How do I do a mock interview by myself

Out loud, recorded, and timed. Rehearsing in your head is not practice, because your inner voice never stumbles or runs long. Record on your phone, hold behavioral answers to ninety seconds, and ask yourself the follow-up after every answer rather than moving on.

How many mock interviews should I do

Enough that the follow-ups stop surprising you, which is usually ten to fifteen reps rather than two or three. Stop rehearsing an answer once it is roughly right and spend the remaining time on the second and third questions, which cannot be scripted.

Are AI mock interviews any good

For volume and for finding where an answer is vague, yes. They probe what you actually said rather than what you meant, and they do not get bored. They cannot judge likeability, industry specifics or the pressure of a real room, so use them for reps and a human for calibration.

What questions should a mock interview include

An opener, three or four behavioral questions across different competencies, a pressure question, a weakness question and a closer. Fifteen with follow-ups runs about forty-five minutes, which is a realistic first round. Practise them in that order rather than at random.

Why do my answers sound worse out loud

Because writing and speaking use different faculties, and the version in your head skips the parts you have not worked out. That gap is the entire reason to practise aloud. Answers that only work when you are fresh and uninterrupted are not ready yet.

Mock Interview Questions to Practice Out Loud

Reading interview answers does almost nothing. The gap between knowing what a good answer contains and producing one out loud, under time, with somebody waiting, is the entire problem. Below: a fifteen-question mock in a realistic order, the follow-up after each, and a way to score yourself that is harsher than being kind to yourself and more useful.

How to run a mock interview alone

Three rules, and they are the whole method.

Out loud, always. Rehearsing in your head is not practice. Your inner voice never stumbles, never says um, and never loses the thread at forty seconds, so it gives you no information about what will happen when your mouth is doing the work. If you only change one thing, change this.

Record it. Voice memos on your phone is enough. You will not want to listen back and you should anyway, because the gap between what you think you said and what you actually said is where all the useful information lives. Most people discover they are twice as long and half as specific as they believed.

Time it. Ninety seconds for a behavioral answer, sixty for an opener. Set a timer where you can see it while you speak. Running long is the most common failure in real interviews, and without a clock in front of you it is completely invisible, because time moves differently when you are the one talking.

One more thing that helps: do it standing up, on a day you are tired. Interviews are rarely at a convenient hour, and an answer that only works when you are fresh is not ready. If you can, practise at roughly the time of day your interview is scheduled, because a nine in the morning slot and a four in the afternoon slot are genuinely different tests.

The 15-question mock, in order

A real interview has a shape. It opens wide, narrows into evidence, applies pressure in the middle, and closes on motivation. Practising questions in a random order trains you for a test nobody administers, and it means you never rehearse the thing that actually goes wrong, which is fatigue in the back half.

Run these in sequence, ninety seconds each, with the follow-up asked immediately after every answer.

Tell me about yourself. Follow-up: you mentioned that project, what was your part in it specifically.
Why this role. Follow-up: what would make you leave within a year.
Walk me through your last twelve months. Follow-up: what would you do differently if you had them again.
Tell me about a time you failed. Follow-up: what was your mistake, rather than the team's. See the full answer.
Tell me about a conflict with a coworker. Follow-up: what did they think you did wrong. See the full answer.
Tell me about a decision you made without enough information. Follow-up: what was your confidence, as a number.
Describe a time you had to learn something fast. Follow-up: what did you get wrong in the first week.
Tell me about a time you disagreed with your manager. Follow-up: did you do it their way, and were they right.
How do you prioritise when everything is urgent. Follow-up: what did you drop last time, and did its owner agree.
Tell me about a time you worked under pressure. Follow-up: what specifically was the pressure, and what dropped.
What is the hardest feedback you have received. Follow-up: what was it, word for word.
Tell me about something you built or changed that nobody asked for. Follow-up: what happened to the work you were supposed to be doing.
Where are you weakest. Follow-up: who would say that about you, and when did they last say it.
What questions do you have for us. Follow-up: none, but notice whether yours were about the work or the perks.
Why should we hire you. Follow-up: what would the strongest other candidate have that you do not.

Set aside forty-five minutes to get through all fifteen with the follow-ups. That is a realistic first round, and running the full sequence once is worth considerably more than running four favourite questions six times each.

The order matters more than people expect. Question four lands differently when you have already spent five minutes establishing who you are, and question fifteen is much harder when you are tired than when it is the only one you practise. If you only have twenty minutes, run one to five in order rather than cherry-picking.

Scoring your own answers

Listen back with a rubric rather than a feeling. Five checks, one point each, per answer.

Length. Did it come in under two minutes. Time it rather than estimating, because the estimate is always generous.
The number. Is there at least one specific figure, date or quantity.
The pronoun. At the decision points, did you say I or we. Count them.
The cost. Does the story contain something that went wrong, or somebody who paid.
The follow-up. Did your answer already contain what the follow-up asked for.

Three or below and the answer needs rebuilding rather than polishing. The fifth check is the one that matters most and the one people skip, because it requires admitting the follow-up found something.

Be harder on yourself than feels comfortable. The temptation when listening back is to score the version you meant to say rather than the one on the recording, and the interviewer only hears the recording.

Track the scores across sessions. What you are looking for is not a rising average, which mostly measures familiarity, but whether the fifth check improves. Answers that already contain what the follow-up wanted are the only ones that hold up when somebody who does this professionally decides to push.

Doing this with an AI interviewer

The obvious limitation of practising alone is that you ask your own follow-ups, which means you only ever get asked about the weaknesses you already know about. That is the specific thing an AI interviewer fixes: it reads what you actually said and probes the part you left vague, including the parts you did not notice you left vague.

What it does well is exactly that. It is available at eleven at night, it does not get bored on the fourth attempt, it will not soften a follow-up to spare your feelings, and it scores content rather than delivery.

What it does badly is worth being honest about. It cannot tell you whether you seemed likeable, which is a real part of a real interview. It does not know your industry's specifics the way a hiring manager in your field does. It will not react to your body language, and it cannot replicate the particular pressure of a person who has the power to reject you sitting three feet away. Anyone selling AI practice as a complete substitute for a mock interview with a human is overselling it.

The sensible use is to run the volume with software and the calibration with a person. Do fifteen reps against an AI interviewer to find where your answers break, fix them, then do one mock with somebody in your field to check whether the fixed version lands.

Where your practice breaks

There is a failure mode specific to practising alone, and it is worth naming because it feels like progress while it is happening.

You run the fifteen questions, your answers get smoother each time, and you conclude you are ready. What has actually happened is that you have memorised a performance. The words come faster, the pauses disappear, and the whole thing acquires the particular fluency of something recited. Interviewers detect that instantly, and it works against you, because a recited answer sounds like it was built to be impressive rather than to be true.

The tell is that a memorised answer cannot be interrupted. Ask a fluent candidate a mid-answer question and watch them either ignore it and finish the paragraph, or lose the thread entirely. Both are worse than a rougher answer from somebody who is actually thinking.

The fix is to practise the follow-ups rather than the answers. Once your ninety seconds is roughly right, stop rehearsing it and spend your time on the second and third questions, which are the ones you cannot script because they depend on what you just said.

A useful drill: take an answer you are happy with, and have someone interrupt you at forty seconds with a question about a detail you mentioned in passing. If you can absorb the interruption, answer it, and get back to your point, the answer is genuinely yours. If you have to restart from the beginning, it is still a performance.

For the questions themselves, grouped by what they test, see behavioral interview questions. For the format underneath them, see the STAR method.