For screening and self-paced responses
Questions branch from previous answers, fields are validated and progress is saved so respondents can return later.
Agent-CustDev
Runs an evidence-based CustDev study from hypothesis to report. Respondents answer in an adaptive form or text chat. The agent probes past behavior and keeps every material finding connected to an exact quote and its source.
InputResearch goal, segments, hypotheses and constraints
During the interviewForm ↔ chat, one question at a time, save and resume
OutputQuotes, pains, contradictions, hypothesis verdicts and a report
Control: findings are not published automatically. A reviewer approves, rejects or refines them. The current version supports forms and text chat.
01Role
It does not replace the researcher. It prevents evidence from getting lost between the interview and the decision.
The agent handles repeatable research work. The team keeps control of the goal, recruitment and final conclusions.
What the agent does
- Structures the study: segments, hypotheses and an editable script
- Runs text interviews: an adaptive form or a focused chat
- Checks questions: blocks leading, hypothetical and compound wording
- Returns to facts: what happened, when, what the person did and what it cost
- Extracts evidence: the exact quote and its source
- Prepares synthesis: pains, opportunities, contradictions and hypothesis verdicts
02Interview
A natural conversation for the respondent. A disciplined method for the study.
The system does not turn the interview into a long questionnaire. It chooses the next question from context while preserving the research boundaries.
For depth, not small talk
The agent asks one question at a time and turns vague statements into a concrete situation, action, workaround or cost.
The format changes; the context does not
Respondents can move from form to chat and back. Existing answers are preserved and are not requested twice.
ResponseThe respondent's own words
ProbeA fact about past behavior
EvidenceQuote and exact source
FindingPain, opportunity or contradiction
ReviewHuman publication decision
03Method
The Mom Test shapes the interview before a question reaches the respondent.
This reduces polite agreement, speculation about the future and answers suggested by the interviewer.
The agent asks for specifics
- “When did this last happen?”
- “How did you handle it before?”
- “What was the hardest part?”
- “How much time, money or effort did it take?”
- “What did you do next?”
The agent avoids steering
- does not pitch a solution during research
- does not ask whether someone would buy an imaginary product
- does not bundle several questions together
- does not treat praise or opinion as behavioral evidence
- does not hide uncertainty when evidence is insufficient
04Result
Every report can be traced back to the respondent's exact words.
Instead of a polished summary, the team receives connected research artifacts. If a finding looks questionable, its basis is visible.
- EvidenceA fact, verbatim quote and exact source.
- Pain and opportunityContext, strength, recurrence in the studied cohort and signs of real effort to solve it.
- Hypothesis verdictSupported, refuted or still uncertain, with supporting and conflicting evidence.
- Interview and cohort reportRecurring patterns, segment differences, exceptions, contradictions and next questions.
“Every Monday we copy numbers from three spreadsheets by hand. It usually takes about two hours, and once a month we have to reconcile everything again.”
Human review
Approve · reject · refine
AI prepares the draft. A reviewer decides whether each material finding belongs in the published report.
05Control
Data rules are agreed before the first interview.
The team defines the permitted environment, participants and review process in advance. Extra access is not treated as a default requirement.
Cloud or local is decided up front
For sensitive studies, work can be designed to stay inside the organization's environment. The right option depends on the data and operating requirements.
Only the materials the study needs
Sources, participants and permitted actions are agreed before launch. Additional data is not connected by default.
The agent prepares; the team decides
The script remains editable, and material findings can be approved, rejected or refined before they enter the final report.
Who benefits most
Teams that run interviews regularly but rely on a few people to maintain research quality.
Founders and product teams
Test the problem before building, see real workarounds and separate strong evidence from general interest.
UX and CustDev researchers
Scale response collection, preserve exact quotes and prepare a reviewable evidence base faster.
Organizations with sensitive data
Agree on the environment, access and review process before interviews are collected.
First step
Choose a ready-made solution or describe your process.
Pick one of the twelve agents and subagents, or fill in the questionnaire for custom development.