Live video quality checks nudge participants on light and framing

A video interview can go wrong before the first question. The participant joins from a dark room, sits half out of frame, or answers from behind the wheel, and you find out when you watch the recording. Now Conveo checks the camera feed when the interview starts and asks the participant to fix it while there is still time.

A silent walkthrough, from a dark room to a checked interview.

How it works

• A check at the start. When a video interview begins, the AI research assistant looks at the participant's camera feed: is there enough light, is their face centered in the frame, and does the setting suit an interview (someone driving does not).

• A nudge in the conversation. If something is off, the AI research assistant asks the participant to adjust, in the flow of the interview: "Before we begin, could you move to a brighter spot?" or "Could you turn on the lights?"

• Two nudges, then on with the interview. The check asks at most twice. After that the interview continues either way. By default it never cancels an interview or screens anyone out.

• See how the check went. Open any interview to inspect its check, including the individual frames that were analyzed.

On for every organization, off in one setting

Video quality checks are in beta and on by default for every organization, so there is nothing to set up. They can be turned off at organization level.

Better light pays off after fieldwork

Conveo's video analysis reads facial expressions and behavior, and it can only read what the camera shows. A lit, centered face gives it more to work with, and gives you clips that belong in a quote reel or a stakeholder deck.

Research teams asked for it in five situations

These come from teams in consumer goods, flavors and fragrance, brewing, retail, insurance and travel. We kept them anonymous.

• Set the scene up front. Teams wanted participants seated somewhere quiet, camera on at eye level, before the first question.

• Uneven footage across a study. Teams saw some recordings come through darker or grainier than others.

• Checks during the interview. Teams asked what protects quality inside the interview, and whether a participant with a problem gets removed. Here they get a nudge and the interview goes on.

• Faces clear enough to read. Teams using expression and emotion analysis asked for a clean view of the face.

• Clips for presentations. Teams sharing reels across their organization wanted footage that holds up on a big screen.

What this means for your research

More interviews start with a face you can see, without you briefing every participant yourself. You still decide which interviews and clips make it into your findings.

The check is a nudge. Some participants can't change their setup, and after two nudges the interview continues, so a few recordings will still come through dim. The hero loop and the short walkthrough show Maya, a person generated with AI to illustrate the check. The participants in the walkthrough's opening and closing scenes are real, from a Conveo-run study. The narrated video uses a demo-account participant, and its dark "before" state was made by darkening that clip.

Video quality checks in about two minutes

A narrated walkthrough of the check, the nudge, the two-nudge limit and where you see the result.

Read the participant experience guide for what participants see, or the emotional analysis guide for what the video analysis reads.