Rehearsing Out Loud Is Not the Same as Knowing the Answer

The candidate who answered only the question I asked

I interviewed someone once who answered every question correctly and told me almost nothing.

The answers were right. They were also exactly as long as the question. I'd ask about a piece of work, and I'd get the piece of work, with no context around it, no detail underneath it, nothing about why it was hard or what else they'd tried. Then I'd ask a follow-up, and the follow-up answer was good. So I'd ask another one, and that was good too. The depth was there the whole time. It just never came out unless I went and got it.

So I kept going, and that interview turned into a long chain of follow-up questions, more than I'd normally need to ask, because every time I pulled on something there was something real behind it.

What stays with me about that interview is that another interviewer could have missed the candidate completely. Whether the depth surfaces depends on the interviewing skills of the person sitting across from you, and those vary a lot. Some interviewers dig. Some take the first answer, write it down, and move to the next question on the sheet.

You don't get to pick which one you get.

It's a presentation skill, and it gets scored like one

What's being tested in a moment like that isn't knowledge. It's the ability to present yourself: to sell your expertise, your skills, the work you've already done. Nobody writes that on the scorecard in those words, but that's what's happening.

From my chair the tells are consistent. Lack of confidence in the answer and in the voice. Lack of details. The answer is technically fine and somehow weightless.

Underneath the uncertainty there's usually something specific going on: people just don't know whether the answer they gave was a good one. They've never heard themselves give it, so they have no reference point. They hedge, and hedging sounds like not knowing.

How much this costs you depends on the role, and this is where other hiring managers disagree with me.

For a junior role I can live with it. If I believe the person can grow in that area, a candidate who needs prompting to show what they know is still a hire. That's a skill you build, and someone on the team will help them build it.

For a senior role, or any role where communicating clearly is already part of the job, it's a no go. The knowledge is there. The problem is that the job is partly the explaining, and if a senior engineer can only surface their reasoning when someone skilled interrogates them, that shows up on an ordinary Tuesday afternoon too, not just in an interview.

The important detail arrives after the interview

Ask me what breaks when someone speaks an answer for the first time in the real interview and it isn't a wrong answer. A wrong answer would be easy to score. What I see instead is confused, uncertain speech, and then something more specific that happens afterwards.

The person answers. Only some time later do they remember the important details.

Sometimes that lands inside the interview. They come back ten minutes on with "actually, on that earlier question, I should have mentioned something." That recovers the score. I don't penalise it, I like it, it tells me they're still thinking about the problem.

The problem is how often the memory arrives after the interview is over, when it doesn't recover anything. The answer that got written down is the thin one, and the good version exists only in the candidate's head, an hour too late to be worth anything.

Reading your notes cannot close that gap. Notes tell you whether you recognise the material. They say nothing about whether you can retrieve it on demand, in order, at speaking speed, while someone watches you and decides. Those are different skills and only one of them is being tested. Saying the answer out loud once moves the retrieval into rehearsal, where being slow costs you nothing.

Will I sound rehearsed? Mostly, no

If you're worried that practising will make you sound scripted, the honest answer isn't the flattering one.

It's hard for me to tell. Sitting on the interviewer side, I can't reliably hear that a candidate prepared. Some people pass an interview without any preparation at all, because they simply speak well, and I have no dependable way to separate them from someone who ran the same story out loud the night before.

Worth saying plainly, because it cuts against what a lot of interview advice promises. It also isn't an argument against preparing. It's an argument about how uneven the starting line already is. Some people don't need the preparation. It's still better to have it.

There is one kind of over-preparation I do catch, and it isn't smoothness. Some people prepare very well and lie. The story is polished and some part of it didn't happen. That version survives the prepared question and dies on the follow-up, because you can't rehearse the follow-ups. I keep asking, past the edge of what was prepared, and the whole thing thins out into generalities.

Which tells you what to rehearse: the project you really worked on, not a better one. Rehearsal makes a true story easier to tell. It doesn't make a false one survivable.

What AI mock interview practice gives you, and what it doesn't

I make a product that does this, so start with what it can't do.

It doesn't give you the emotions, or any of the human stuff. The interviewer who goes quiet at the wrong moment, the one whose face you can't read, the weight of a conversation you care about. No practice session reproduces that, and I'd be lying if I said otherwise.

What does transfer is narrower and duller than the marketing usually suggests. Repetition. Preparation. Being asked the same questions and answering them again, mechanically repeating one process until it stops being unfamiliar. That's the whole claim, and it happens to be the exact thing missing when you prepare by reading, the thing that decides whether the important detail shows up during the interview or an hour after it.

You don't need a product for it either. Anything that makes you say the answer out loud, on demand, more than once will do the mechanical part. What software gives you is that it will ask the same question a fourth time without getting tired of it.

Start with where you're going, not with your notes

If you're two weeks out from an onsite and your current plan is to re-read what you wrote down, the first thing I'd change isn't rehearsal.

Investigate the context around the vacancy first. The role, the company, what they're actually hiring this person to do. Understand clearly where you're going before you decide what to prepare, because otherwise you're preparing in general, and preparing in general is close to not preparing.

When I was preparing for my own interviews, that was the hardest part. Not the answers. Knowing which questions would be asked. Once I found examples of the real questions the rest was mechanical, and I practised against them out loud.

So the order is: work out where you're walking in, find the questions that role actually gets asked, then say your answers to something that can hear them. Repeat until the answer stops rearranging itself on the way out. That's usually when you discover your best story takes a long time to get going, or that you've been leaving out the part that makes it interesting.

Your notes will tell you that you know the material. You already know the material. The only thing that tells you whether it comes out in the right order, at speed, in front of someone who is judging you, is having said it before.