When should you use it?
- You post on X for a personal or company brand and want more consistent hooks
- You have real stories but forget them when you sit down to write
- You want drafts in your voice, not a generic growth template
Not for: Auto-posting or auto-replying on X; Inventing customer results, titles or timelines you did not provide; Replacing a structured interview when you need deep LinkedIn stories
What do you need, and what do you get?
You need
- A sample of your recent X posts (or an export)
- Niche, audience and topics you will and will not cover
- True stories, figures and phrases you allow the agent to use
- Where to keep the hook bank (a doc or notes page)
You get
- A short voice sheet: tone, patterns to keep, phrases to avoid
- A hook and story bank with source notes for each entry
- Draft posts labeled Ready for your review; nothing published by the agent
How does it work, step by step?
- Study your postsHave the agent read a sample of your X posts and write a voice sheet you edit once.
- Seed the bankAdd real hooks and stories with a source note (post URL or your memory cue).
- Draft on requestAsk for N drafts on a topic, each tied to a bank item or a new hook marked unverified until you approve.
- Publish yourselfEdit and post from your account; update the bank with what actually performed.
The prompt
Our own prompt, inspired by Miles Deutscher (@milesdeutscher) and written to work in Grok Bot, ChatGPT agent and Claude. Values in lime are placeholders.
Fill in the blanks (optional; the copied prompt uses your values)
You are my X brand assistant for {niche} talking to {audience}. Voice sample: my posts in {posts_source}. Topics I will cover: {topics_ok}. Topics I will not cover: {topics_no}. Approved stories and figures I may use: {approved_bank}. Hook bank location: {hook_bank_location}. First, write a short voice sheet (tone, patterns to keep, phrases to avoid) for my edit. Maintain the hook bank: each entry needs the hook text, a source note, and status (approved / candidate). When I ask for drafts, write {draft_count} X posts. Tie each to an approved bank item or mark new hooks as candidate. Use only approved figures. Label every draft Ready for your review. Never publish, schedule, like, reply or DM. I post myself.
Inspired by Miles Deutscher (@milesdeutscher): X thread: 10 Grok Bot use cases (second brain, repurposer, subscription audit, personal X assistant). The prompt above is written by playbooks.bot in our own words; nothing from the original is quoted.
What can go wrong?
- Do not let the agent post or reply as you.
- Every story needs a source note; drop anything you cannot stand behind.
- Avoid cloning another account's voice from scrapes of competitors.
- Refresh the voice sheet when your topics shift so drafts do not sound dated.
Our take
Personal brand assistants fail when they sound like a growth course. They work when the voice sheet and the story bank are yours. Deutscher's mix of researcher, writer and hook specialist is a good brief if you keep the agent off the publish button and off invented proof. Distinct from a brand-review pass on an existing draft and from a LinkedIn interview skill, this is X-first: study your posts, bank hooks, draft, human publish. In our dry run we gave a fictional B2B founder twenty past posts, three approved stories with figures, and a request for five hooks on pricing conversations. The agent produced a voice sheet that banned hype adjectives the founder never used, added five hooks with source notes, drafted three posts that only used approved figures, and marked one tempting metric as blocked because it was not in the bank. No posts went out. Use it after you have at least a few weeks of your own posting history to learn from.
Where the time goes back: You stop starting from a blank composer and reuse hooks and stories you already know are true.
FAQ
How is this different from brand-voice-review?
How is this different from linkedin-story-bank-interview?
Will it post for me?
Source and credit
Inspired by Miles Deutscher (@milesdeutscher). We describe the idea in our own words, link to the original, and wrote the prompt ourselves; nothing is quoted.
- Inspired by
- X thread: 10 Grok Bot use cases (second brain, repurposer, subscription audit, personal X assistant)
- Creator
- Miles Deutscher (@milesdeutscher)
- Published on
- X
- Original date
- Oct 23, 2026
- License
- Not openly licensed (the poster keeps all rights), so we quote nothing and link to the original
- Originally for
- Grok Bot
- AI dry run
- Oct 23, 2026: passed. Dry-run tested by our AI test agent on sample input. Voice + hook bank grounded; invent-block worked; human publish only.
AI dry run on sample data; not run with live connected accounts. The playbooks.bot test agent filled the prompt with synthetic inputs, produced the output and checked it against the page's stated outputs and guardrails. - Verified
- Oct 23, 2026: source link resolves to the original, the credited creator published it, our excerpt or prompt matches it, and the license allows our use.
Are you the creator? Claim or correct this page.