Prompt Patterns for Product Discovery

This is an intermediate course designed for product managers who already use chat-based AI tools and know discovery frameworks like Jobs-to-be-Done. You should be comfortable having a multi-turn conversation with Claude, ChatGPT, or Gemini before enrolling. The course does not teach foundational prompting concepts or discovery frameworks from scratch—it builds on both to create discovery-specific prompt patterns you can use immediately.

You will leave with a tested, revisable personal prompt library covering four recurring discovery tasks: JTBD extraction from interview transcripts, comparative feature hypothesis generation, AI persona evaluation using a bias-aware rubric, and ticket-to-opportunity prioritisation using chained prompts. Every template is built and tested within the course, so the library is ready to use from day one.

No coding or API access is required. Every technique in this course is designed for the consumer chat interfaces—Claude.ai, ChatGPT, and Google Gemini—using plain text prompts you can paste directly into the chat window. The course explicitly avoids API-level features like schema enforcement or prompt objects, because those require developer access this course does not assume.

The course contains 11 lessons across 5 modules with a total narration time of 105 minutes. All lessons are lectures; there are no quizzes. Plan for additional time to apply each prompt template to your own discovery materials as you go—the hands-on application is where the learning consolidates.

No—and deliberately so. The course covers how to evaluate AI-generated personas critically, including the evidence showing that lightweight AI personas exhibit sycophancy and perform less accurately for underrepresented user groups. The course teaches you to use AI to accelerate real discovery work, not to substitute AI output for real-user research. Learners who want to replace user interviews with synthetic users will not find support for that approach here.