio deliverable: sanitized process diagram, automation export, synthetic test fixtures, runbook and an evidence-based demo.
Module 01 · 10 prompts
Discovery & Scoping
Define the business problem before choosing a tool.
Intake interview
Use this when you need to: Conduct a structured discovery interview for [client/business].
Act as a senior AI automation consultant and practical implementation partner. Task: Conduct a structured discovery interview for [client/business]. Ask about current process [PROCESS], actors, volume, systems, exceptions, pain points, budget, security and success criteria. Group questions by topic and produce a blank answer sheet. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Problem statement
Use this when you need to: Turn these client notes [NOTES] into a concise problem statement, affected users, existing bottlenecks, measurable outcome, assumptions and open questions.
Act as a senior AI automation consultant and practical implementation partner. Task: Turn these client notes [NOTES] into a concise problem statement, affected users, existing bottlenecks, measurable outcome, assumptions and open questions. Do not invent facts. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Process inventory
Use this when you need to: Create an inventory of recurring processes for [BUSINESS TYPE].
Act as a senior AI automation consultant and practical implementation partner. Task: Create an inventory of recurring processes for [BUSINESS TYPE]. Columns: trigger, owner, current tool, frequency, minutes per run, error risk, sensitive data, potential for automation and evidence needed. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Opportunity scorecard
Use this when you need to: Evaluate these candidate workflows [LIST] by impact, feasibility, data quality, integration availability, risk, maintenance and reversibility.
Act as a senior AI automation consultant and practical implementation partner. Task: Evaluate these candidate workflows [LIST] by impact, feasibility, data quality, integration availability, risk, maintenance and reversibility. Explain weights and rank with uncertainty. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Stakeholder map
Use this when you need to: Map requesters, approvers, operators, IT/security and affected customers for [WORKFLOW].
Act as a senior AI automation consultant and practical implementation partner. Task: Map requesters, approvers, operators, IT/security and affected customers for [WORKFLOW]. Give concerns, decisions and sign-off criteria for each. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
AS-IS workflow
Use this when you need to: Translate [CURRENT STEPS] into numbered AS-IS workflow with inputs, decisions, handoffs, delays, exceptions, outputs and unknowns.
Act as a senior AI automation consultant and practical implementation partner. Task: Translate [CURRENT STEPS] into numbered AS-IS workflow with inputs, decisions, handoffs, delays, exceptions, outputs and unknowns. Identify where human confirmation is necessary. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
TO-BE workflow
Use this when you need to: Propose a future-state automation for [GOAL] using [TOOLS].
Act as a senior AI automation consultant and practical implementation partner. Task: Propose a future-state automation for [GOAL] using [TOOLS]. Include triggers, actions, branches, human approvals, logging, failure routes and rollback. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Scope boundary
Use this when you need to: Create in-scope, out-of-scope, dependencies, exclusions and acceptance criteria for automating [PROCESS].
Act as a senior AI automation consultant and practical implementation partner. Task: Create in-scope, out-of-scope, dependencies, exclusions and acceptance criteria for automating [PROCESS]. Explicitly mark capabilities not verified. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Baseline measurement
Use this when you need to: Design a measurement plan for current [PROCESS]: cycle time, touch time, error rate, throughput, exception volume and customer impact.
Act as a senior AI automation consultant and practical implementation partner. Task: Design a measurement plan for current [PROCESS]: cycle time, touch time, error rate, throughput, exception volume and customer impact. Include sample-size caveats. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Discovery summary
Use this when you need to: Write a client-ready discovery report from [NOTES]: objective, current state, opportunity, constraints, required access, next decisions and unresolved questions.
Act as a senior AI automation consultant and practical implementation partner. Task: Write a client-ready discovery report from [NOTES]: objective, current state, opportunity, constraints, required access, next decisions and unresolved questions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 02 · 10 prompts
Workflow Design & Logic
Turn requirements into reliable automation specifications.
Trigger specification
Use this when you need to: Specify trigger for [WORKFLOW]: event source, schedule/webhook/polling, authentication, payload example, deduplication key, retries and latency target.
Act as a senior AI automation consultant and practical implementation partner. Task: Specify trigger for [WORKFLOW]: event source, schedule/webhook/polling, authentication, payload example, deduplication key, retries and latency target. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Input schema
Use this when you need to: Design an input data contract for [USE CASE]: fields, type, required/optional, example, validation, privacy class and missing-value behavior.
Act as a senior AI automation consultant and practical implementation partner. Task: Design an input data contract for [USE CASE]: fields, type, required/optional, example, validation, privacy class and missing-value behavior. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Decision tree
Use this when you need to: Build a decision tree for [WORKFLOW] from [RULES].
Act as a senior AI automation consultant and practical implementation partner. Task: Build a decision tree for [WORKFLOW] from [RULES]. Include precedence, edge cases, fallbacks and questions for ambiguous rules. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Swimlane breakdown
Use this when you need to: Write a textual swimlane for [PROCESS] across [ACTORS/SYSTEMS], listing trigger, owner, steps, handoffs, wait states and approvals.
Act as a senior AI automation consultant and practical implementation partner. Task: Write a textual swimlane for [PROCESS] across [ACTORS/SYSTEMS], listing trigger, owner, steps, handoffs, wait states and approvals. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
State machine
Use this when you need to: Design states and transitions for [ENTITY] in [WORKFLOW].
Act as a senior AI automation consultant and practical implementation partner. Task: Design states and transitions for [ENTITY] in [WORKFLOW]. Include permitted transitions, idempotency, timeouts, retries and terminal states. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Exception register
Use this when you need to: List likely failures for [AUTOMATION]: detection, severity, user-visible effect, safe fallback, responsible owner and recovery playbook.
Act as a senior AI automation consultant and practical implementation partner. Task: List likely failures for [AUTOMATION]: detection, severity, user-visible effect, safe fallback, responsible owner and recovery playbook. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Human-in-loop gate
Use this when you need to: Design human approval checkpoints for [HIGH-IMPACT ACTION] including approval context, audit log, timeout and reject path.
Act as a senior AI automation consultant and practical implementation partner. Task: Design human approval checkpoints for [HIGH-IMPACT ACTION] including approval context, audit log, timeout and reject path. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
SLA and SLO
Use this when you need to: Propose measurable reliability, processing time and recovery objectives for [WORKFLOW], based on [BUSINESS REQUIREMENTS], with explicit assumptions.
Act as a senior AI automation consultant and practical implementation partner. Task: Propose measurable reliability, processing time and recovery objectives for [WORKFLOW], based on [BUSINESS REQUIREMENTS], with explicit assumptions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Versioning plan
Use this when you need to: Define configuration, schema and workflow versioning strategy for [AUTOMATION].
Act as a senior AI automation consultant and practical implementation partner. Task: Define configuration, schema and workflow versioning strategy for [AUTOMATION]. Explain backward compatibility, migration and rollback. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Architecture review
Use this when you need to: Critique [ARCHITECTURE DESCRIPTION] for bottlenecks, single points of failure, privacy, observability, cost and maintainability.
Act as a senior AI automation consultant and practical implementation partner. Task: Critique [ARCHITECTURE DESCRIPTION] for bottlenecks, single points of failure, privacy, observability, cost and maintainability. Prioritize remedies. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 03 · 10 prompts
n8n, Make & Zapier
Platform-specific build planning without fabricated app features.
n8n node plan
Use this when you need to: Design an n8n workflow for [TRIGGER] to [OUTCOME].
Act as a senior AI automation consultant and practical implementation partner. Task: Design an n8n workflow for [TRIGGER] to [OUTCOME]. Provide node order, purpose, field mappings, expressions as illustrative pseudocode, error handling and items to verify against installed node versions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Make scenario blueprint
Use this when you need to: Design a Make scenario for [PROCESS].
Act as a senior AI automation consultant and practical implementation partner. Task: Design a Make scenario for [PROCESS]. Specify modules, filters, routers, bundle mapping, error handlers, scheduling and operations cost assumptions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Zapier Zap design
Use this when you need to: Design a Zapier Zap for [PROCESS].
Act as a senior AI automation consultant and practical implementation partner. Task: Design a Zapier Zap for [PROCESS]. Specify trigger, actions, paths/filters, field mappings, approval step and plan/feature availability checks. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Platform comparison
Use this when you need to: Compare n8n, Make, Zapier and custom code for [PROJECT] using hosting, connector support, privacy, budget, volume, debugging, lock-in and governance.
Act as a senior AI automation consultant and practical implementation partner. Task: Compare n8n, Make, Zapier and custom code for [PROJECT] using hosting, connector support, privacy, budget, volume, debugging, lock-in and governance. Do not assert current prices. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Node-by-node QA
Use this when you need to: Review this workflow configuration [EXPORT/TEXT] for incorrect fields, unsafe assumptions, branching errors, duplicate sends and missing error routes.
Act as a senior AI automation consultant and practical implementation partner. Task: Review this workflow configuration [EXPORT/TEXT] for incorrect fields, unsafe assumptions, branching errors, duplicate sends and missing error routes. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Expression helper
Use this when you need to: Write an expression or transformation for [PLATFORM/VERSION] converting [INPUT EXAMPLE] to [EXPECTED OUTPUT].
Act as a senior AI automation consultant and practical implementation partner. Task: Write an expression or transformation for [PLATFORM/VERSION] converting [INPUT EXAMPLE] to [EXPECTED OUTPUT]. Explain null handling and how to test it. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Webhook setup plan
Use this when you need to: Explain secure webhook setup for [PLATFORM] and [SOURCE]: validation, signature checking if supported, test event, secret storage, retry response and troubleshooting.
Act as a senior AI automation consultant and practical implementation partner. Task: Explain secure webhook setup for [PLATFORM] and [SOURCE]: validation, signature checking if supported, test event, secret storage, retry response and troubleshooting. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Pagination blueprint
Use this when you need to: Design pagination handling for [API] on [PLATFORM], including cursor/offset, rate limits, max-page safety, deduplication and resume checkpoints.
Act as a senior AI automation consultant and practical implementation partner. Task: Design pagination handling for [API] on [PLATFORM], including cursor/offset, rate limits, max-page safety, deduplication and resume checkpoints. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Batching strategy
Use this when you need to: For [VOLUME] records on [PLATFORM], plan batch size, parallelism, wait intervals, error isolation, checkpoints and cost monitoring.
Act as a senior AI automation consultant and practical implementation partner. Task: For [VOLUME] records on [PLATFORM], plan batch size, parallelism, wait intervals, error isolation, checkpoints and cost monitoring. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Migration checklist
Use this when you need to: Plan migration of [WORKFLOW] from [SOURCE TOOL] to [TARGET TOOL] with feature parity, credentials, sample dataset, shadow run, cutover and rollback.
Act as a senior AI automation consultant and practical implementation partner. Task: Plan migration of [WORKFLOW] from [SOURCE TOOL] to [TARGET TOOL] with feature parity, credentials, sample dataset, shadow run, cutover and rollback. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 04 · 10 prompts
APIs, Webhooks & Integrations
Connect systems safely and document contracts.
API research brief
Use this when you need to: Create an API due-diligence checklist for [SERVICE] and [USE CASE]: official docs, auth scopes, endpoints, limits, webhooks, pagination, sandbox, data residency and terms.
Act as a senior AI automation consultant and practical implementation partner. Task: Create an API due-diligence checklist for [SERVICE] and [USE CASE]: official docs, auth scopes, endpoints, limits, webhooks, pagination, sandbox, data residency and terms. Mark anything unverified. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
HTTP request design
Use this when you need to: Draft an illustrative HTTP request for [ENDPOINT DOCS] given [PAYLOAD].
Act as a senior AI automation consultant and practical implementation partner. Task: Draft an illustrative HTTP request for [ENDPOINT DOCS] given [PAYLOAD]. Show method, headers placeholders, body, success/error response and validation. Never fabricate credentials. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
OAuth mapping
Use this when you need to: Explain the OAuth authorization flow needed for [SERVICE/APP TYPE], least-privilege scopes, redirect URI, token storage/refresh, revocation and test steps.
Act as a senior AI automation consultant and practical implementation partner. Task: Explain the OAuth authorization flow needed for [SERVICE/APP TYPE], least-privilege scopes, redirect URI, token storage/refresh, revocation and test steps. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Webhook receiver
Use this when you need to: Design a production-minded webhook handler for [EVENT]: signature verification where available, timestamp freshness, idempotency, queueing, ack and observability.
Act as a senior AI automation consultant and practical implementation partner. Task: Design a production-minded webhook handler for [EVENT]: signature verification where available, timestamp freshness, idempotency, queueing, ack and observability. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Data mapping table
Use this when you need to: Map fields from [SOURCE SAMPLE] to [DESTINATION SCHEMA].
Act as a senior AI automation consultant and practical implementation partner. Task: Map fields from [SOURCE SAMPLE] to [DESTINATION SCHEMA]. Include transforms, type conversion, missing values, validation and rejected-record treatment. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Rate-limit plan
Use this when you need to: Prepare 429/rate limit handling for [API DOCS], including documented headers, exponential backoff with jitter, queueing and maximum retry policy.
Act as a senior AI automation consultant and practical implementation partner. Task: Prepare 429/rate limit handling for [API DOCS], including documented headers, exponential backoff with jitter, queueing and maximum retry policy. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Error response map
Use this when you need to: Classify [API ERROR SAMPLES] into retry, fix-input, auth-refresh, stop-and-alert and manual escalation; explain why.
Act as a senior AI automation consultant and practical implementation partner. Task: Classify [API ERROR SAMPLES] into retry, fix-input, auth-refresh, stop-and-alert and manual escalation; explain why. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Sync conflict policy
Use this when you need to: Define source of truth and conflict rules for a two-way sync between [SYSTEM A] and [SYSTEM B], including deletes, timestamps and auditability.
Act as a senior AI automation consultant and practical implementation partner. Task: Define source of truth and conflict rules for a two-way sync between [SYSTEM A] and [SYSTEM B], including deletes, timestamps and auditability. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Integration test cases
Use this when you need to: Generate happy path, auth failure, timeout, invalid payload, partial update, duplicate delivery and permission-change tests for [INTEGRATION].
Act as a senior AI automation consultant and practical implementation partner. Task: Generate happy path, auth failure, timeout, invalid payload, partial update, duplicate delivery and permission-change tests for [INTEGRATION]. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
API handoff document
Use this when you need to: Produce developer handoff for [INTEGRATION]: endpoints, schema, auth placeholders, example payloads, failure policy, monitoring and open dependencies.
Act as a senior AI automation consultant and practical implementation partner. Task: Produce developer handoff for [INTEGRATION]: endpoints, schema, auth placeholders, example payloads, failure policy, monitoring and open dependencies. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 05 · 10 prompts
AI & LLM Workflows
Use language models with guardrails and evaluation.
AI task suitability
Use this when you need to: Assess whether [TASK] truly needs an LLM or a deterministic rule.
Act as a senior AI automation consultant and practical implementation partner. Task: Assess whether [TASK] truly needs an LLM or a deterministic rule. Compare accuracy, cost, latency, explainability and harm from errors. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
System prompt draft
Use this when you need to: Draft a bounded system prompt for an assistant that [ROLE/TASK], using [APPROVED SOURCES].
Act as a senior AI automation consultant and practical implementation partner. Task: Draft a bounded system prompt for an assistant that [ROLE/TASK], using [APPROVED SOURCES]. Define allowed actions, refusal/escalation cases, output format and uncertainty handling. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Extraction prompt
Use this when you need to: Extract [FIELDS] from [INPUT TEXT] into valid JSON matching [SCHEMA].
Act as a senior AI automation consultant and practical implementation partner. Task: Extract [FIELDS] from [INPUT TEXT] into valid JSON matching [SCHEMA]. Return null when missing; never infer unprovided personal facts; include evidence snippets where permitted. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Classification rubric
Use this when you need to: Create category definitions and decision rubric for [LABELS] on [DATA TYPE], with borderline examples and an abstain/needs-review option.
Act as a senior AI automation consultant and practical implementation partner. Task: Create category definitions and decision rubric for [LABELS] on [DATA TYPE], with borderline examples and an abstain/needs-review option. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
RAG pipeline plan
Use this when you need to: Design retrieval-augmented answers from [DOCUMENT COLLECTION].
Act as a senior AI automation consultant and practical implementation partner. Task: Design retrieval-augmented answers from [DOCUMENT COLLECTION]. Include chunking, metadata, permission filtering, citations, freshness and insufficient-evidence response. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Tool-calling safety
Use this when you need to: Specify least-privilege tool calling for [ACTIONS], including argument validation, preview-before-write, approval for irreversible actions and audit trails.
Act as a senior AI automation consultant and practical implementation partner. Task: Specify least-privilege tool calling for [ACTIONS], including argument validation, preview-before-write, approval for irreversible actions and audit trails. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Hallucination audit
Use this when you need to: Evaluate these AI outputs [OUTPUTS] against [SOURCE/TRUTH SET].
Act as a senior AI automation consultant and practical implementation partner. Task: Evaluate these AI outputs [OUTPUTS] against [SOURCE/TRUTH SET]. Flag unsupported claims, omissions, contradictions and severity; propose corrected outputs. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Model evaluation set
Use this when you need to: Design evaluation dataset for [AI TASK] with representative cases, rare edge cases, expected labels, scoring rubric and leakage prevention.
Act as a senior AI automation consultant and practical implementation partner. Task: Design evaluation dataset for [AI TASK] with representative cases, rare edge cases, expected labels, scoring rubric and leakage prevention. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Prompt injection defense
Use this when you need to: Threat-model untrusted emails, pages or documents entering [AI WORKFLOW].
Act as a senior AI automation consultant and practical implementation partner. Task: Threat-model untrusted emails, pages or documents entering [AI WORKFLOW]. Add instruction hierarchy, content isolation, allowlists and approval gates. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
AI fallback path
Use this when you need to: Design fallback when the model fails, times out, produces invalid JSON, exceeds budget or has low confidence.
Act as a senior AI automation consultant and practical implementation partner. Task: Design fallback when the model fails, times out, produces invalid JSON, exceeds budget or has low confidence. Specify retries, deterministic checks and human review. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 06 · 10 prompts
CRM, Sales & Lead Operations
Automate sales processes with consent and clear data rules.
Lead capture map
Use this when you need to: Design a consent-aware lead capture workflow from [FORM] to [CRM], with normalization, duplicate checks, source attribution and confirmation.
Act as a senior AI automation consultant and practical implementation partner. Task: Design a consent-aware lead capture workflow from [FORM] to [CRM], with normalization, duplicate checks, source attribution and confirmation. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Lead routing
Use this when you need to: Define routing rules for [LEAD FIELDS] across [OWNERS], considering territory, availability, priority, fairness and fallback queue.
Act as a senior AI automation consultant and practical implementation partner. Task: Define routing rules for [LEAD FIELDS] across [OWNERS], considering territory, availability, priority, fairness and fallback queue. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Lead enrichment check
Use this when you need to: Propose legitimate enrichment steps for [BUSINESS LEADS], recording provenance, lawful basis, confidence and outdated data flags.
Act as a senior AI automation consultant and practical implementation partner. Task: Propose legitimate enrichment steps for [BUSINESS LEADS], recording provenance, lawful basis, confidence and outdated data flags. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
CRM duplicate logic
Use this when you need to: Design duplicate detection using [EMAIL/DOMAIN/PHONE/NAME] with cautious fuzzy matching and a human review queue for uncertain merges.
Act as a senior AI automation consultant and practical implementation partner. Task: Design duplicate detection using [EMAIL/DOMAIN/PHONE/NAME] with cautious fuzzy matching and a human review queue for uncertain merges. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Follow-up sequence
Use this when you need to: Draft a respectful follow-up automation for opted-in [AUDIENCE], with stop rules, preferences, reply detection, caps and handoff.
Act as a senior AI automation consultant and practical implementation partner. Task: Draft a respectful follow-up automation for opted-in [AUDIENCE], with stop rules, preferences, reply detection, caps and handoff. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Deal stage transition
Use this when you need to: Define deal-stage automation for [PIPELINE] with required fields, evidence, alerts and prevention of accidental stage changes.
Act as a senior AI automation consultant and practical implementation partner. Task: Define deal-stage automation for [PIPELINE] with required fields, evidence, alerts and prevention of accidental stage changes. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Meeting handoff
Use this when you need to: Create workflow from booked meeting to CRM update, owner notification and preparation brief.
Act as a senior AI automation consultant and practical implementation partner. Task: Create workflow from booked meeting to CRM update, owner notification and preparation brief. Avoid putting sensitive notes in broad channels. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Sales summary
Use this when you need to: Summarize [APPROVED CRM DATA] into pipeline changes, stalled deals, follow-up priorities and caveats.
Act as a senior AI automation consultant and practical implementation partner. Task: Summarize [APPROVED CRM DATA] into pipeline changes, stalled deals, follow-up priorities and caveats. Separate data from recommendations. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Unsubscribe handling
Use this when you need to: Design reliable unsubscribe and suppression propagation across [TOOLS], including audit history and prevention of future sends.
Act as a senior AI automation consultant and practical implementation partner. Task: Design reliable unsubscribe and suppression propagation across [TOOLS], including audit history and prevention of future sends. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
CRM UAT plan
Use this when you need to: Provide user acceptance tests for [CRM AUTOMATION], including wrong owner, duplicate, opt-out, failed API and partial sync.
Act as a senior AI automation consultant and practical implementation partner. Task: Provide user acceptance tests for [CRM AUTOMATION], including wrong owner, duplicate, opt-out, failed API and partial sync. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 07 · 10 prompts
Marketing & Content Operations
Coordinate publishing while protecting editorial quality.
Content intake
Use this when you need to: Create a request intake form for [TEAM]: audience, goal, channel, assets, claims evidence, approvals, deadline and accessibility needs.
Act as a senior AI automation consultant and practical implementation partner. Task: Create a request intake form for [TEAM]: audience, goal, channel, assets, claims evidence, approvals, deadline and accessibility needs. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Brief generator
Use this when you need to: Turn [CAMPAIGN NOTES] into editorial brief: goals, target persona, angle, sources, CTA, voice, exclusions and review responsibilities.
Act as a senior AI automation consultant and practical implementation partner. Task: Turn [CAMPAIGN NOTES] into editorial brief: goals, target persona, angle, sources, CTA, voice, exclusions and review responsibilities. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Calendar workflow
Use this when you need to: Design campaign calendar automation across [TOOLS] with draft, review, scheduled, published and correction states.
Act as a senior AI automation consultant and practical implementation partner. Task: Design campaign calendar automation across [TOOLS] with draft, review, scheduled, published and correction states. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
SEO QA checkpoint
Use this when you need to: Build SEO pre-publication checklist for [PAGE] covering intent, originality, evidence, metadata, links, accessible headings and structured-data eligibility.
Act as a senior AI automation consultant and practical implementation partner. Task: Build SEO pre-publication checklist for [PAGE] covering intent, originality, evidence, metadata, links, accessible headings and structured-data eligibility. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Social repurpose
Use this when you need to: Turn [APPROVED CONTENT] into platform-specific draft variants without inventing facts, with editorial review and channel constraints.
Act as a senior AI automation consultant and practical implementation partner. Task: Turn [APPROVED CONTENT] into platform-specific draft variants without inventing facts, with editorial review and channel constraints. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Asset approval
Use this when you need to: Map asset request, design, legal/brand review, revision, final export and publication with version tracking.
Act as a senior AI automation consultant and practical implementation partner. Task: Map asset request, design, legal/brand review, revision, final export and publication with version tracking. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Analytics digest
Use this when you need to: Create a weekly performance summary workflow using [METRICS] and [DATE RANGE], showing change, sources, limitations and suggested tests.
Act as a senior AI automation consultant and practical implementation partner. Task: Create a weekly performance summary workflow using [METRICS] and [DATE RANGE], showing change, sources, limitations and suggested tests. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
UTM governance
Use this when you need to: Propose standardized UTM naming and validation rules for [CAMPAIGNS], preserving existing analytics conventions.
Act as a senior AI automation consultant and practical implementation partner. Task: Propose standardized UTM naming and validation rules for [CAMPAIGNS], preserving existing analytics conventions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Publishing rollback
Use this when you need to: Design monitoring for failed or incorrect scheduled posts, including alert, pause, correction, audit log and responsible approver.
Act as a senior AI automation consultant and practical implementation partner. Task: Design monitoring for failed or incorrect scheduled posts, including alert, pause, correction, audit log and responsible approver. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Experiment tracker
Use this when you need to: Set up a content A/B test workflow with hypothesis, one primary metric, randomization constraints, duration and interpretation caveats.
Act as a senior AI automation consultant and practical implementation partner. Task: Set up a content A/B test workflow with hypothesis, one primary metric, randomization constraints, duration and interpretation caveats. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 08 · 10 prompts
E-commerce & Customer Operations
Order and support flows with financial safety gates.
Order alert
Use this when you need to: Design new-order processing from [STORE] to [OPERATIONS TOOL], including deduplication, PII minimization and alert escalation.
Act as a senior AI automation consultant and practical implementation partner. Task: Design new-order processing from [STORE] to [OPERATIONS TOOL], including deduplication, PII minimization and alert escalation. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Inventory sync
Use this when you need to: Create an inventory synchronization design for [CHANNELS] with authoritative source, reservation timing, stale data alerts and oversell prevention.
Act as a senior AI automation consultant and practical implementation partner. Task: Create an inventory synchronization design for [CHANNELS] with authoritative source, reservation timing, stale data alerts and oversell prevention. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Abandoned-cart ethics
Use this when you need to: Plan consent-aware recovery for [STORE], with preference checks, frequency caps and clear opt-out.
Act as a senior AI automation consultant and practical implementation partner. Task: Plan consent-aware recovery for [STORE], with preference checks, frequency caps and clear opt-out. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Return workflow
Use this when you need to: Map return request to eligibility check, human approval, labels, updates and refund handoff; never auto-refund without explicit authorization.
Act as a senior AI automation consultant and practical implementation partner. Task: Map return request to eligibility check, human approval, labels, updates and refund handoff; never auto-refund without explicit authorization. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Ticket triage
Use this when you need to: Classify support tickets [SAMPLES] by intent, urgency and sentiment with confidence and review flag; avoid interpreting emotion as fact.
Act as a senior AI automation consultant and practical implementation partner. Task: Classify support tickets [SAMPLES] by intent, urgency and sentiment with confidence and review flag; avoid interpreting emotion as fact. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Shipping notification
Use this when you need to: Design status notifications from [CARRIER] to [CUSTOMER CHANNEL] with verified carrier events, duplicate suppression and exception messages.
Act as a senior AI automation consultant and practical implementation partner. Task: Design status notifications from [CARRIER] to [CUSTOMER CHANNEL] with verified carrier events, duplicate suppression and exception messages. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Product data QA
Use this when you need to: Audit [CATALOG EXPORT] for missing attributes, contradictory descriptions, taxonomy consistency and unsupported claims.
Act as a senior AI automation consultant and practical implementation partner. Task: Audit [CATALOG EXPORT] for missing attributes, contradictory descriptions, taxonomy consistency and unsupported claims. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Escalation playbook
Use this when you need to: Define escalation for lost shipment, payment issue, complaint and potential fraud without exposing customer data in alerts.
Act as a senior AI automation consultant and practical implementation partner. Task: Define escalation for lost shipment, payment issue, complaint and potential fraud without exposing customer data in alerts. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Review moderation
Use this when you need to: Design review routing for spam, safety report, legitimate criticism and response drafting; prohibit fake reviews and suppression of genuine criticism.
Act as a senior AI automation consultant and practical implementation partner. Task: Design review routing for spam, safety report, legitimate criticism and response drafting; prohibit fake reviews and suppression of genuine criticism. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Ops dashboard brief
Use this when you need to: Specify store-ops dashboard for order volume, backlog, inventory mismatches, response SLA and fulfillment failures with metric definitions.
Act as a senior AI automation consultant and practical implementation partner. Task: Specify store-ops dashboard for order volume, backlog, inventory mismatches, response SLA and fulfillment failures with metric definitions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 09 · 10 prompts
Data, Sheets & Reporting
Build dependable ETL and reporting automations.
Data audit
Use this when you need to: Profile [DATASET] for schema, duplicates, nulls, anomalies, source freshness, identifiers and sensitive columns.
Act as a senior AI automation consultant and practical implementation partner. Task: Profile [DATASET] for schema, duplicates, nulls, anomalies, source freshness, identifiers and sensitive columns. Ask for data if not supplied. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
ETL map
Use this when you need to: Design extract-transform-load flow from [SOURCES] to [DESTINATION] with validation, checkpoints, lineage and quarantine for bad rows.
Act as a senior AI automation consultant and practical implementation partner. Task: Design extract-transform-load flow from [SOURCES] to [DESTINATION] with validation, checkpoints, lineage and quarantine for bad rows. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Spreadsheet formula
Use this when you need to: Write [GOOGLE SHEETS/EXCEL] formula for [TASK] using [COLUMN LAYOUT], with example, locale caveat and error handling.
Act as a senior AI automation consultant and practical implementation partner. Task: Write [GOOGLE SHEETS/EXCEL] formula for [TASK] using [COLUMN LAYOUT], with example, locale caveat and error handling. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Record normalization
Use this when you need to: Specify normalization rules for [FIELDS] (dates, names, currencies, status) without silently overwriting raw values.
Act as a senior AI automation consultant and practical implementation partner. Task: Specify normalization rules for [FIELDS] (dates, names, currencies, status) without silently overwriting raw values. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Data reconciliation
Use this when you need to: Design source vs destination reconciliation using counts, hash/key matching, totals, tolerances and exception logs.
Act as a senior AI automation consultant and practical implementation partner. Task: Design source vs destination reconciliation using counts, hash/key matching, totals, tolerances and exception logs. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Incremental sync
Use this when you need to: Plan incremental updates using [UPDATED_AT/ID] with overlap window, late-arriving events, deletes and safe backfill.
Act as a senior AI automation consultant and practical implementation partner. Task: Plan incremental updates using [UPDATED_AT/ID] with overlap window, late-arriving events, deletes and safe backfill. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
KPI definition
Use this when you need to: Define [KPI LIST] with exact numerator, denominator, filters, timeframe, timezone and limitations.
Act as a senior AI automation consultant and practical implementation partner. Task: Define [KPI LIST] with exact numerator, denominator, filters, timeframe, timezone and limitations. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Report narrative
Use this when you need to: Turn [VERIFIED METRICS] into an executive summary distinguishing observations, hypotheses and recommended actions.
Act as a senior AI automation consultant and practical implementation partner. Task: Turn [VERIFIED METRICS] into an executive summary distinguishing observations, hypotheses and recommended actions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Anomaly alert
Use this when you need to: Specify anomaly detection for [METRIC] with baseline, seasonality, threshold, false positives and incident response.
Act as a senior AI automation consultant and practical implementation partner. Task: Specify anomaly detection for [METRIC] with baseline, seasonality, threshold, false positives and incident response. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Backfill strategy
Use this when you need to: Prepare a backfill procedure for missing [DATE RANGE] records: estimate scope, dry run, idempotency and reconciliation.
Act as a senior AI automation consultant and practical implementation partner. Task: Prepare a backfill procedure for missing [DATE RANGE] records: estimate scope, dry run, idempotency and reconciliation. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 10 · 10 prompts
Testing, Debugging & Reliability
Find defects before they affect users.
Test matrix
Use this when you need to: Build test matrix for [WORKFLOW] covering happy path, boundaries, empty inputs, duplicates, auth expiration, downtime and replays.
Act as a senior AI automation consultant and practical implementation partner. Task: Build test matrix for [WORKFLOW] covering happy path, boundaries, empty inputs, duplicates, auth expiration, downtime and replays. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Fixture generator
Use this when you need to: Generate synthetic test payloads matching [SCHEMA] including valid, malformed and edge cases.
Act as a senior AI automation consultant and practical implementation partner. Task: Generate synthetic test payloads matching [SCHEMA] including valid, malformed and edge cases. Use fictional data only. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Bug triage
Use this when you need to: Analyze [ERROR LOG + EXPECTED BEHAVIOR] to produce probable causes, reproduction steps, tests and confidence; do not claim certainty.
Act as a senior AI automation consultant and practical implementation partner. Task: Analyze [ERROR LOG + EXPECTED BEHAVIOR] to produce probable causes, reproduction steps, tests and confidence; do not claim certainty. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Idempotency review
Use this when you need to: Review [AUTOMATION] for duplicate event risks and propose stable identifiers, write checks and replay tests.
Act as a senior AI automation consultant and practical implementation partner. Task: Review [AUTOMATION] for duplicate event risks and propose stable identifiers, write checks and replay tests. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Retry policy
Use this when you need to: Set retry/backoff strategy for [ACTION] distinguishing transient errors from permanent and non-idempotent operations.
Act as a senior AI automation consultant and practical implementation partner. Task: Set retry/backoff strategy for [ACTION] distinguishing transient errors from permanent and non-idempotent operations. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Dead-letter queue
Use this when you need to: Design failed-job quarantine with payload minimization, reason, retry eligibility, alert owner and retention.
Act as a senior AI automation consultant and practical implementation partner. Task: Design failed-job quarantine with payload minimization, reason, retry eligibility, alert owner and retention. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Regression plan
Use this when you need to: Create regression tests after [CHANGE] focused on impacted paths and critical untouched workflows.
Act as a senior AI automation consultant and practical implementation partner. Task: Create regression tests after [CHANGE] focused on impacted paths and critical untouched workflows. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Load test brief
Use this when you need to: Plan ethical load testing on authorized [SYSTEM] with representative volume, safe limits, rollback and success metrics.
Act as a senior AI automation consultant and practical implementation partner. Task: Plan ethical load testing on authorized [SYSTEM] with representative volume, safe limits, rollback and success metrics. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Monitoring design
Use this when you need to: Define logs, traces, correlation IDs, metrics and actionable alerts for [AUTOMATION], excluding secrets and excess PII.
Act as a senior AI automation consultant and practical implementation partner. Task: Define logs, traces, correlation IDs, metrics and actionable alerts for [AUTOMATION], excluding secrets and excess PII. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Incident postmortem
Use this when you need to: Write a blameless incident report from [TIMELINE/LOGS]: impact, detection, causes, mitigation, lessons and owners.
Act as a senior AI automation consultant and practical implementation partner. Task: Write a blameless incident report from [TIMELINE/LOGS]: impact, detection, causes, mitigation, lessons and owners. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 11 · 10 prompts
Security, Privacy & Governance
Keep automations safe and auditable.
Data classification
Use this when you need to: Classify fields in [WORKFLOW] by public, internal, confidential and regulated according to [POLICY], with retention and access recommendations.
Act as a senior AI automation consultant and practical implementation partner. Task: Classify fields in [WORKFLOW] by public, internal, confidential and regulated according to [POLICY], with retention and access recommendations. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Least privilege
Use this when you need to: Review access scopes for [INTEGRATIONS]; suggest minimum permissions, owners, credential rotation and separation of environments.
Act as a senior AI automation consultant and practical implementation partner. Task: Review access scopes for [INTEGRATIONS]; suggest minimum permissions, owners, credential rotation and separation of environments. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Secret management
Use this when you need to: Design handling for API keys, OAuth tokens and webhook secrets in [PLATFORM] without hardcoding or sharing raw credentials.
Act as a senior AI automation consultant and practical implementation partner. Task: Design handling for API keys, OAuth tokens and webhook secrets in [PLATFORM] without hardcoding or sharing raw credentials. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Privacy impact
Use this when you need to: Create preliminary privacy-impact assessment for [AUTOMATION]: data subjects, purpose, lawful basis questions, processors, transfers and mitigations.
Act as a senior AI automation consultant and practical implementation partner. Task: Create preliminary privacy-impact assessment for [AUTOMATION]: data subjects, purpose, lawful basis questions, processors, transfers and mitigations. Flag legal review. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Approval matrix
Use this when you need to: Classify automation actions as safe auto-execute, notify-only, human approve or prohibited, based on reversibility and impact.
Act as a senior AI automation consultant and practical implementation partner. Task: Classify automation actions as safe auto-execute, notify-only, human approve or prohibited, based on reversibility and impact. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Audit log schema
Use this when you need to: Define audit log for [WORKFLOW] with event ID, actor, action, version, timestamp, status and redaction requirements.
Act as a senior AI automation consultant and practical implementation partner. Task: Define audit log for [WORKFLOW] with event ID, actor, action, version, timestamp, status and redaction requirements. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Retention policy
Use this when you need to: Propose retention and deletion workflow for [DATA TYPE] based on [ORGANIZATION POLICY], including backups and vendor propagation.
Act as a senior AI automation consultant and practical implementation partner. Task: Propose retention and deletion workflow for [DATA TYPE] based on [ORGANIZATION POLICY], including backups and vendor propagation. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Prompt data protection
Use this when you need to: Audit [AI PROMPT/WORKFLOW] for accidental disclosure of secrets, customer PII and confidential business data; suggest redacted substitutions.
Act as a senior AI automation consultant and practical implementation partner. Task: Audit [AI PROMPT/WORKFLOW] for accidental disclosure of secrets, customer PII and confidential business data; suggest redacted substitutions. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Vendor assessment
Use this when you need to: Create due-diligence questionnaire for [AUTOMATION VENDOR]: security, subprocessors, export, uptime, logging, deletion and contract review.
Act as a senior AI automation consultant and practical implementation partner. Task: Create due-diligence questionnaire for [AUTOMATION VENDOR]: security, subprocessors, export, uptime, logging, deletion and contract review. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Threat model
Use this when you need to: Threat-model [AUTOMATION] for spoofed events, leaked tokens, privilege escalation, injected content, replay and data exfiltration.
Act as a senior AI automation consultant and practical implementation partner. Task: Threat-model [AUTOMATION] for spoofed events, leaked tokens, privilege escalation, injected content, replay and data exfiltration. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Module 12 · 10 prompts
Delivery, Pricing & Portfolio
Package work clients can assess and maintain.
Feasibility memo
Use this when you need to: Draft a feasibility memo for [CLIENT PROJECT] including objective, solution options, dependencies, risk, assumptions and go/no-go recommendation.
Act as a senior AI automation consultant and practical implementation partner. Task: Draft a feasibility memo for [CLIENT PROJECT] including objective, solution options, dependencies, risk, assumptions and go/no-go recommendation. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Proposal draft
Use this when you need to: Write scoped automation proposal from [DISCOVERY NOTES] with deliverables, exclusions, milestones, acceptance tests, client duties and change control.
Act as a senior AI automation consultant and practical implementation partner. Task: Write scoped automation proposal from [DISCOVERY NOTES] with deliverables, exclusions, milestones, acceptance tests, client duties and change control. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Effort estimate
Use this when you need to: Break down [PROJECT] into discovery, build, test, deployment, documentation and support.
Act as a senior AI automation consultant and practical implementation partner. Task: Break down [PROJECT] into discovery, build, test, deployment, documentation and support. Estimate ranges with assumptions, not invented guarantees. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
ROI model
Use this when you need to: Make a transparent ROI model from [BASELINE VOLUME], [TIME SAVED], [HOURLY COST], platform cost and maintenance; show sensitivity and non-financial risks.
Act as a senior AI automation consultant and practical implementation partner. Task: Make a transparent ROI model from [BASELINE VOLUME], [TIME SAVED], [HOURLY COST], platform cost and maintenance; show sensitivity and non-financial risks. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Statement of work
Use this when you need to: Create SOW for [AUTOMATION] specifying systems, access, integrations, output, acceptance criteria, ownership, security and support boundaries.
Act as a senior AI automation consultant and practical implementation partner. Task: Create SOW for [AUTOMATION] specifying systems, access, integrations, output, acceptance criteria, ownership, security and support boundaries. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Demo script
Use this when you need to: Build a 5-minute demo for [AUTOMATION] including before/after, sample input, live trigger, edge case, logs and measurable benefit if verified.
Act as a senior AI automation consultant and practical implementation partner. Task: Build a 5-minute demo for [AUTOMATION] including before/after, sample input, live trigger, edge case, logs and measurable benefit if verified. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Runbook
Use this when you need to: Write operational runbook for [WORKFLOW]: owner, start/stop, monitoring, known failures, manual fallback, escalation, rollback and contacts.
Act as a senior AI automation consultant and practical implementation partner. Task: Write operational runbook for [WORKFLOW]: owner, start/stop, monitoring, known failures, manual fallback, escalation, rollback and contacts. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Handover checklist
Use this when you need to: Prepare handoff checklist with exported workflow, account ownership, secrets transfer procedure, documentation, training and sign-offs.
Act as a senior AI automation consultant and practical implementation partner. Task: Prepare handoff checklist with exported workflow, account ownership, secrets transfer procedure, documentation, training and sign-offs. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Case study template
Use this when you need to: Turn [ACTUAL PROJECT EVIDENCE] into honest case study: challenge, approach, tech stack, screenshots, test evidence, verified results and limitations.
Act as a senior AI automation consultant and practical implementation partner. Task: Turn [ACTUAL PROJECT EVIDENCE] into honest case study: challenge, approach, tech stack, screenshots, test evidence, verified results and limitations. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
Interview rehearsal
Use this when you need to: Act as remote AI automation hiring manager.
Act as a senior AI automation consultant and practical implementation partner. Task: Act as remote AI automation hiring manager. Ask 12 progressively harder questions on [PORTFOLIO], troubleshooting, security and stakeholder communication; evaluate my answers with rubric. Project context: [DESCRIBE YOUR BUSINESS, TOOLS, USERS AND GOAL]. Constraints: [BUDGET, DEADLINE, DATA POLICY, REGION, PLATFORM VERSIONS]. Output requirements: 1. Start with what is known, assumptions and the smallest set of necessary clarifying questions. Continue with a clearly labeled provisional solution if answers are missing. 2. Give a usable deliverable specific to this task, with numbered steps, examples or tables where appropriate. 3. Include failure/edge cases, security and privacy controls, a human approval step for consequential or irreversible actions, and a simple verification/test checklist. 4. Clearly distinguish illustrative configuration from verified platform functionality. Do not invent data, API capabilities, metrics or client results. 5. Close with acceptance criteria and the next implementation action.
Replace placeholders · verify before deployment
No prompts matched. Try a different keyword or select “All 120”.
Full project walkthrough: build a responsible lead-intake automation
- Use /AUTO-001, /AUTO-004 and /AUTO-008 to learn requirements, choose the opportunity and agree scope.
- Use /AUTO-011, /AUTO-012 and /AUTO-017 to define trigger, field contract and approval gates.
- Use /AUTO-021, /AUTO-022 or /AUTO-023 for a platform-specific blueprint, then /AUTO-035 for mapping.
- Use /AUTO-061, /AUTO-062, /AUTO-064 and /AUTO-069 for consent, routing, duplicate logic and opt-outs.
- Use /AUTO-091, /AUTO-094, /AUTO-099 and /AUTO-105 for testing, replay protection, monitoring and credential handling.
- Use /AUTO-112, /AUTO-116 and /AUTO-117 for proposal, demo and runbook.
Acceptance standard: sample input produces one authorized CRM record; duplicates do not double-create; invalid consent stops sends; failed sync alerts an owner; secrets never enter logs.
Frequently asked questions
Are these shortcodes native AI commands?
No. They are index labels for locating and reusing the complete prompts in this handbook.
Do I need paid automation software?
Not necessarily. Choose based on verified connector availability, hosting, security, support and total costs; a prototype can use synthetic data.
Can I use the examples in an AI remote-job portfolio?
Yes, as personal or practice projects when labeled honestly. Never present sample outcomes as real client achievements.
Will copying a prompt create a deployed automation?
No. The prompt produces a plan or draft, which still requires implementation, access controls, tests and review.