Text your ideal customers a link. An AI interviewer calls them by voice, asks the follow-ups a good researcher would, and turns every answer into evidence.
Five comparable core questions for everyone. Follow-up probes that change as the AI learns. Five to ten minutes, whenever they're ready.
A short, personal SMS says why they fit, how long it takes and what they earn: $2–8 depending on the length and quality of their answers.
The AI greets them by name, asks every core question in order, probes when an answer is vague, and never pitches.
Every word is saved against the guide version that was used, scored for quality and ready to quote.
We read competitors, pricing pages and reviews, then draft an ICP where every claim cites a source.
The voice AI talks to the people your ICP describes. This is where the truth comes from.
A scorecard marks each assumption supported, contradicted or unclear, with quotes linked to the call.
After every third interview the AI finds patterns, then rewrites its follow-ups to test them, and to try to prove them wrong.
“How do you track receipts today?”
Starting probe from the ICP's open questions.
“When did your accountant last reject a categorised receipt?”
3/3 mentioned accountant pushback.
“Would you switch if your accountant approved the tool first?”
Tests whether it's a dealbreaker, not just a pain.
Per completed interview, set by the AI from the length and quality of the answers. You see every amount before you pay.
A full sample project with interviews, patterns and an insights report is ready when you sign in.
Start a project