Latest / Future of Work Tech with Fexingo: Remote Tools, AI Productivity, and Workplace Software / How AI Is Rewriting Your Job Interview Script
Transcript
- Lucas: So there's this stat I've been turning over in my head: candidates who used an ai powered interview coach improved their offer rate by 18 percent. That's from a study out of Stanford's digital economy lab, published last quarter. Luna: Eighteen percent is huge. But what exactly does an AI interview coach do? Is it like a chatbot asking you questions? Lucas: It's more sophisticated than that. These platforms — there are a handful now, Yoodli is one, another called Interviewing.io — they run a simulated interview where an AI avatar asks you questions. But the real magic is in the feedback loop. Lucas: They analyze your speech patterns: filler words like 'um' and 'like', your pace, your word choice. Some even track facial expressions through your webcam and tell you if you looked too tense or not engaged enough. It's essentially a mirror that talks back. Luna: I've actually seen this in action. A friend of mine was interviewing for a product manager role at a FAANG company, and she used one of these tools for two weeks straight. She said the first time she saw her own recording she was shocked at how many 'ums' she used. Luna: She landed the job, by the way. But she also said it felt a little weird — like she was being trained to perform rather than to have a real conversation. Lucas: That's the tension exactly. The tools can make you more polished, but there's a risk of over-optimization. You end up sounding like you're reading from a script that the AI helped you craft. And hiring managers, especially experienced ones, can smell that a mile away. Lucas: But the companies making these tools argue they're democratizing access. Historically, if you had connections, you could do mock interviews with friends in the industry or with a career coach. That costs money or social capital. An AI coach is available 24/7 for a fraction of the cost. Luna: The Stanford study — did it control for other factors? Like, maybe the people who used the coach were just more motivated in general? Lucas: Good question. The study was a randomized controlled trial. One group got access to the AI coach, the other group had mock interviews with human peers. Both groups had the same number of practice sessions. So the improvement seems tied to the quality of the feedback from the AI. Lucas: The AI gave more granular feedback: 'You paused for 2.4 seconds after that question — that's too long, it signals uncertainty.' A human peer might not catch that. So the AI is effectively teaching you the hidden rules of interview performance. Luna: Which brings up the fairness question. If only some candidates use these tools, doesn't that create a new kind of advantage? It's like the SAT prep arms race all over again. Lucas: Exactly. And there's a flip side too: companies are also using AI to screen candidates. So you've got AI on both sides of the table. The candidate is trained by AI to say the right things, and the recruiter uses AI to parse those responses for keywords. Luna: It starts to feel like a game of cat and mouse. Or maybe more like two AIs negotiating with each other. Lucas: Right. And that's where the ethics get murky. One concern is that the AI coach might reinforce bias. If it trains everyone to sound like the typical successful candidate from the past, it could homogenize the workforce. You lose the quirky thinker who might be brilliant but doesn't fit the mold. Luna: There was a case recently — I think it was in the news — where a company's hiring AI was biased against certain speech patterns. If the coaching tool is trained on similar data, it could be coaching candidates to suppress their natural accent or dialect. Lucas: That's a real risk. And some of these platforms are starting to address it. Yoodli, for instance, now has a feature that flags bias in your own language — not just your filler words but whether you're using gendered or exclusionary terms. So the tool can actually help you become more inclusive. Lucas: But the deeper question is: should we be optimizing for interview performance at all? Maybe the interview process itself is flawed, and we're just perfecting a broken system. Luna: That's a really good point. I think a lot of companies are moving toward skills-based assessments and work samples instead of behavioral interviews. But those haven't replaced the traditional interview yet. Lucas: No, they haven't. And until they do, candidates will keep looking for an edge. Which is why these tools are growing fast. A report from MarketsandMarkets projects the AI interview coaching market will hit $1.2 billion by 2029, up from about $300 million this year. Lucas: And speaking of edges — and this is going to sound like a plug, but I mean it genuinely — if you're finding these conversations useful, one way you can keep the show ad-free and independent is by supporting us. It's buy me a coffee dot com slash fexingo. Luna: Yeah, honestly, even a one-time contribution helps cover the production costs. We try to keep the tech talk practical and grounded, and listener support is what lets us do that without sponsors. Lucas: Alright, back to the interview tech. One development I'm watching is the integration of these coaching tools directly into job platforms. LinkedIn already has a feature called Interview Prep that gives you feedback on recorded answers. It's basic right now, but you can see where this is going. Luna: So eventually, you might practice your interview on the same platform where you'll submit your application. That's a seamless experience, but it also means LinkedIn has data on how you perform under pressure. Lucas: Exactly. And they could use that data to rank candidates or feed into their hiring products. That's a whole other privacy discussion. But for now, the immediate question for most job seekers is simply: should I use one of these tools? Lucas: My take is: use it as a diagnostic, not a crutch. Do a couple of sessions, identify your tics — like overusing 'basically' or trailing off at the end of sentences — then practice on your own. Don't let the AI script your answers. Luna: I think that's solid advice. And for hiring managers listening, maybe the lesson is to design interviews that are harder to game. Instead of asking 'Tell me about a time you led a team,' give them a real problem to solve on the spot. Lucas: Yeah, that's the direction the best companies are moving. And ironically, AI can help there too — by generating realistic case studies or coding challenges that are unique to each candidate. Luna: So the arms race continues, but maybe it's shifting from who can perform best in an interview to who can solve the actual problems the job requires. That's a win, I think. Lucas: It would be. But old habits die hard. Until we see a mass shift to skills-based hiring, the AI interview coach is here to stay. And honestly, if it helps someone who's been out of the job market for a while regain their confidence, that's not nothing. Luna: Agreed. Confidence matters. And if a tool can give you that extra boost, even if it's just by pointing out that you say 'um' too much, it's probably worth a try. Lucas: Right. And as always, we'll keep watching how this evolves. Next episode, I want to dig into something a bit under the hood: how companies are using AI to write interview questions — and whether those questions are actually better than the ones humans write. Luna: That sounds like a natural follow-up. I'm curious to see if ai generated questions are less biased or just differently biased. Lucas: That's the million-dollar question. We'll find out together. Thanks for listening — we'll catch you next time.