AI Prompts for Demand-First Product Building
Fill in the blanks and copy — 8 staged AI prompts that run the content-first playbook behind apps like Stella: find where attention already is, turn content into proven demand, then build the smallest product that captures it.
Fill in the blanks below — the prompt updates live. These work with any AI chat tool (Claude, ChatGPT, Gemini, etc.); nothing here is tied to one specific model.
How it works
This set turns a specific playbook into runnable prompts: the one used by creators who build an audience first and a product second. The claim that makes it worth trying is that building software is now cheap, so the scarce resource is attention — and content is how you both find demand and deliver it.
Run them roughly in order. Prompts 1 and 2 are about finding a repeatable content format where attention already exists. Prompt 3 is the factory — 30 pieces to publish before you build anything. Prompt 4 is the one that decides whether there is a product here at all: it reads your own comments for the moment someone asks "is there an app for this". Prompts 5 and 6 turn a validated friction into the smallest possible app and wire the content straight to it. Prompt 7 keeps you honest about which experiments are working, and Prompt 8 lets you reverse-engineer any app or creator you admire.
Every prompt is told to flag survivorship bias and over-optimism — a "viral" result that was luck, a friction too vague to build against, a plan that depends on a hit rather than a repeatable action. Keep those flags.
For the content side in depth, see the Reels strategy prompt generator. For pricing the product once it exists, see the regional pricing calculator and, for an app store launch, the app store optimization prompts. To start a business without an existing audience, see the online-business prompt generator.
FAQ
What is demand-first (or distribution-first) product building?
It reverses the usual order. Instead of building a product and then hunting for users, you build an audience and publish content first, watch which posts get saved and which comments ask "is there an app for this", and only then build the smallest product that removes that specific friction. The content becomes both the demand test and the distribution channel. It is also called content-led growth or a demand-first startup.
Do these prompts work with ChatGPT, Claude, and Gemini?
Any AI chat tool. They are plain planning and analysis briefs with a fixed output structure — nothing is model-specific. Paste your details, run the prompt, and check the output against what you can actually verify.
Can AI really build the product in two days like the Stella story says?
AI coding tools can compress the build of a genuinely small MVP to a few days, and Stella's founder describes vibe-coding the first version in about two days with the full app taking two to three months. But "cheap to build" is the part AI changed; distribution did not get cheaper. The prompts plan and draft — they do not shoot videos, reply to comments, or grow an audience, which is where the months actually go.
Is the $340K/month figure in the Stella story real?
Treat it as a marketing claim. A headline MRR number does not tell you gross versus net, churn, refunds, app-store fees, how much rode on a single viral spike, or how large the audience was before the product launched. Copy the mechanism — build distribution, validate with content, then monetise a proven behaviour — not the revenue number.
Which prompt should I run first?
Prompt 1 (Attention Map) if you have not picked a niche yet. If you already post content, start with Prompt 2 (Viral Pattern Teardown) on a post that worked and Prompt 4 ("I Wish This Existed" Miner) on your own comments — those two tell you whether there is demand before you spend a day building. Prompts 5 and 6 come once a real friction has surfaced.
How is this different from the online-business prompt generator?
The online-business prompts assume you are deciding what business to start and how to launch it. These assume the product comes last: the sequence is audience, then content, then a validated friction, then a deliberately tiny app. If you have a distribution channel (or are willing to build one) before you have a product idea, start here. See also the Reels strategy prompts for the content half.