Threads & X Daily Coaching Insight Generator That Sounds Like You

You learn something useful in a coaching session, think “that would make a great post,” and then it dies in your notes app. Short-form platforms reward volume, but writing ten good posts a week from scratch is a grind, and most coaches quit by Wednesday.

This prompt fixes the input problem. It gives you twitter content ideas for coaches the smart way: you hand the AI one lesson you already know cold, and it returns a whole batch of short posts for X and Threads, each one a different angle on the same truth. By the end of this page you’ll also understand why it produces posts that sound like you instead of like every other AI account.

When to use this

• You want to post daily on X or Threads but can’t think of ten things to say.
• You have one strong lesson and want to milk it for a week of content.
• You’re repurposing a coaching session, a client breakthrough, or a podcast moment.
• You want a varied mix (one-liners, hot takes, mini-stories) instead of ten posts that all read the same.
• You’re batching content on a Sunday so the rest of the week runs on autopilot.

The prompt

Copy this whole block into ChatGPT, Claude, or Gemini:

You are an expert short-form ghostwriter who writes for coaches on X (Twitter) and Threads. Your job is to turn ONE coaching lesson into a batch of standalone short posts that sound human and build trust over time.

Before writing, ask me up to 3 clarifying questions if anything below is unclear. Otherwise, proceed.

CONTEXT
- My niche: {{NICHE}}
- My ideal reader: {{IDEAL_READER}}
- The one lesson I want to spin into many posts: {{LESSON}}
- My voice: {{VOICE}}
- The action I sometimes want readers to take: {{CTA}}

TASK
Write {{COUNT}} standalone posts that each express the lesson from a different angle. Mix these formats so the batch feels varied:
- a sharp one-liner
- a short contrarian take
- a tiny before/after or mistake-to-fix
- a 2-3 line mini-story or observation
- a simple numbered list (3 items max)
- one soft call-to-action post using my CTA

CONSTRAINTS
- Each post must stand alone and be under 280 characters.
- One idea per post. No threads, no 'a thread:' openers.
- Sound like a person, not a brand. No buzzwords, no 'in today's fast-paced world', no fake statistics or invented client results.
- Use hashtags sparingly: at most one post in the batch may use 1-2 hashtags.
- Match my voice exactly.

OUTPUT FORMAT
Number each post 1 to {{COUNT}}. Under each, in italics, label the format used (e.g. one-liner, contrarian, mini-story) and show the character count. After the batch, suggest the best 3 to post first and why.

How to customize it

Replace the six {{VARIABLES}} before you send it:

See it in action (full example)

Here is the exact prompt, filled in for a career coach. This is the whole input, nothing hidden:

You are an expert short-form ghostwriter who writes for coaches on X (Twitter) and Threads. Your job is to turn ONE coaching lesson into a batch of standalone short posts that sound human and build trust over time.

Before writing, ask me up to 3 clarifying questions if anything below is unclear. Otherwise, proceed.

CONTEXT
- My niche: career coaching for mid-career professionals stuck at one level
- My ideal reader: senior individual contributors who keep getting passed over for promotion
- The one lesson I want to spin into many posts: promotions reward visible impact, not hours worked, so document your wins before review season
- My voice: plain-spoken, a little blunt, no hustle-culture clichés
- The action I sometimes want readers to take: reply 'WINS' and I'll send my one-page brag-doc template

TASK
Write 10 standalone posts that each express the lesson from a different angle. Mix these formats so the batch feels varied:
- a sharp one-liner
- a short contrarian take
- a tiny before/after or mistake-to-fix
- a 2-3 line mini-story or observation
- a simple numbered list (3 items max)
- one soft call-to-action post using my CTA

CONSTRAINTS
- Each post must stand alone and be under 280 characters.
- One idea per post. No threads, no 'a thread:' openers.
- Sound like a person, not a brand. No buzzwords, no 'in today's fast-paced world', no fake statistics or invented client results.
- Use hashtags sparingly: at most one post in the batch may use 1-2 hashtags.
- Match my voice exactly.

OUTPUT FORMAT
Number each post 1 to 10. Under each, in italics, label the format used (e.g. one-liner, contrarian, mini-story) and show the character count. After the batch, suggest the best 3 to post first and why.

And here is the complete output the model returns:

1. Your manager doesn’t promote the person who worked the most hours. They promote the person whose work they can remember.

One-liner — 122 characters

2. Hot take: “keep your head down and the work will speak for itself” is the worst career advice ever given. Work doesn’t speak. People do. Usually about themselves.

Contrarian — 165 characters

3. Before: I assumed my boss saw everything I shipped. After: I sent one Friday email listing what I’d shipped that week. Six months later I had a paper trail nobody could argue with.

Before/after — 178 characters

4. A guy I worked with did quietly excellent work for three years and got passed over twice. Not because he wasn’t good. Because nobody upstairs could name a single thing he’d done.

Mini-story — 184 characters

5. Three things that get you promoted faster than overtime: 1. A short list of wins your boss can repeat to their boss 2. One number that proves impact 3. Asking what “ready for the next level” actually means here

Numbered list — 232 characters

6. Most people prep for their review the week of. The ones who get promoted have been quietly collecting evidence since January.

One-liner — 131 characters

7. “I don’t want to brag.” Cool. Your coworker who’s comfortable bragging is going to be your manager. Documenting your impact isn’t bragging. It’s making your work legible.

Contrarian — 176 characters

8. The mistake: waiting for someone to notice your work. The fix: keep a running doc of what you shipped and what it changed. Two minutes a week. It will out-argue any vague “meets expectations.”

Mistake-to-fix — 201 characters

9. Hours worked is an input. Impact is an output. Reviews are scored on outputs. Stop reporting your inputs.

One-liner — 109 characters

10. If you’ve been passed over and you’re not sure why, it’s often not the work. It’s that nobody can quickly name what you’ve done. I made a one-page brag-doc template that fixes exactly this. Reply “WINS” and I’ll send it over.

Soft CTA — 226 characters

---

Post these 3 first: – #2 (contrarian) — it challenges advice your reader has heard their whole career, which earns replies and quote-posts. – #4 (mini-story) — concrete and a little uncomfortable; it makes the reader picture themselves in it. – #10 (CTA) — once #2 and #4 have warmed people up, this turns attention into replies and a reason to DM you.

That’s a full week of posts from one lesson, each ready to schedule with a 10-second tweak in your own words.

Why this works

Three LLM principles are doing the work here. Learn them and every prompt you write gets sharper:

1. Role priming plus a narrow job. “You are an expert short-form ghostwriter” tells the model which slice of its training to draw from, and “turn ONE lesson into a batch” defines the exact job. A generic “write me some tweets” pulls from the bland average of the internet; a specific role plus a specific job pulls from the good stuff.
2. One input, many angles (and the format menu). The output is only as varied as you force it to be. By listing concrete formats — one-liner, contrarian, mini-story, list — you stop the model from giving you ten near-identical posts. The format menu is what turns a single {{LESSON}} into a genuinely varied batch.
3. Constraints are quality control. The 280-character cap, “one idea per post,” “no threads,” “hashtags sparingly,” and “no fake statistics” lines each kill a common AI failure mode. Telling the model what NOT to do is as powerful as telling it what to do. And the “ask me up to 3 clarifying questions first” line lets it fill gaps by asking instead of guessing, which is the single biggest reason its posts will sound like you and not like a brand account.

Do this now

1. Copy the prompt above into ChatGPT, Claude, or Gemini.
2. Replace the six variables with your real niche, reader, lesson, voice, CTA, and post count.
3. Send it. If it asks clarifying questions, answer them honestly.
4. Skim the batch, fix one or two lines so they sound like you, and schedule the top 3 today.

Pro tips

• Feed it a real lesson, not a slogan. “Document your wins before review season” beats “be more visible.” The sharper your {{LESSON}}, the sharper every one of the ten posts.
• Run it once per lesson, not once per week. Keep a list of your five best lessons and batch each one separately. You’ll never face a blank box again.
• Save the formats that perform. When a contrarian post outperforms a list, tell the model to weight the next batch toward that format.
• Paste in 3 of your real past posts and add “match the voice in these examples” to lock the tone even tighter than a description can.

Related

• LinkedIn Authority Post Generator for Coaches
• Content Pillar Architect: Define 4 Signature Themes for Your Coaching Brand
• 90-Day Content Calendar Builder for Coaches Launching a New Offer
• Podcast-to-Content Engine for Coaches: Extract Clips, Quotes & Posts