Draft a runbook: how to do it with AI
Last checked 2026-10-15
IntermediateBot with human review 🛠️ IT ops & DevOps
Turn a messy procedure into a step-by-step runbook.
1. Bot automation level
Bot with human review The bot does the work, but facts, numbers or code should be checked.
2. Before you hand it to the bot
- Gather inputs: files, data samples, links and any templates.
- Access: make sure you can use the chosen bot/plan (see limits below).
- Context: tell the bot audience, purpose, constraints and examples of good output.
- Privacy check: remove personal, confidential or regulated data unless your plan/workspace is approved for it.
- Success criteria: write down what 'done' looks like (length, format, tests passing, fields filled).
3. Copy-paste prompts
Main prompt
Turn these notes into a runbook with rollback steps.
Beginner-friendly
Turn these notes into a runbook with rollback steps. I'm new to this: explain each step in plain words and ask me for anything you need before starting.
Check-your-work
Turn these notes into a runbook with rollback steps. When you finish, list any assumptions you made, anything you're unsure about, and what I should double-check before using the result.
Free tiers that can handle this (verified): ChatGPT (incl. Work & Codex), Claude (incl. Claude Code), Google Gemini, Mistral Le Chat. Others: check plan.
4. Step-by-step
- Collect the current notes, commands, screenshots and who does each step.
- Ask for a runbook template: purpose, prerequisites, steps, verification, rollback, escalation contacts.
- Have the bot flag gaps or ambiguous steps as questions.
- Walk through the runbook on a test system and fix mistakes.
- Store it in your wiki with an owner and review date.
5. Which bots and plans can do it
| Tool | What's different here | Plan limits | Source |
|---|---|---|---|
| ChatGPT (incl. Work & Codex) | Work can produce finished documents. | Free tier has limited deep research, Codex and Work; scheduled tasks need Go or higher (per chatgpt.com/pricing). | help.openai.com/en/articles/20001275-chatgpt-work-and-codex |
| Claude (incl. Claude Code) | Artifacts/Docs help keep a structured document. | Claude Code and Research need Pro or higher; computer use is Pro/Max only (research preview). | code.claude.com/docs/en/desktop |
| Google Gemini | — | Free tier includes Deep Research, Canvas, Gems; video generation needs AI Plus or higher. | gemini.google/overview/deep-research/ |
| Microsoft Copilot | Drafting inside Word with Microsoft 365. | Copilot in Word/Excel/PowerPoint/Outlook needs an eligible Microsoft 365 plan. | www.microsoft.com/en-us/microsoft-copilot/for-individuals/features |
| Mistral Le Chat | — | Free tier has limited messages, searches and coding sessions; Pro raises limits. | mistral.ai/pricing |
Free or paid?
- ChatGPT (incl. Work & Codex): Free plan: yes. unlimited text chats; limited uploads, images, voice, deep research and Codex. Paid: price not verified, so check the official site. Its free tier can handle this task.
- Claude (incl. Claude Code): Free plan: yes. chat, web search, files & code execution, Artifacts, connectors. Claude Code and Research need Pro. Paid: the lowest price we could verify on the official pricing page is $20/month (other plans may exist). Its free tier can handle this task.
- Google Gemini: Free plan: yes. Flash-Lite model, image generation, Deep Research, Live, Canvas, Gems. Video needs AI Plus or higher. Paid: the lowest price we could verify on the official pricing page is $99.99/month (other plans may exist). Its free tier can handle this task.
- Microsoft Copilot: Free: not verified, so check the official site. Paid: the lowest price we could verify on the official pricing page is $23.5/month (other plans may exist). For this task, check the plan.
- Mistral Le Chat: Free plan: yes. limited messages, searches, coding sessions and basic images. Paid: the lowest price we could verify on the official pricing page is $14.99/month (other plans may exist). Its free tier can handle this task.
6. Risks and what to double-check
- A runbook that hasn't been rehearsed may fail when needed.
- Remove credentials from source notes.
- Test scripts in a non-production environment first; least-privilege credentials.
- AI can hallucinate facts — verify anything that matters.
- Don't share secrets or other people's personal data.