Field notes on product discovery - finding, framing, and building on real customer problems.
PMs who write specs inside an AI agent keep hitting the same ceiling: the chat connects to nothing, remembers nothing, and can't say who actually said what. The fix isn't better prompting - it's grounding the agent in a research corpus it can cite.
Online research turns up two things that look identical on the page - a wishlist and a real felt pain. The discipline that tells them apart has to run at the moment you capture the quote, before an accumulating pattern makes you fall for it.
Discovery happens at every kickoff, then delivery pressure wins and the calls go quiet. Why episodic discovery fails the strategic decisions that arrive between projects - and how to make every user contact compound instead of evaporate.
Sales talks to your customers more than you do, and what comes back is feature requests with the evidence stripped out. How PMs can keep the customer's verbatim words attached as requests travel to the roadmap.
Deadlines don't wait for synthesis. Why the rigor-versus-speed dilemma is built on a false assumption, and how PMs under time pressure can turn a backlog of interviews into evidence-linked problems.
A field guide to Sales Safari: the worldview behind observing communities for raw customer language, and a step-by-step loop for capturing quotes with the CLRA extension and synthesizing them into problems worth building on.
