100 Posts

You don't run out of content because you're not creative. You run out because you've been asking AI the lazy way. This turns one prompt into 100 ideas actually worth filming.

Type "give me 100 content ideas" into any AI and you get slop. Tips nobody saves. Hooks that sound like a robot in a blazer. "The power of consistency." "5 tips to boost your reach." You've seen them, you've scrolled past them, and so has your audience.

The problem was never the AI. It was the ask. A generic prompt with no context and no standard gets you generic garbage back, every time. Great ideas come from three things: context (it knows your business), constraints (it models what's already winning instead of inventing filler), and a bar (it kills the slop on sight). This prompt gives it all three.

Run it in order. Prime it once, run the engine, then turn the winners into posts.

Do this once. Every prompt after it gets sharper because it's answering as something that actually knows you.

Save this to memory. My niche: ___. Who I help: ___. What I sell: ___. My 5 content pillars: ___. My voice: candid, real, no corporate tone, no hype. Three posts that did well for me: ___. Remember all of this for every future request.

You're my content strategist and you know my business. Give me 100 short-form content ideas I could film this week. Not tips. Ideas that stop my exact person mid-scroll and make her feel seen.

Rules, follow all of them:
1. Use the 70/20/10 mix: 70 ideas built on hook shapes already going viral in my niche (mistake-callout, POV, "things I don't do anymore," the 5-levels framework, contrarian hot-take, before/after), 20 built on my own story or expertise, 10 weird experiments I'd never think of.
2. No slop. Every hook needs one concrete detail, number, or emotion. Banned: "in today's world," "unlock," "game-changer," "boost your," "X tips to." If a distracted person wouldn't stop, cut it and try again.
3. Open loop on every one. The hook implies a payoff it doesn't give away, so they have to watch or comment to get it.
4. Tag each idea SAVE (teaches a tactic worth keeping) or FOLLOW (personality, story, or tribe that earns the follow). I need both.
5. Write every hook in my voice, using what you know about me.

Give it to me as a table: number, hook (the on-screen text), format (talking head / b-roll / carousel), pillar, SAVE or FOLLOW, and the one-line payoff. Then pick the 10 you'd bet money on and tell me why.

A list of 100 ideas is still homework. These two follow-ups turn the winners into something you can film today.

Take your top 10. For each, write the on-screen hook and a one-line subhook underneath it, the way a scroll-stopping reel stacks them.

Now turn idea number 3 into a 45-second script in my voice. Hook in the first line, one idea, one clear CTA. Keep my phrases where they're good.

It models what's already winning. The 70/20/10 mix stops the AI from inventing generic ideas and forces it to build on hook shapes that are already proven in your space, then bend them to your story. That one rule is the difference between a fresh idea and a recycled one.

It self-scores for two jobs. Tagging each idea SAVE or FOLLOW is the part nobody does. Some content earns a save (the tactic), some earns a follow (you). Knowing which is which is how you grow instead of just getting bookmarked.

It bans the tells. The blocked phrases are the exact words that make content read as AI-written. Cutting them on sight is what keeps it in your voice.

This is only as good as Step 1. An AI with no memory of your business hands back the same slop as everyone else, dressed up in a table. The 100 ideas are downstream of the context you give it. Skip the priming and you're back where you started. That is the whole secret: the prompt isn't magic, the memory behind it is.

If you can't tell whether a hook was written for you or for anyone, it's slop. Every idea that survives should have one specific detail only you could have said. That specificity is what stops the scroll.