What do you still not know about this role that would matter to your decision?

A late-stage question with a different purpose

By the time somebody asks this, usually in a final round or just before an offer, the assessment of your skills is largely finished. The question is doing two things. It is a genuine offer of information, because a hiring manager who has decided they want you would rather resolve your doubts now than lose you in week six. It is also the last behavioural signal of the process: whether you are evaluating this job or merely trying to be selected for it.

Candidates lose ground here by being agreeable. "You have covered everything, it sounds great" is intended as a compliment and is received as an absence of judgement. A person about to reorganise their working life around a decision has open questions, and a senior person has specific ones. Saying you have none suggests either that you are not really deciding or that you have not thought about what this role would be like on a Tuesday in November.

Distinguish what you still do not know from what you can look up

The material for this answer is not a list of clever questions. It is an honest audit of your remaining uncertainty, and it will be different after five conversations than it was at the screen. Work out which unknowns would actually change your answer, and separate them from the ones that are merely interesting.

Things you can find out yourself — what the product does, who the competitors are, what the last funding round was — should already be done, and asking them wastes your one good slot. Things nobody can answer honestly in an interview, such as whether you will enjoy the culture, are better approached obliquely by asking about mechanisms rather than atmosphere. What is left is the useful set: how work arrives and gets decided, what the team's last quarter actually looked like, what happened to the person who held this role before, and what would make the manager consider the hire a mistake.

A question that gets a brochure answer	A question that gets you information
"What is the culture like here?"	"When two teams want the same thing and neither will move, who decides, and how long does that usually take?"
"Is there a good work-life balance?"	"How often was the on-call rota paged out of hours last month, and who covered the follow-up work?"
"What are the growth opportunities?"	"Who was promoted out of this team in the last year, and what did they do to get there?"
"Why is the role open?"	"Is this a backfill? If so, what did the previous person go on to do?"
"What are the biggest challenges?"	"What is the thing about this team you would fix first if you had another engineer tomorrow?"

The right-hand column works because each question asks for a fact or an event rather than a characterisation. Facts are harder to smooth over, and the way an interviewer handles one they would rather not answer is frequently more informative than the answer itself.

Ask the right person, and say why it matters

Directing every question at whoever happens to be in front of you is a small error with a large cost. Recruiters know process, dates, bands and policy. Hiring managers know priorities, the team's real state, and what success looks like in six months. Future peers know the things nobody else will say out loud: what the code is like, how planning really works, what happened during the last crunch. Someone two levels up knows where the team sits in the company's plans and how safe its funding is. Asking a peer about compensation bands, or a recruiter about architectural direction, signals that you have not registered who you are talking to.

Attach a short reason to your question as well. "I ask because the last team I joined had three product owners in a year and I want to understand how decisions get made here" turns an interrogation into a conversation and lets the interviewer answer the concern rather than the sentence. It also gives them the chance to reassure you, which is what they want to be doing at this stage.

Where this backfires on strong candidates

Two mistakes are common in otherwise excellent finals. The first is the audit. A candidate arrives with eleven prepared questions and works through them in order, taking notes, without responding to anything they hear. The intent is thoroughness; the effect is that the interviewer feels processed, and any warmth built up over five rounds cools in ten minutes. Three real questions asked with follow-through beat a list every time, and if the conversation only supports one, ask the one.

The second is asking a question you have already had answered, particularly one answered earlier in the same loop. It reads as inattention, and in a debrief it is the sort of small thing that gets mentioned. If you want to return to something because the earlier answer was vague, say so explicitly: "your colleague mentioned the reorganisation last week and I did not quite understand what changed for this team". Naming the reason turns a repeat into a legitimate probe.

There is also a trap in what you ask about too early relative to the stage. Detailed questions about policy, leave, notice and working arrangements are entirely reasonable and belong with the recruiter around offer time. Bringing them to a hiring manager mid-loop shifts their attention from whether they want you to whether you want the job on your own terms, which is a conversation you would rather have from a position of having been chosen.

Treat the answers as data

Whatever you ask, listen to the shape of the reply. A manager who cannot say what the previous post-holder went on to do, or who describes the team's last quarter in entirely abstract terms, or who reacts badly to a factual question about on-call load, has told you something about the role that no amount of positive framing will cancel. That is the actual value of this moment, and it only exists if your questions were specific enough to be evaded.

The point of the last ten minutes is not to look interested. It is to leave with the two or three facts you would regret not knowing on your first Monday.