How to test a vibe-coded prototype with real users
A session kit for testing an AI-built prototype: task scenarios, a script, an observation sheet, interview questions and a findings table.
Read articleHow to plan a study, ask questions that get real answers, and turn what customers tell you into decisions you can stand behind.
A session kit for testing an AI-built prototype: task scenarios, a script, an observation sheet, interview questions and a findings table.
Read articleA worksheet and a small pilot plan to judge whether a practice common among your best branches or reps is a cause of their success.
Read articleChurn interview questions as a full 30-minute script, plus a worksheet for weighing rival explanations, so you know which cause of leaving to fix first.
Read articleTurn your feature list into an assumption map, run interviews about past behaviour and workarounds, and use a rubric to pick what to build.
Read articleUse an evidence ladder, a past-behaviour interview guide and a stop worksheet to judge whether a problem is real enough for the next test.
Read articleA first-week timeline worksheet, a copyable interview guide and a matching table that show which step between sign-up and value to fix first.
Read articleTurn your interview transcripts into a finding you can defend: a worked example with a coding table, a counterexample and a copyable template.
Read articleWrite eligibility rules from your decision, then use a copyable screener, invitation messages and channel table to recruit the right interviewees.
Read articleA 45-minute session plan with planted flaws, a coding sheet and interview questions to test whether people rely on your AI feature's output.
Read articleA seven-line user-readiness checklist to complete before you invite a first group to your AI-built app, with go or fix-first rules.
Read articleStrong operators do opposite things. A comparison table and three-way sort show what to trial as a shared principle, what fits some sites and what to test.
Read articleLearn to tell AI interviews, AI analysis of real material and synthetic users apart, with an evidence-origin table, a checklist and a slide label.
Read article