AI Agents & Automation
Agents and automations are only as good as the instructions, tools and guardrails behind them. These prompts help you map a repetitive task into a workflow, decide between a script, a no-code flow and an agent, write system prompts and tool descriptions, and plan evaluation and safety checks. You will find designs for email triage, meeting-to-task pipelines, invoice handling and research assistants, each with human approval points and clear limits. Fill in the placeholders with your own process, test with harmless data first, and keep a person in the loop for anything that sends, pays or deletes. The prompts draft plans and specifications; they do not build or run the automation for you.
Start with these
A short list of well-specified prompts from this page, picked by our checklist score (specific inputs, clear output format, concrete steps). It is a heuristic, not a verdict.
-
Risk Register for an AI Feature
Create a risk register for an AI feature that lists what can go wrong, how likely it is and what reduces each risk.
-
Decision Log for Automated and AI-Assisted Decisions
Design a decision log for AI-assisted decisions with fields for suggestion, evidence, human choice, override reason and later outcome.
-
Agent Pilot Plan: Safe Rollout to a Small Group
Plan an AI agent pilot for a small group, with success measures, safeguards and a clear stop rule before wider rollout.
-
Inbox Triage Agent Specification With Human Approval Points
Specify an inbox triage agent that sorts email, drafts replies and asks for human approval before it sends anything.
-
Bias and Fairness Check for a Screening Workflow
Look for uneven outcomes in an AI-assisted screening process.
-
Document Intake and Filing Assistant
Name, tag and file incoming documents consistently.
Prompts
-
System Prompt for a Single-Purpose Assistant
Write a clear system prompt with scope, tone, limits and fallback behaviour.
-
Tool Description Writer for Function Calling
Write tool names, descriptions and parameters a model can use correctly.
-
Research Agent Instructions With Source Rules
Instruct an agent to research a question and cite what it found.
-
Planner and Executor Prompt Pair
Split a task into a plan step and an execution step with checks between them.
-
Memory and Context Handover Prompt
Summarise a long working session so a fresh assistant can continue it.
-
Browser Task Agent Instructions With Safety Limits
Tell a web agent how to complete a task and what it must never do.
-
Data Cleaning Agent Prompt With Audit Trail
Clean a table while recording every change.
-
Scheduling Assistant Prompt With Conflict Handling
Instruct an assistant to propose meeting times without double booking.
-
Agent Handoff Note Between Two Tools
Write the message one agent passes to the next so work is not lost.
-
System Prompt for a Scheduling Assistant With Clear Limits
Write a system prompt that is specific about tools, tone and refusal behaviour.
-
Tool Description Writer: Clear Schemas for an Agent's Functions
Write function descriptions an agent can use correctly.
-
Memory Notes Design for a Long-Running Assistant
Decide what an assistant should remember, update and forget.
-
Agent Hand-Off Prompt: Passing Work Between Specialist Agents
Write the message format one agent sends to another so nothing is lost.
-
Planner and Executor Prompts for a Multi-Step Task
Separate planning from doing so an agent's steps can be reviewed.
-
Retrieval Instructions: How an Agent Should Use Search Results
Write instructions that make an assistant ground answers in retrieved text.
-
Error Handling Instructions for an Assistant That Uses Tools
Tell an agent exactly what to do when a tool fails or returns nonsense.
-
Confirmation-Before-Action Prompt Pattern
Write a pattern that makes an agent show exactly what it is about to do.
-
Persona and Boundary Prompt for a Customer-Facing Assistant
Write a prompt that sets tone and limits without sounding robotic.
-
Output Format Contracts: Make Agent Responses Machine-Safe
Define strict output formats and what to do when the model breaks them.
-
Regression Check Before Changing a Prompt
Compare old and new prompt outputs on the same inputs.
-
Hallucination Check for Summaries
Compare a model summary with its source and flag unsupported statements.
-
Prompt Injection Test Cases for My Assistant
Write defensive test inputs to see whether an assistant ignores hidden instructions.
-
Human Review Sampling Plan for Automated Output
Decide how many automated results a person should check and how often.
-
Failure Log and Root Cause Template for AI Workflows
Record each failure so patterns point to the real fix.
-
Compare Two Models on My Own Tasks
Run a fair side-by-side test using your real work, not benchmarks.
-
Acceptance Criteria for an Automation Project
Write measurable pass conditions before building starts.
-
Monitoring Dashboard Spec for a Live AI Feature
List the signals to watch once users depend on the feature.
-
Agent Test Case Suite Designer
Build a set of test cases that cover normal use, edge cases and failures.
-
Defensive Red-Team Checklist for a Customer Support Agent
Test your own agent for prompt injection and data leakage, safely.
-
Evaluation Rubric for Agent Answers
Write a rubric that scores helpfulness, accuracy, safety and honesty.
-
Golden Examples Set: Collecting and Maintaining Reference Answers
Build a small set of reference interactions and keep it current.
-
Regression Log for Prompt Changes
Track every prompt change and its effect on test results.
-
Agent Pilot Plan: Safe Rollout to a Small Group
Plan an AI agent pilot for a small group, with success measures, safeguards and a clear stop rule before wider rollout.
-
AI Use Policy for a Small Team
Write plain rules on what staff may and may not do with AI tools.
-
Risk Register for an AI Feature
Create a risk register for an AI feature that lists what can go wrong, how likely it is and what reduces each risk.
-
Human-in-the-Loop Approval Design
Decide where a person must approve, and make approval easy to do well.
-
Data Handling Rules for Prompts
Define what may be pasted into AI tools and how to mask the rest.
-
Model Output Disclosure Wording
Write honest notices when content or decisions involve AI.
-
Vendor Questionnaire for an AI Tool
Ask the right questions before adopting a third-party AI service.
-
Bias and Fairness Check for a Screening Workflow
Look for uneven outcomes in an AI-assisted screening process.
-
Incident Response Plan for an AI Mistake
Prepare steps for when an automated output causes harm or embarrassment.
-
Permissions Review for Connected Tools
Check what an assistant can reach and trim it to what it needs.
-
Decision Log for Automated and AI-Assisted Decisions
Design a decision log for AI-assisted decisions with fields for suggestion, evidence, human choice, override reason and later outcome.
-
Acceptable Use Rules for Customer-Facing AI Features
Define what your own AI feature will and will not do for customers.
-
Agent Permission Matrix and Least-Privilege Review
Define exactly what each agent may access and change.
-
Audit Log Specification for AI Actions
Define what an assistant must log so actions can be reviewed later.
-
Human Escalation Rules for an AI Assistant
Decide when the assistant must stop and hand over to a person.
-
Data Handling Rules for Assistants: What Goes In, What Stays Out
Write plain rules about the data an assistant may process.
-
Kill Switch and Rollback Plan for Automations
Plan how to stop an automation quickly and undo its effects.
-
Responsible Use Policy for Teams Using AI Assistants
Draft a short, usable policy for staff.
-
Map a Repetitive Task Into an Automation Spec
Turn a manual routine into a step-by-step spec an automation tool can follow.
-
Choose Between a Script, a No-Code Flow and an AI Agent
Decide the simplest tool that can do the job reliably.
-
Email Triage Workflow With Human Approval
Design a flow that sorts incoming email and drafts replies for review.
-
Meeting-to-Tasks Pipeline
Turn meeting transcripts into owners, deadlines and follow-up messages.
-
Invoice and Receipt Processing Flow
Plan an automation that reads invoices and prepares entries for review.
-
Customer Inquiry Routing Flow
Route form and chat inquiries to the right team with a clear first reply.
-
Content Repurposing Workflow
Automate turning one long piece into several formats with review points.
-
Weekly Report Compilation Flow
Collect numbers from several sources into one report with checks.
-
Lead Follow-Up Sequencing With Guardrails
Plan timed follow-ups to new leads without spamming or overpromising.
-
Document Intake and Filing Assistant
Name, tag and file incoming documents consistently.
-
Social Listening Digest Flow
Summarise mentions into a daily digest with an escalation rule.
-
Onboarding Checklist Automation for New Hires
Trigger tasks, accounts and welcome messages from a single hire record.
-
Inbox Triage Agent Specification With Human Approval Points
Specify an inbox triage agent that sorts email, drafts replies and asks for human approval before it sends anything.
-
Meeting-to-Action-Items Agent Workflow
Design a workflow that turns a transcript into owners, dates and follow-ups.
-
Research Assistant Agent Loop With Source Tracking
Define a plan, search, read, note and verify loop that never invents a source.
-
Customer Email Drafting Agent With Policy Lookup and Escalation Rules
Specify a support drafting agent that checks policy before it writes.
-
Spreadsheet Reconciliation Agent Plan
Plan an agent that matches two data sources and explains every mismatch.
-
Weekly Report Agent: From Raw Updates to a One-Page Summary
Design an agent that collects updates and writes a consistent weekly report.
-
Lead Qualification Agent Rubric and Hand-Off Rules
Specify how an agent scores inbound leads and when it must hand off.
-
Document Intake Agent for Forms, Receipts and Contracts
Plan how an agent extracts fields from documents and flags uncertainty.
-
Personal Admin Agent Boundaries and Permissions Plan
Decide what a personal assistant agent may read, draft and do.
-
Multi-Step Trip Planning Agent With Budget Guardrails
Design an agent that plans a trip in stages and checks costs as it goes.
-
Content Repurposing Agent: One Source, Many Formats
Plan an agent that turns one long piece into posts, an email and a script, while keeping facts straight.
-
Support Ticket Routing Agent With Confidence and Fallbacks
Design a router that assigns tickets to the right team and says when it is unsure.
-
Invoice Processing Agent With Approval Thresholds
Plan how an agent reads invoices, checks them and routes approvals.
-
Calendar and Scheduling Agent Rules
Write the rules for an agent that proposes meetings and protects focus time.
-
Data Cleaning Agent: Plan, Rules and a Change Log
Plan an agent that cleans a messy table without silently changing meaning.
-
Hiring Screen Assistant With Fairness Guardrails
Design an assistant that helps structure applications without making decisions.
No prompts match those filters.