What Do Participants Really Think of AI-Moderated Interviews? Turns Out, They Prefer Them!

Participants feel less judged and more candid in AI-moderated interviews. Here is what Strella's sessions show, and where humans still lead.

Melissa Tovin

Insights teams weighing AI moderation often worry that participants will hold back with a machine and the answers will come back thin. The evidence points the other way: participants open up more readily to AI-moderated interviews on sensitive topics because nobody is forming an opinion of them while they talk. This piece draws on Strella’s published feedback from 50 participants, and it closes with the work where human moderation is still the right choice.

What the participant evidence shows

The evidence below covers how willing participants were to open up, how engaged they were, and what they said they liked about the format.

Willingness to open up

Participants open up more readily to an AI moderator on sensitive topics, because nobody is forming an opinion of them while they talk. Strella’s published feedback from 50 participants comes from 50 US working professionals with five or more years of experience, gathered in June 2026. Four themes recurred:

  • Participants felt comfortable rather than awkward.
  • The interview moved at a natural pace.
  • They felt more able to be honest.
  • They reported less pre-interview anxiety and less self-censorship on sensitive topics.

From the participant’s side, there is no scheduling pressure, and they talk at their own pace and in their own words, even while explaining a decision they are not proud of. In their own words:

  • “With AI there’s no judgment and there’s just recognition. With a human, you kind of get into that” (Pebbles M, 47, Physician Practice Manager)
  • “There’s almost more pressure when it’s a person.” (Bryant K, 36, Business Operations Manager)
  • “I would say it’s actually less anxiety-driven to have an AI moderator as opposed to a human” (Bryant K, 36, Business Operations Manager)
  • “Sometimes it’s easier than actually having a moderator, just because I sometimes get a little nervous with a moderator” (Tracy B, 62, HR Generalist)

A short answer tells a team which box a participant ticked. A participant explaining the circumstances behind a decision tells insights, research, product and strategy teams which message to build and which to retire.

Engagement and completion

Strella research on 800 AI-moderated sessions shows how participants experienced the format; the satisfaction, recommendation and repeat figures below come from Strella’s own sessions and participant feedback, not an independent audit.

  • Engagement: seven in ten Strella sessions ran longer than expected, which gives the moderator more room to follow up and hear the reasoning behind an answer.
  • Satisfaction: participants rated the experience 4.7 out of 5, which suggests they found the format easy to take part in.
  • Recommendation: asked how likely they were to recommend it, participants averaged 9.3 out of 10, with no one scoring as a detractor, so none of these participants left the experience unhappy enough to speak against it.
  • Willingness to repeat: 96% chose “sign me up again” when offered another AI-moderated interview, which makes it practical to re-run the same guide with new groups.

What participants say they like

Participants reported liking two things about the format:

  • Interviews on their own time: Participants open the link when it suits them, with no calendar coordination and no moderator waiting, and take the conversation at their own speed.
  • A conversation built around them: The moderator follows up on what the participant said a moment ago, so people feel heard. As participants put it:    
    • “I love human interaction, but for this type of interview, I would rather talk to an AI moderator.” (Pebbles M, 47, Physician Practice Manager)
    • “I was under the assumption that the AI moderator would sound more robotic. But it didn’t” (Steven C, 42, IT Security Professional)

Want to see what an AI-moderated interview is really like? Try an interview:

Where participants still prefer a human

Human moderation is the right choice for studies that turn on trauma, vulnerability, crisis or deep cultural meaning. On Strella, human-moderated interviews are live on the same platform as AI-moderated interviews, so a team can run a human-moderated study alongside an AI-moderated study and analyze both together in the same workspace. Practitioner research supports drawing the line there.

A 2026 practitioner study of AI-moderated interviews with Afro-descendant and Latine respondents examined cultural considerations in AI-moderated qualitative interviews. It found that human-moderated interviews drew more detailed and personal responses, which makes human moderation the right choice where cultural nuance and emotional weight are high.

The same boundary sits inside any study. Explaining a decision is the kind of topic a study assumes participants will discuss openly with an AI moderator. If the brief turns to a participant describing financial hardship, a trained moderator who can slow down and notice distress belongs in the room.

  • If the value of a study sits in how a person tells a story that carries emotional or cultural weight, a human should moderate it.
  • If the value sits in hearing many people explain a decision, the AI moderator can carry it, and on sensitive decisions participants report more candor with it.

Addressing the researcher objection

A trained human moderator still reads emotion and improvises better than an AI moderator, and the rapport objection is right about that half and wrong about the other. The wrong half is the assumption that warmth is what produces candor on sensitive topics.

Lucas et al. 2014, published in Computers in Human Behavior more than a decade ago and still holding, told participants that either a computer or a human was running their interview. Those who believed a computer was running it feared disclosure less and worried less about how they came across, whatever was actually running the interview. They were rated as more willing to disclose when they believed no one was judging them.

The probing half deserves a straight answer too.

  • Strella’s AI moderator runs a live voice-to-voice conversation and chooses each follow-up question from what the participant just said; it does not walk a fixed branching tree or fire standard probes. When a participant gives an answer the guide did not anticipate, the moderator can follow up on it.
  • Researchers set the objectives, edit the generated discussion guide, organize it into sections, add conditional branching and embed stimuli, and the guide builder flags leading questions before launch (a flag, not a guarantee of bias-free research).

The moderator can still let go of a thread that a seasoned moderator would have chased. That is a reason to keep a person moderating studies where the most important threads are still unknown, and to keep a researcher reviewing the source interviews behind any finding.

To hear how the moderator follows up on your own answers, including unexpected ones, try an interview.

What this means for data quality

More candid, engaged participants give the team richer verbatims, and each one links back to the participant’s own words and video. Insights, research, product and strategy teams receive explanations of why customers made a decision, in the customers’ own sentences. Those explanations can be grouped into themes by research objective, and highlight reels can be cut from the recordings. Traceability lets anyone on the team open the interview behind a finding and check it.

A more candid answer is still not an audited one, which is why the quality controls sit underneath the synthesis. Strella layers its own checks onto whatever a panel provider already runs:

  • Screening and fact-checking: Strella screens participants and fact-checks their screener answers before the interview.
  • Post-interview fraud checks and human review: Strella checks each session on completion, response quality, consistency, location, IP and timing. A person reviews flagged sessions, adding a judgment layer on top of automated checks.

Candid answers from many people, each checked and traceable to its source, are also what give AI moderation its own place in a research practice. An AI-moderated interview is a distinct method in qualitative research in which an AI moderator holds a live, voice-to-voice conversation with a real participant and asks follow-up questions based on what the participant said, with every insight linking back to the interview it came from. It sits between survey counts and a handful of human-moderated in-depth interviews, returning the reasons from hundreds of customers, each traceable to the customer’s own words. The team chooses between the three by what the decision needs, and the middle option exists because neither of the others can do its job.

Where AI moderation fits in a research practice

AI moderation fits as the method a team reaches for when it needs qualitative depth from many people on a continuous cadence.

  • A team can re-run the same guide with each new group of subscribers and compare the reasons month over month without waiting for an annual study.
  • Hundreds of interviews can run in parallel, which keeps a monthly study practical.

Human-moderated and AI-moderated sessions live in one platform and one repository, so a natural-language query across studies pulls from both. Teams can also query any historical interviews and customer calls they have uploaded. Teams choose the moderation method study by study, and the findings end up in the same place either way. For a team without the bandwidth to keep a monthly study running in-house, Strella’s research team can scope, recruit, run and synthesize an AI-moderated study through the fully managed service level.

Surveys give counts, while AI-moderated interviews, in which the AI moderator talks with each participant, asks follow-up questions based on what they said and links every insight back to its interview, give explanations from more people than a human moderator could reach. If the decision needs counts, use a survey; if it needs explanations from many people, use AI-moderated interviews; and if the answer depends on a moderator’s judgment in the room, a human moderator leads it on the same platform.

Bring your research question to us and we will talk it through. See Strella in action.

FAQ

Do participants open up in AI moderated interviews?

Yes, by their own account. Participants in Strella’s feedback said they felt less judged speaking with the AI moderator and more open on sensitive subjects than they would be with a person.

Do participants need to be told they are talking to an AI?

Yes. The industry’s code of standards requires telling participants at the start of research that they are talking to an AI. Teams should say so up front.

How does engagement compare to surveys?

The interview captures the reasoning that a survey cannot. A survey records a ranking or a checkbox answer; the interview records the reasoning behind it, in the participant’s own words, with follow-up on anything unexpected.

How does engagement compare to human-moderated interviews?

Participants describe feeling less judged and more candid with the AI moderator than they would be with a person; no head-to-head rating comparison exists. The two methods do different jobs: AI moderation hears many people explain a decision, and a human moderator handles studies where judgment in the room matters. Both run on Strella, and teams choose per study.

What happens if the AI moderator misses a follow-up?

The researcher can see it and act on it. Every finding links to its source interview, so a reviewer can open the session, hear what was missed and tighten the guide for the next wave.