Sense · Core skill lab
Replace feature opinions with recent behaviour, repeated pressure, and a decision you can defend.
Your work saves in this browser.
Field tools
Copy these into your interview, agent, review, or working document. They are specific to this repetition.
Use the same spine in every interview; follow the participant when their real workflow becomes specific.
1. Tell me about the last time [behaviour] happened. 2. What triggered it? 3. What did you do first? What happened next? 4. Show me the tool, message, or document if you can. 5. Where did it slow down or break? 6. What did you do to recover? 7. What was the consequence? 8. When did this happen before that? 9. What have you already tried to change?
Label each note before synthesising so inference does not masquerade as evidence.
[B] observed behaviour [Q] verbatim quote [F] fact supplied by participant [I] your interpretation [A] assumption still untested Only B, Q, and corroborated F can support the decision.
Calibrate judgment
Three agency account leads were interviewed about client handoffs.
Narrow the bet to distributed agencies without a mandatory live kickoff. All three rebuilt missing delivery context 2–5 times per week; two showed the duplicate briefs. One co-located agency removed the problem with a meeting, which tells us the constraint is asynchronous context—not handoffs in general.
Why it works: It identifies repeated behaviour, includes a contradiction, narrows the segment, and changes the build decision.
Users loved the idea and said an AI handoff assistant would save time. We should build an MVP and ask for feedback.
Why it fails: Interest is not behaviour, no current workaround is visible, and there is no disconfirming evidence or decision threshold.
Review → revise → repeat
Check only standards your current artifact actually meets. Then record one consequential revision before exporting it.
