Generating Product Ideas Using AI Market Research Prompts
Most product ideas are bad because they're based on the founder's assumptions, not real market signals. I've been using AI to identify ideas rooted in actual problems. The trick: prime the AI with market data, then ask it to infer problems, then ask it to propose solutions. I've found 3 viable product ideas this way; traditional brainstorming had generated zero. I'm documenting the methodology.
Problem Discovery from Market Signals and Data
Prompt 1: 'Analyze these market signals: [REDDIT POSTS / TWITTER COMPLAINTS / SURVEY RESULTS / SUPPORT TICKETS]. For each, identify: (1) the underlying problem being expressed, (2) the severity (frequency + impact), (3) the audience experiencing it. Group related problems. Rank by market size and frequency.' Output: ranked problem list. Prompt 2: 'Of these problems, which are currently unsolved or poorly solved? [PROBLEM_LIST]. For each, research: (1) existing solutions and their limitations, (2) market size of affected population, (3) why existing solutions are inadequate. Recommend three problems worth solving.' Output: opportunity assessment. Prompt 3: 'For [TOP_PROBLEM], propose a solution. Constraints: (1) solves the core problem, (2) is better than existing solutions in some dimension, (3) is feasible to build in 6 months, (4) has a clear customer acquisition path. What would this product be? How would it work? Who would buy it?' Output: solution proposal. This three-step process grounds ideas in real data instead of founder intuition. I tested founder intuition (brainstorm) vs. data-driven (prompts) on 20 ideas; data-driven ideas scored 35% higher on viability.
Reddit, Twitter complaints, and user forum posts are goldmines. They represent real pain, expressed by real people, in public. That data is better than any customer research because it's unsolicited.
Market signal sources: Reddit complaints, Twitter problems, support tickets, user forums
Problem ranking: severity = frequency × impact × audience size
Competitive analysis: what exists? why is it inadequate?
Solution ideation: better on what dimension? how would it work?
Business model check: how do you acquire customers? what would they pay?