In-Depth Interviews: What They Are, When to Use Them and How to Run Them at Scale

In-depth interviews (IDIs) are one-on-one qualitative interviews in which a moderator asks a participant open questions and follows up on their answers to understand their experiences, motivations and language in detail. They usually run from a quarter of an hour to an hour and follow a discussion guide, but the follow-up questions change with every participant.

IDIs are how you learn why customers churn or what a message makes people picture, and run well, they surface things nobody knew to ask about.

This guide covers when to choose in-depth interviews over focus groups or surveys, how many to run, and what changes when an AI interviewer runs the sessions.

What makes an in-depth interview different

The defining feature is the follow-up. A survey asks everyone the same fixed questions. An in-depth interview asks an open question, listens, and decides what to ask next: a probe on a vague answer, a request for an example, a question about something unexpected.

That is why IDIs work for professional and expert audiences as well as consumers. Our study of 76 market researchers used in-depth interviews with US researchers, most with eight or more years of experience. Structured questions showed that 91% are open to AI in their workflow while only 29% trust AI-generated insights as much as traditional research. The open conversation explained the gap. Researchers would hand AI transcription, theming and deck formatting, but deciding what the findings mean stayed with them. A survey would have produced the two percentages without the line they draw between them.

In-depth interviews vs focus groups vs surveys

A quick way to decide:

• Choose in-depth interviews when you need individual reasoning, the topic is personal or sensitive, participants are hard to get in one room, or you want to avoid people influencing each other.

• Choose focus groups when the group dynamic is the point, for example watching how people build on each other's ideas or react to a concept together.

• Choose a survey when you already know the answer options and need to size them.

Focus groups have a known weakness: one confident voice can steer the room, and quieter participants agree rather than disagree. Individual qualitative interviews remove that pressure, though they still carry the moderator's influence. For more on when to measure and when to explain, see our guide to qualitative vs quantitative research.

Sensitive topics and the moderator effect

The person asking the questions shapes the answers. A participant who wants to come across well may soften an opinion about money, health or politics. So decide carefully who, or what, moderates.

Our AI interviewer vs human moderator study asked 50 US adults who had recently taken part in a human-moderated session to compare it with an AI-moderated one. 92% were very or extremely comfortable sharing honest thoughts with the AI interviewer, the same figure as for their human session. When honesty differed, it shifted toward AI: 34% said they were more honest with AI and 6% with a human. The main reason, cited by 60%, was that an AI interviewer feels less judgmental and safer to be candid with.

Topic mattered. 58% would choose an AI interviewer for political or religious topics. Health was the human moderator's strongest topic at 32%, though AI still edged ahead at 40%, and participants wanted someone who could recognize distress and refer them to help. The trade-off was warmth: 26% found AI more impersonal and 22% felt it lacked empathy or nuance. Our article on types of bias in research covers this moderator effect alongside other common biases.

How many in-depth interviews do you need?

The traditional answer is "until you stop hearing new things." For a single, fairly uniform audience, new themes often slow down after a dozen or two interviews. That rule of thumb was shaped as much by moderator time and budget as by method. A human moderator can only run a few sessions a day.

Larger samples change what you can do with the data. With enough interviews you can count how often each theme appears, compare segments and report percentages alongside quotes. In our apparel brand positioning study, 151 US clothing shoppers each did an interview. At that size the study could show that 94% rated comfort and fit as extremely or very important, but only 38% mentioned comfort unprompted when describing how they choose a brand. That gap, between what people say matters when asked and what they raise on their own, needs both a structured question and an open conversation from the same person, across enough people to count.

Practical guidance on sample size:

• Exploratory work with one audience: a smaller sample often maps the main themes.

• Comparing segments: plan enough interviews per segment to see whether themes differ.

• Quantifying themes: go larger, with structured questions in the same interview.

How to run in-depth interviews well

1. Start from the decision. Write down what the team will do differently depending on the answer, then write the guide around it.

2. Open broad, then narrow. Ask about behavior and context before showing any stimulus or asking for ratings.

3. Write probes, not just questions. Plan the follow-ups you want when an answer is vague: "Can you give me an example?" or "What happened next?"

4. Ask why after every rating. A score without the reason is hard to act on.

5. Keep it short enough to hold attention. In our restaurant message testing study, each of 96 diners did a 15 to 20 minute AI-moderated video interview and rated and explained 12 messages.

6. Plan the readout early. In the market researchers study, 58% of researchers who described stakeholder pushback called it a communication or framing problem.

Our qualitative interview guide goes into writing the discussion guide itself.

Running in-depth interviews at scale with AI moderation

An AI interviewer runs the same kind of session a human moderator would: it asks an open question, listens, and follows up on what that participant said. Because there is no moderator calendar, hundreds of interviews can run at once instead of eight a day, and participants join by video or voice on their own schedule.

That makes customer interviews at survey scale practical. On Listen, you can run in-depth interviews with real people in 100+ languages, with structured questions and AI-moderated follow-ups in the same session, and every theme links back to the participant's clip.

Researchers still own the study. In the market researchers study, 68% were uncomfortable with product managers running their own research. The design of the guide and the interpretation of the findings stay with the people trained to do them.

FAQ

What are in-depth interviews in research?

In-depth interviews are one-on-one qualitative interviews where a moderator asks open questions and follows up on each answer. They are used to understand individual experiences, motivations and language in more detail than a survey or focus group allows.

How long should an in-depth interview be?

Most run from a quarter of an hour to an hour, depending on the number of topics and stimuli. Shorter, focused sessions keep attention high, and a well-written guide with planned probes gets more depth per minute than a long, loose one.

What is the difference between in-depth interviews and focus groups?

An in-depth interview is one participant and one moderator, so each person's reasoning comes through without group influence. A focus group brings several participants together, which is useful for watching group reactions but lets dominant voices shape the discussion.

How many customer interviews are enough?

For exploratory work with one audience, a modest sample often covers the main themes. If you want to compare segments or report how common each theme is, plan a larger sample, which AI moderation makes affordable.

Can an AI interviewer run qualitative interviews?

Yes. An AI interviewer asks open questions and follows up on each participant's answers. In one study, 92% of participants were very or extremely comfortable sharing honest thoughts with an AI interviewer, the same as with a human moderator.