Business idea generator: 50+ AI prompts that actually produce real ideas
Most "business idea generators" spit out 1,000 generic ideas you'll never build. These prompts do the opposite: they help you find one idea that's specific, validated, and matched to who you are. Copy any prompt into ChatGPT, Claude, or Gemini.
Why this is different from every other "idea generator"
- • No random lists. Every prompt forces specificity — a real user, a real pain, a real wedge.
- • Built for outsiders. Domain expertise beats AI cleverness. These prompts mine yours.
- • Pre-validated. Each prompt asks for distribution and monetization, not just "what."
- • Free, forever. No login, no upsell, no "premium prompt pack."
Solve a pain you already have
The strongest founders solve pains they personally feel. These prompts mine your own life for ideas.
The 'this week' prompt
List 10 things that annoyed me at work or in my personal life in the last 7 days. For each, suggest a software product that would have made it go away, who would pay for it, and what they pay for today instead.
The 'spreadsheet workflow' prompt
Ask me 5 questions about my current job or industry. Based on my answers, identify 3 spreadsheet- or email-based workflows that could be replaced by a focused SaaS product, and for each one give me: target user, MVP scope, pricing, and the first 50 customers I could reach.
The 'old industry, new tools' prompt
I have domain expertise in [INDUSTRY]. Give me 10 ideas for AI-powered tools that would only make sense to someone inside this industry. For each, explain the wedge a domain expert has over a generic AI startup.
Steal from existing markets
Take a proven business model and apply it to a niche the incumbents ignore.
The 'Notion for X' prompt
Pick 5 successful horizontal SaaS products (Notion, Stripe, Calendly, Loom, Figma). For each, name 3 specific verticals where a focused, opinionated version would beat the generic tool. For the best 3, sketch the MVP.
The 'unbundle Reddit' prompt
Browse 10 niche subreddits (suggest some I should look at based on my background). Identify the 3 most-upvoted complaints from the last 6 months. For each, propose a paid product that would solve it.
The 'micro-SaaS gap' prompt
List 10 categories where existing software costs $100+/mo but the buyer only uses 10% of the features. For each, propose a $19/mo micro-SaaS that does only the 10% — and the audience that would happily switch.
Ride a trend (without becoming a wrapper)
Trends create new pains. The trick is to solve a specific pain, not just package the trend.
The 'new behavior' prompt
What behaviors became normal in the last 24 months that didn't exist before (remote-first, AI coding agents, creator monetization, etc.)? For each, name 3 painful side effects and a software product that fixes them.
The 'regulation tailwind' prompt
List 5 new or upcoming regulations in any country (privacy, AI, climate, financial). For each, suggest a SaaS product that helps small businesses comply, and estimate the willingness to pay.
The 'AI commodity' prompt
Now that LLM calls cost almost nothing, list 10 things that used to require a human professional that could be productized as a $30/mo tool. Avoid generic chatbots. Be specific about the user, the trigger, and the output.
Validate before you build
These prompts pressure-test an idea you already have.
The 'kill my idea' prompt
Here is my idea: [PASTE]. Argue, in detail, why this is a bad idea — go after the audience, the willingness to pay, the distribution, the moat, and the timing. Be brutal. Then suggest the 3 sharpest pivots that would fix the biggest weaknesses.
The 'pre-mortem' prompt
Imagine it's 12 months from today and my SaaS [PASTE IDEA] has failed. Write the post-mortem. What were the 5 most likely reasons? For each, what should I test in the next 2 weeks to find out if it's true?
The 'competitor recon' prompt
Find the top 5 competitors for [IDEA]. For each: pricing, target customer, top user complaint from G2/Trustpilot/Reddit, and the gap I could exploit. End with a positioning statement for my product in one sentence.
How to use these prompts (the right way)
- 1. Pick the prompt set that matches where you are — solving your own pain, attacking a known market, riding a trend, or pressure-testing an existing idea.
- 2. Run 2–3 prompts. Don't run all 12. The point is one strong idea, not a list.
- 3. Push back on the AI. "Why is that audience specific enough?" "Who would actually pay $30/mo for that?" Force it to defend its answers.
- 4. Take the best 1–2 ideas and run them through the ShiporDrop quiz. AI is good at brainstorming, terrible at honest scoring.
- 5. Pick the highest-scoring idea. Do 5 customer interviews. Then consider building.
Got an idea? Find out if it's worth building.
16 questions · 8 dimensions · <4 minutes · Free
Score my idea →