Behavioral Interviews and the Rehearsed Answer

When a test matters enough, people start preparing for it. And if they can prepare enough, the test stops measuring what it was designed to measure and starts measuring the preparation.

I think that's what has happened to the behavioral interview.

The idea behind it was sensible. Past behavior predicts future behavior, so instead of asking candidates what they would do, you ask what they did. Tell me about a time you handled a conflict, or led a team when nobody knew the answer, or recovered from a failure. The research generally supported this: structured behavioral interviews did better than loose, unstructured conversations for most kinds of roles. Companies standardized the questions and trained interviewers to push for specifics, on the theory that specifics are hard to fake. Someone who really did the thing can tell you the details. Someone bluffing can't.

That theory held as long as preparation was limited. It doesn't hold when a candidate can rehearse every common behavioral question with an AI coach, get feedback on each answer, and do it again as many times as they like before the interview.

What makes this worse is what the coaching focuses on. The prep tools teach candidates to add concrete details, mention numbers, and turn every failure into a tidy story about what they learned. In other words, they train exactly the thing interviewers were told to look for. The part of the answer we trusted most has become the easiest part to practice. What the interviewer hears now is a well-rehearsed version of authenticity, and it can sound more convincing than the real thing.

You can see the effect if you know where to look. Suppose a hiring manager remarks that the last few hires "interviewed really well but seem to struggle with ambiguity." Most teams would treat that as a problem with the manager's judgment. I suspect it's more often a problem with the method. And it's being noticed at the worst possible time, after the person has already started.

So what still works? The useful question is what a candidate can't rehearse, and the answer is anything that depends on what they just said.

A fixed list of questions with a fixed scoring rubric can always be prepared for, however well it's designed, because the candidate knows roughly what's coming. A conversation that follows the thread of the answer is different. If the interviewer picks up on an odd detail or a small inconsistency and asks about that, the candidate can't have prepared the next question, because it didn't exist until they answered the last one.

Here's an example. A candidate has rehearsed a smooth answer to "Tell me about a time you led through ambiguity." They deliver it well. Then the interviewer asks, "When you say the team was uncertain, what did you actually know, and what were you guessing at?" There's no script for that. Whatever comes next shows you how the person thinks when they haven't practiced, which is the thing you were trying to learn in the first place.

Work samples and paid trials are even harder to fake, since the candidate is doing real work under realistic conditions. They're still the best predictors we have. But they cost a lot to run and suit some roles better than others. Nobody is going to run a paid trial on every applicant in a 400-person pipeline. They belong later in the process, once the pool is small.

That leaves the top of the funnel, which is where we've spent our time at Asendia. From the applicant's side, it goes like this. They apply, and that same evening their phone rings. The caller is software, and it listens to what they say rather than checking answers against a behavioral template. If they give a polished answer, the next question picks up a detail from it and goes somewhere the rehearsal didn't. So they end up talking about what they actually did, in their own words, which is what a behavioral interview was supposed to get at in the first place.

On the recruiter's side, what arrives is a shortlist based on what candidates actually said, with verbatim quotes and a qualification summary written into the ATS. Agencies running big campaigns use it to handle hundreds of applications without adding recruiters, since the AI has every first conversation and the recruiters pick up from there. The piece on agentic recruiting goes into why that first conversation shapes so much of what happens later in the pipeline.

None of this means behavioral interviewing was a bad idea. It rested on an assumption that candidates would arrive lightly prepared, so that asking for specifics would separate real experience from generic preparation. That assumption no longer holds. If you're using the same interview framework you used in 2021 and your recent hires don't match how well they interviewed, this is probably a large part of why.

You don't need to give up on structure. You need to stop handing candidates a fixed target to prepare for. Adaptive conversations can do that at the top of the funnel, and work samples can do it further down.

Ready to transform your hiring strategy? Schedule a Demo with our founders today!