What to ask
Mine the last time it happened. Map the forces. Don't pitch, don't lead, don't hypothesize.
Your job: build a guide that mines past behavior, not hypothetical futures.
Design my interview guide Read the customer forces in `business_model://current`. Draft a three-act problem-discovery guide for recent switchers. Flag any leading, future-facing, or socially-desirable question, and give a follow-up probe for the three moments most likely to surface a workaround. claude mcp add --transport http leanspark https://leanspark.ai/mcp Connect first →
What you get back
Builds your three-act interview guide and strips out every leading or future-facing question, so you mine real behavior — and gives you the follow-up probes for the moments most likely to surface a workaround. Then you run the conversation and listen.
A three-act interview guide with the leading questions flagged and rewritten.
A good interview is a guided story, not a questionnaire. Your job is to get the person re-living a specific, recent episode — and then to get out of the way. The structure that works is three acts: set the scene (what was going on when the problem showed up), walk the struggle (what they did, step by step, including the workarounds), and surface the switch — the switching trigger that made them finally change, and what almost stopped them.
Map the four forces
Underneath every switch are four forces. Listen for all of them:
- Push — what was broken or frustrating about the old way.
- Pull — the outcome they wanted and couldn’t get.
- Friction — the effort and worry of changing: migrating, learning, “will this even work for me?”
- Inertia — the habit and comfort of the status quo: “it’s not great, but I know how it works.”
The switch equation. People only switch when push plus pull outweighs friction plus inertia — and inertia is heavy, so a merely attractive pull rarely beats it. If you can’t hear all four forces in the story, you don’t yet understand why they did — or didn’t — move.
The switching trigger. The four forces are the steady-state pressure. The switching trigger is the spark — the specific moment the old way failed badly enough to make them act (“the strap tore on the train”). It’s the single most predictive thing in the interview: no trigger, and you’re hearing a nice-to-have, not a switch.
When those USERcycle founders finally switched, you could hear all four: the push was a conversion rate that wouldn’t budge; the pull was finally seeing why it wouldn’t; the friction was ripping out analytics they’d already wired in; the inertia was the dashboard they already knew by heart. Miss one and you’re guessing.
The three ways to ruin an interview
- Leading. “Don’t you hate it when…?” hands them the answer. Ask open: “Walk me through what happened.”
- Hypothetical. “Would you use a tool that…?” invites a polite lie. Stay in the past: “What did you do the last time?”
- Pitching. The moment you describe your idea, the interview is over — they stop telling you the truth and start protecting your feelings. Don’t sell. Listen.
Look for workarounds, not “problems.” Never use the word “problem.” Listen for struggles, pet peeves, and workarounds — calling three times to confirm a booking, waking up early to absorb delays. When someone goes out of their way, that effort is the problem worth solving.
Related: the interview-prep MCP recipe runs this chapter’s exercise from any MCP-connected assistant.