Instagram says AI can build you a business in 24 hours. It can't.
The viral "build a business in 24 hours" prompt list produces ten beautiful documents and zero customers. Here's where each prompt earns its keep, and where it lies.
Prompt lists for Claude and ChatGPT have been flooding my social media lately. The latest one, maybe you've seen it too, promises a complete business in 24 hours: ten prompts that supposedly walk you from "idea" to "launch." Check your idea. Define your niche. Build your offer. Name your business. Write your sales page. Plan your content. Write your emails. Set your pricing. Launch.
It's slick. It's encouraging. And it's the reason a lot of smart people are going to spend this weekend building beautiful documents for businesses that will never earn a dollar.
I want to walk you through why, because the same logic applies to almost every "AI will do it for you" promise you'll see this year. And why this kind of content isn't making you more productive. It's steering you wrong.
The signal
The thing those prompts produce is artifacts. A named business. A one-page plan. Landing page copy. A 7-day content calendar. Five welcome emails. A pricing recommendation.
Here's the test. Open any business you actually admire. Ask yourself: did that business succeed because the founder had great artifacts, or because the founder had paying customers who kept coming back?
Artifacts are the easy part. Customers are the hard part. AI is genuinely good at the easy part, which is exactly why this content travels so well. It feels like progress.
The breakdown
Before we get to what actually works, let me show you what I mean. Here's every prompt from "AI can build you a business in 24 hours," what it's good for, and where it lies to you:
| The prompt | What it's good for | Where it lies to you |
|---|---|---|
| Idea analysis | Pressure-test your assumptions | AI can't see your market. It pattern-matches from training data. |
| Niche definition | Forces you to write it down | Doesn't replace 10 real customer conversations |
| Offer creation | A v1 draft to react to | The offer that sells is the one customers shape, not the one AI drafts |
| Business names | Brainstorm fuel | The good ones come from domain availability and trademark search, not vibes |
| One-page plan | A useful artifact | Plans without distribution are wishes |
| Landing page copy | A solid starting draft | Copy converts when it echoes customer language you collected, not AI's guesses |
| 7-day content plan | A decent scaffold | Posting isn't an audience. You need a channel that compounds. |
| Welcome sequence | A real time-saver | Only matters once leads exist |
| Pricing | Frameworks, not numbers | Pricing is discovered through conversations and rejections |
| Launch blueprint | Checklist value | Launches work when the audience already exists |
The level-set
Three things to remember the next time a viral prompt list shows up in your feed.
One. AI is excellent at the second half of work: drafting, formatting, scaffolding. It needs raw material to do that, and the raw material has to come from you. Your customer conversations. Your judgment. The specific words your clients use. No prompt can manufacture that. It can only echo it back to you better.
Two. Sharing prompts is useful. Prompt engineering is a real skill. GenAI results vary enormously based on how you prompt. The caution isn't "don't use prompts." It's "don't fall for the fix-all prompt." Before you copy any viral list, ask: what is this actually doing differently from what I've already been doing? Most of the time, the answer is nothing. It's the same advice with new emoji.
Three. The prompts worth your time are the ones that do something you couldn't do alone: synthesize across formats, surface a counter-argument, structure a decision you've been avoiding. Everything else is repackaging.
AI doesn't replace your expertise. It accelerates it. The operators getting the most out of AI right now aren't the ones running the most prompts. They're the ones with the most lived experience to feed in.
Even cleaner. Pivot is a credible source, the mention is in passing, and crediting him properly is the right move. Here's the revised section:
One I'm trying right now
Speaking of prompts that do something you couldn't do alone. Scott Galloway mentioned something called the "Council of 5" Claude skill on a recent Pivot episode. I went looking for it and the Instagram post I landed on claimed it would "stop Claude from hallucinating." Big promise. Exactly the kind of thing this issue is warning you about.
So I looked closer. Here's what it actually is.
Skills are a real Claude feature, released by Anthropic for Pro, Max, Team, and Enterprise users. They let you load reusable instruction sets into Claude so you don't have to re-explain context every time. Think of a skill as a saved playbook Claude follows when a certain kind of task comes up.
"Council of 5" is a community-built skill (not from Anthropic) based on an idea from Andrej Karpathy. When you invoke it, Claude takes whatever decision you're stuck on and runs it through five distinct advisor perspectives, then synthesizes the responses. You get a strategist's view, a skeptic's view, an operator's view, and so on, instead of one averaged answer.
Does it stop hallucination? No. Nothing does. But does it do something I couldn't do on my own? Yes. It surfaces angles I'd miss when I'm too close to a decision.
That's the test. The Council of 5 passes it. Most viral prompt lists don't.
I'll write up how I'm using it in a future issue once I've put it through enough real decisions to tell you what works and what doesn't. For now, if you have a Pro account and you want to try it, search "Council of 5 Claude skill" on GitHub. (And if you don't have a Pro account, ignore this entire paragraph. Free Claude won't run it.)
— Anne-Cécile
P.S. If you want a second pair of eyes on something you're working through in your business, hit reply. Could be strategy, operations, data, or just a decision you've been circling. I read every one.
