In this article
- Why the first answer is usually pay
- Sixteen questions, grouped by what each one is for
- When the decision actually formed
- What the job was actually like
- Who knew, and when
- What would have changed it
- The job you are going to
- Three of those do the real work
- Run it mid-notice, and keep the manager out of the room
- What the answers are actually for
- Where this stops working
This is written for the person running the exit interview, not the person sitting one. If you are about to be asked these questions rather than ask them, what follows is the other side's reasoning.
Aon's 2026 survey of more than 1,200 Indian organisations puts overall attrition at 18%. A 200-person company at that rate loses 36 people a year. At 45 minutes each that is 27 hours of exit interviews, and if yours run like most of them, a good share of those hours will produce some version of “better opportunity” and then stop.
The questions below are built to get past that answer. They are not better-worded versions of “why are you leaving”. They are questions whose honest answer costs the person nothing, which turns out to be most of the trick.
Why the first answer is usually pay
Pay is the safest thing to say. It is specific, it is nobody's personal failing, it cannot really be argued with, and it does not require someone to criticise a person whose reference they may need next month.
That last part carries more weight in India than most advice written elsewhere assumes. Somebody on a 60 or 90 day notice period is not leaving on Friday. They are in the building for another two or three months, their full and final settlement has not been processed, and the manager they might criticise is often the one who signs it off. The statutory side of an exit has its own timelines, and all of them run past the day of this conversation.
Gallup's 2026 report, using 2025 data, found that 53% of employees in India say it is currently a good time to find a job where they live. The person in front of you probably agrees. They may also have friends at the company you are about to lose them to.
So the problem is not that people lie in exit interviews. They answer the question you asked, in the safest register available, and pay is almost always available.
Most question sets make this worse by asking for judgements. “How would you describe the company culture” invites a verdict, and a verdict is exactly what somebody protecting a reference will not give you. Ask about events instead. An event is a fact about a Tuesday rather than an opinion about a company, and people hand those over without much hesitation.
Sixteen questions, grouped by what each one is for
Use ten or twelve of these, not all sixteen. The grouping matters more than the wording, so change the phrasing to sound like you.
When the decision actually formed
- When did you first think about leaving? Not when you started applying. When the thought first turned up.
- What was going on that week?
- Was there a particular day you can point at?
- If that week had gone differently, would you still be here?
What the job was actually like
- Describe a normal Tuesday from your last month here.
- What did you spend time on that you did not expect to?
- What was in the job description that never really materialised?
- When did you last get four uninterrupted hours?
Who knew, and when
- Who here knew you were unhappy before you resigned?
- When did your manager last ask you something like what I am asking now?
- If you had wanted to raise this three months ago, where would you have gone?
What would have changed it
- What is the one thing that, had it been different, would most likely have kept you?
- Did you ever ask for it?
- What happened when you asked?
The job you are going to
- What does the new role give you that this one did not?
- What are you giving up to take it?
Three of those do the real work
Most published question sets are variations on the same six themes, and an AI summary of them will give you those six in about a second. These three are the ones that tend to be missing.
“Who here knew you were unhappy before you resigned?” asks whether the information existed inside the company before the resignation did. If the answer is a name, you did not have an information problem. You had a routing problem, and that is fixable. If the answer is nobody, that is a different and more expensive finding.
“Did you ever ask for it, and what happened?” separates two failures that look identical in a spreadsheet. A company that did not know is in better shape than a company that was told and did nothing. Both show up as one line in the attrition report.
“What are you giving up to take it?” is the only question in the set that asks what was working. People answer it honestly because it is not a complaint, and the answer tells you what to protect for everyone still here.
Run it mid-notice, and keep the manager out of the room
The last day is the worst possible slot. Their laptop is going back, the farewell lunch has happened, and the conversation is competing with the logistics of leaving. You will get warmth and very little information.
Somewhere in the middle of the notice period works better. The decision is settled enough that they are not defending it, and far enough from the final handover that the conversation is not an interruption.
Whoever runs it should not be the person they might need to criticise. On a small team that may mean a founder rather than HR, or HR rather than the founder, depending on who the manager is. The test is simple. Can this person say “my manager” in this room without calculating what it costs.
Say out loud what happens to the answers. Who reads them, whether anything is attributed, and what reaches the person's manager. If you cannot answer that, work it out before the interview rather than during it.
What the answers are actually for
Not for keeping this person. That decision was made weeks ago, probably before you knew there was anything to decide.
They are for two things. The first is classification: voluntary or not, regretted or not, and which of a small set of drivers it belongs to. Pick five or six drivers, use the same ones every time, and resist adding a seventh. The second is the pattern that only appears at twenty interviews rather than one. Two exits in a quarter from the same team is not a coincidence, and burnout is the leading indicator of attrition that usually sits underneath it.
That pattern is the part worth your time. One exit interview is an anecdote about a person. Twenty, classified the same way, is a measurement, and measurement is where employee retention work has to start. The questions above are designed to make that classification honest rather than convenient.
Where this stops working
A good exit interview cannot save the person sitting in it. Nothing here will change that, and Wanderfly cannot change it either. By the time this conversation happens, the only window in which anything could have shifted has closed. We build a product for the months before this, and it has nothing useful to offer a reader on the day somebody hands in their notice.
The honest limit of the exercise is this. You will get better data than you are getting now, you will get it one person at a time, and you will get it after each decision is already final. That is worth doing. It is not a retention programme, and treating it as one is how companies end up with four years of well-organised exit data and the same attrition rate they started with.
If your exit interviews keep returning the same driver and nothing changes, the problem is not your question set. Better questions produce better answers. They do not produce the willingness to act on them.
Questions people ask
What questions should you ask in an exit interview?
Ask about events rather than opinions. When the thought of leaving first turned up and what was happening that week, what a normal Tuesday looked like in their last month, who inside the company knew before the resignation, whether they ever asked for the thing that would have kept them and what happened when they did. Questions that invite a verdict on the culture tend to get a safe answer, because the person is usually still serving notice and may need a reference.
Should HR or the manager run the exit interview?
Neither, as a rule. It should be run by somebody the person is not being asked to criticise. If the manager is the subject, HR should run it. If HR owns the process the person is unhappy with, a founder or a department head is better. On a team of thirty this is a practical constraint rather than a policy, and the test is whether the person can say “my manager” in that room without working out what it will cost them.
When in the notice period should the exit interview happen?
Mid-notice rather than on the last day. In India that usually means somewhere in the second month of a 60 or 90 day notice period. By then the decision is settled enough that they are not defending it, and the conversation is not competing with handover logistics and a farewell lunch. The last day produces warmth and very little information.
Do employees tell the truth in exit interviews?
Mostly yes, to the question you actually asked. The common failure is not dishonesty, it is that people answer in the safest available register, and pay is almost always available as an answer. It is specific, it blames nobody, and it cannot be argued with. Questions about events rather than judgements get past this, because an event is a fact about a particular week rather than a verdict on a company.
What should you do with exit interview data?
Classify it and then look at the pattern, not the individual. Use five or six fixed drivers, record whether the exit was regretted and whether it was voluntary, and keep the categories stable so that twenty interviews are comparable. One exit interview is an anecdote about a person. Twenty, classified the same way, is a measurement. Do not use it to try to retain the person in the room, because that decision was made weeks earlier.
Is an exit interview mandatory in India?
No. There is no statutory requirement to conduct an exit interview in India, and an employee is under no obligation to attend one or to answer anything in it. The obligations that do attach to an exit are separate, and concern notice periods, full and final settlement and the relevant statutory filings. Treating an exit interview as a condition of releasing any of those is a bad idea for reasons that go well beyond the quality of the answers.
