Customer interview guide generator

By the Gather team · Published August 18, 2026 · Last updated August 18, 2026 · Free, no signup, nothing stored

A good interview guide is mostly probes. This generates a structured guide with warm-up, core questions, laddering follow-ups, and a close, plus the probe library that turns a shallow answer into a useful one.

The structure that works

Warm up, two minutes. Get them talking about themselves before you ask anything that matters. People give short answers when they are still deciding how much effort this deserves.

Context, five minutes. Their role, their situation, what changed recently. This is where triggers surface without you asking about them directly.

Core, fifteen to thirty minutes. Three to five real questions, each with laddering probes. Not fifteen questions.

Close, three minutes. What did I not ask that I should have. This question routinely produces the best material in the interview.

Probes matter more than questions

The question gets you a surface answer. The probe gets you the reason. Keep these in front of you:

• Tell me more about that.
• Why did that matter? Then ask why again.
• What would have had to be true for you to do the opposite?
• Walk me through what actually happened.
• You said X earlier and Y just now. Help me square those.
• What did you expect, and what happened instead?

That last pair, contradiction chasing, is where the most useful material usually is. Most moderators skip it because it feels confrontational. It is not, if you ask it plainly.

Things that ruin an interview

• Leading questions. Asking whether speed is important gets you yes. Asking what mattered most gets you the truth.
• Talking too much. If you are speaking more than a quarter of the time, you are interviewing yourself.
• Filling silence. Wait. The three seconds after they finish is often when the real answer arrives.
• Asking about the future. People are poor forecasters of their own behaviour. Ask what they did, not what they would do.

Doing this at scale

Fifteen interviews at 30 minutes is a week of somebody's life, plus scheduling and analysis. That is why most teams run far fewer than they should. AI moderation runs the same guide with the same probes at a scale a human cannot, and the follow-ups are written from what the respondent actually said. More on how that works and on the platform.

Frequently asked questions

How do I write a customer interview guide?

Start with three to five core questions tied to a decision, then write laddering probes under each. Add a short warm-up, a context section, and a close. Probes matter more than questions.

How many questions should a customer interview have?

Three to five core questions for a 30 minute interview. More than that and you get shallow answers to all of them because you are managing time instead of listening.

What is laddering in an interview?

Repeatedly asking why to move from a surface answer, usually a feature, to the underlying driver, usually what the person is protecting or trying to achieve.

How long should a customer interview be?

Thirty minutes is the practical sweet spot for most B2B research. Longer interviews get better depth but much worse recruitment rates, especially for senior audiences.

Should I ask what customers would do in future?

Avoid it. People are poor forecasters of their own behaviour. Ask what they actually did last time and walk through it step by step.

Is this interview guide generator free?

Yes. No signup, nothing stored, downloadable as markdown.