What happens when an AI interview hits its time limit

On 28 September 2026 we changed one number in our AI interviewer and deliberately left another one alone. The change was for conversational jobs, where the interviewer works through the hiring team’s own questions in a natural back-and-forth. Those plans can run to about 18 minutes when a team sets ten or more questions, and our standard wrap-up at 15 minutes would have cut into them. So for conversational jobs the wrap-up moved to 17 minutes. The number we left alone was the end: every interview still closes at 20 minutes, whatever style it is.

That pair of decisions is the whole design of the AI interview time limit in HireQwik: a soft limit that can move with the interview, and a hard limit that never does. This post explains, for our own AI interviewer, exactly what happens at each one, what the candidate hears, how an interview that reaches the limit is scored, and what a recruiter should look at when it happens often. If you are a candidate, the short version is reassuring: running out of time is not a fail.

How long is an AI interview, and what is the time limit?

A HireQwik AI interview usually takes 10 to 20 minutes. At 15 minutes, or 17 on conversational jobs, the interviewer tells the candidate it is wrapping up and moves to closing. At 20 minutes it thanks them and closes the call. The interview is then scored on what was covered, and unasked questions do not count against the candidate.

Other AI interview tools work differently, so it is worth being precise. There is no countdown on any single answer. Nothing on the screen ticks down. The limit is on the whole interview, it is the same for every candidate on every job, and the candidate hears about it before it arrives.

The timeline of a HireQwik interview

Here is what happens, in order, from the moment the interview starts.

Two details in that table do most of the work. First, if the interview finishes normally before 15 minutes, nothing else happens: no heads-up and no second goodbye. Most screens end well before the limit. Second, the score is not produced during the call. The evaluation runs after the call ends, which is why the verdict lands within minutes of the candidate leaving, not while they are still talking.

What the candidate hears at each limit

The heads-up and the close sound different on purpose.

On a standard interview, the heads-up is a fixed sentence: “Just a heads up, we’re running a bit long. Let me wrap up with a couple quick closing points.” The candidate can talk over it. If they are mid-answer, they can finish their point; the sentence is a signal, not a stop.

On a conversational interview, there is no fixed sentence. The interviewer is told to finish the current topic in one short sentence, ask whether the candidate has any questions, and close, without mentioning a time limit or that it is ending early. The reasoning: in a flowing conversation, announcing the clock makes the last minutes feel rushed, and leaving it unsaid lets the interview end the way a conversation does.

The close at 20 minutes is the same everywhere: “Alright, we need to wrap up here. Thank you so much for your time, it was great chatting with you! The team will review everything and get back to you soon. Have a great day!” Unlike the heads-up, this line cannot be interrupted, and once it is spoken the interviewer does not add a goodbye of its own. A few seconds later the call ends.

We chose that wording carefully. The interviewer does not tell the candidate how they did, and it does not promise a date it cannot know. It promises what is true: people on the hiring team will review the interview. If you want the human version of this ending, with scripts, we wrote how to end an interview as the interviewer.

Why an AI interview needs a hard limit at all

It is fair to ask why an AI, which does not get tired, needs a time limit. There are three reasons, and the candidate is the first.

Candidates plan around the time you give them. The invitation says the interview takes about 10 to 20 minutes, and people book that time between shifts, lectures and other interviews. A 35-minute interview breaks that promise. Among North American candidates who dropped out of a hiring process in Talent Board’s 2018 survey, 37% named disrespect for their time in interviews as a reason. A limit that holds for every candidate is the simplest protection against that.

The next candidate’s slot depends on it. Candidates book interview slots, and a slot has a fixed number of seats. When our booking system decides how many seats a slot has left, it plans for a typical call of 7 to 10 minutes and counts an interview that could still be running from the previous slot at half weight. That planning only works because no interview can run past 20 minutes. We explained the seat counting in detail in what a full interview slot actually means.

A fair comparison needs the same conditions. Structured interviewing rests on asking every candidate the same questions and grading them on an identical scale. The same time limit is part of the same conditions. A candidate who got 35 minutes to talk is not being compared fairly with one who got 15.

There is also a plainer engineering reason. A voice conversation between a person and an AI can go wrong in ways neither side intends: a confused exchange that loops, a candidate who never says goodbye. Without a limit, those calls would run until someone noticed. With one, the worst case is 20 minutes.

How an interview that reaches the limit is scored

This is the part candidates worry about, and the answer is the same every time: reaching the limit is not a fail.

When an interview reaches 20 minutes, it is saved as completed and goes through the same scoring as any other interview. Each question that was asked is scored from what the candidate actually said, with their own words quoted as evidence. The score and the recommendation (Strong Go, Go, On Hold or No Go) are worked out exactly as they would be for an interview that ended at minute 12. We broke that down in how an AI interview score is calculated.

Questions that were never reached are simply not scored. They do not count as zero. And if the interview ended with too little of the screen covered to judge the candidate fairly, the result is marked Incomplete rather than given a low score. The rule behind that, and what HR should do with an Incomplete row, is in what an incomplete AI interview means.

The honest trade-off is this: a candidate who used most of the time on two questions has less evidence behind their score than one who answered six. The score is fair to what was said, but there is less to say it from. That is a reason for recruiters to look closely when interviews hit the limit often.

What HR should check when interviews hit the limit

An occasional interview that reaches 20 minutes is normal: a talkative candidate, a reconnect, a long question about the company. If it happens on many interviews for one job, the setup is usually the cause, not the candidates. Check these, in this order:

1. Count the questions. On a conversational job, the planned length rises with the question count: around 12 minutes for five questions or fewer, 15 for six to nine, 18 for ten and above. A job with twelve questions is planned to run close to the limit before any candidate opens their mouth. Trim to the questions that decide the role.
2. Look for compound questions. “Tell me about your last role, why you left and what you want next” is three questions. Split it, or cut two parts.
3. Read two or three transcripts. If candidates are spending minutes asking about the company, answer the common questions (location, shifts, rounds) in the invitation so the interview time goes on the screen itself. If the interviewer is repeating itself, tell us; that is a bug, not a setup problem.
4. Check the deal-breakers come first. On structured jobs, a candidate who cannot meet a hard requirement should find out in the first minutes, not after fifteen.
5. Compare scores. If interviews that hit the limit score noticeably lower than the rest, the time limit is costing those candidates evidence, and the question list needs work.

The fastest fix is almost always the first one. A screen with six well-chosen questions beats a screen with twelve, because every question gets a real answer.

Can a job have a longer interview?

Not today, and that is deliberate. The heads-up moves between 15 and 17 minutes depending on the interview style, but the 20-minute close is the same for every job and every account. Hiring teams sometimes ask for 30 or 40 minutes for senior roles, and the honest answer is that a longer AI interview is usually the wrong tool for that problem.

A first-round screen exists to decide one thing: is this candidate worth a longer conversation with a person? Twenty minutes is enough for that if the questions are well chosen, and the extra time a senior role needs is better spent in a hiring manager round, where someone who knows the work can follow the candidate’s thinking. Stretching the screen tends to produce the same evidence at greater length, and it costs every candidate who did not need it.

If a job keeps hitting the limit, the fix is in the question list, not the clock. For later rounds that genuinely need forty-five minutes, the human version of the same structure, with a soft limit and a hard end, is laid out in how long an interview should last.

What to tell candidates about the time limit

Candidates handle a time limit well when they know about it in advance and badly when it surprises them. Put three facts in the invitation or the briefing:

• The length. “About 10-20 minutes, spoken, in your browser.” Our candidate briefing uses the same length.
• That there is no timer on each answer. Candidates who have used one-way video tools often expect a countdown, and the expectation alone makes them rush.
• That the interviewer will tell them when it is wrapping up. “If you are still talking near the end, the interviewer will let you know and close politely. That is not a fail.”

That last sentence removes most of the anxiety. A candidate who hears “we’re running a bit long” knows what it means because you told them in advance.

How this differs from one-way video interview timers

Many one-way video tools put a countdown on each answer: a question appears, a timer starts, and the recording stops when it runs out. That has one real advantage. Every answer is the same length, so reviewers can compare them quickly.

It also has a cost, and candidates feel it. A per-answer timer punishes the person who pauses to think, cuts off the end of a good answer, and turns the interview into a race against a clock on the screen. It is the opposite of what a spoken conversation is like.

We chose a whole-interview limit instead. Each answer can be as long as it needs to be, the interviewer can follow up on a short one, and the time pressure only arrives near the end, with a spoken warning first. The trade-off is that one very long answer can crowd out a later question, which is why the question list matters so much. We compared the two formats more broadly in AI voice screening versus one-way video interviews.

The short version for recruiters and candidates

For candidates: a HireQwik interview takes 10 to 20 minutes, there is no timer on any answer, the interviewer will tell you when it is wrapping up, and reaching the end of the time is never a fail.

For recruiters: the heads-up at 15 (or 17) and the close at 20 are the same for everyone, an interview that reaches the limit is scored on what was said, and interviews that hit the limit often are a sign the question list is too long.

If you want to hear the timing from the candidate’s side, take the five-minute live demo. It is shorter than a real screen, but the pacing and the close work the same way.