Every Interview Is Checked Before You Pay for It

Research is only worth acting on if the interviews behind it are real and substantive, so User Intuition checks quality in two places. Interview controls apply to every study, including studies with your own customers: laddering 5–7 levels deep, AI voice detection on every voice interview, automatic scoring on length, depth and coverage, human review by our team, which listens to every call with a panelist, and full review of every recording by yours. Panel controls apply on top whenever we recruit for you. Every panelist verifies their email, phone number and LinkedIn profile and passes a video identity interview, more than 100 attributes are tracked across studies, invitations go only to people who match your criteria, and an IP check runs before every interview. Interviews that fall short are never billed, and anyone flagged for fraud or low-quality answers is removed from the panel immediately.

How does User Intuition protect interview quality?

These controls apply to every study, whether participants come from our panel or from your own customer list.

Laddering. The moderator follows up on each answer, 5–7 levels deep: what someone chose, why they chose it, and the need behind that reason. This depth is also difficult to fake. A respondent reading from a script or an AI tool struggles to keep their reasons consistent that many questions down. How laddering works.

Video interviews. When a study runs on video, you can watch each participant answer in real time and judge their engagement for yourself: whether they are paying attention, reacting naturally and talking to the moderator rather than reading. Video also confirms that a real, identifiable person is on screen, and we check it for signs of outside help, such as eyes repeatedly moving to another screen. Answers given on camera are much harder to fake.

AI voice detection. Every voice interview is screened for AI-generated answers, using three measures. Answer similarity scoring compares how answers are worded and structured across the interview. Latency analysis measures how long each answer takes to begin. Linguistic analysis checks whether answers stay thematically and emotionally consistent, and looks for the markers that separate genuine human answers from generated text. A real person's wording and timing change from question to question; when someone relays answers from an AI tool, both stay unusually uniform. We also flag the signs of a coached respondent: long pauses after each question and the flat delivery of text being read aloud.

Real-time engagement monitoring. Every interview is monitored as it happens: how answer length develops, how engagement varies across topics, and whether follow-up questions produce real elaboration or deflection. The moderator also compares what a participant claimed in the screener with what they say in the conversation, and flags contradictions for review before findings are delivered.

Quality scoring. Every interview is scored on three measures: its length, the depth of the answers, and how fully it covers your research objectives. An interview that falls short is marked Poor or Fair and is never billed. That happens automatically, so you never have to ask for a refund. When we recruit, flagged interviews are replaced from an over-recruit buffer, so your sample size and timeline hold.

Human review by our team. A member of our team listens to every call with a panelist from our panel. For studies with your own participants, we listen to a random sample of calls. This review tests the automated screening against human judgment, and we use what it finds to keep recalibrating our quality scoring, so it gets more accurate over time.

Full review by your team. Every interview is recorded and transcribed, so your team can review all of them and reject any that don't meet your standard. Rejected interviews are not billed.

How does User Intuition protect panel quality?

When we recruit participants for you, they pass these controls in addition to the interview checks above.

Verification at signup. Every panelist verifies their email address, phone number and LinkedIn profile when they join, and device fingerprinting and digital identity checks stop the same person from registering more than one account.

Video identity interview. Before anyone is accepted onto the panel, they complete a video interview that our team reviews to confirm they are a real person and who they say they are. In every study after that, accent, language fluency and, on video, visible identity are checked against the profile each panelist claims.

Tracking across studies. We track more than 100 attributes for each panelist over time and check their answers against them in every study, so inconsistencies surface. Someone can't be a carpenter in one study and an engineer in the next. Anyone whose answers show studies across incompatible professional sectors is flagged.

Invitation only. Participants have no list of studies to browse. We use those tracked attributes to invite only panelists who match your exact criteria, so people who don't qualify never learn a study exists and can't misrepresent themselves to get in. This is how we filter out professional respondents: there is no open study board to game. Niche audiences, such as executives, clinicians and narrow B2B roles, are hand-recruited separately for each study.

Recruit-time screening. Before every interview, we check the participant's IP address, and the interview won't load unless it passes. We block VPNs, data-center addresses and addresses previously flagged for fraud, and the location must match your study: if you recruit in New York City, only New York City IP addresses get through. We also block bots with challenge questions and behavioral analysis and suppress duplicates by matching email, IP address and device fingerprint. Professional respondents who hop between survey panels are kept out by invitation-only recruitment.

Removal. Anyone flagged for fraud or low-quality answers is removed from the panel immediately and is not invited again.

Quality and fraud prevention FAQ