For 20 years now I've been advising people on preparing résumés and getting ready for interviews. And this past year — actually, the past couple of years — I've started sending everyone to language models first, and to come back to me if they still have questions. Well, people have been coming back less and less. Today, with ChatGPT or any other language model, a person can prepare a résumé, get ready for an interview, and gather information about a company. That used to take completely different resources. What's more, AI can even help you pass an interview successfully if you use prompts during it, but it's unlikely to make you successful in the job if they hire, so to speak, not you but your avatar. And there's one more idea that has run like a thread through my whole professional career.
You can't assess a person in one interview, or in two, and you can't fully assess their professional work in that time either. So look at interviews more broadly, more calmly. It's not your last bet. That's what today's episode is about: interviews through the candidate's eyes, and AI. Hi everyone! I've been interviewing people for twenty years. Less so lately, but I still do. I don't know how many thousands of interviews I've conducted in my life, how many people I've hired myself, how many got jobs with my involvement and went on to build their careers.
I've never counted, though it would probably be really interesting. But what matters most to me right now? That AI is changing a great deal in the interview process, and yet the foundation, the basics, remain what they were and still are. What does that mean, when we talk about interviews in general? Let me give one example. Cover letters were never important to me; I never paid attention to them. Maybe that was a professional quirk of mine, maybe in some ways a mistake, but I always did read cover letters, letters from candidates, but I didn't use them as any kind of criterion, as a parameter I applied in my assessment.
And just recently I came across a study from Yale University where they analyzed the Freelancer platform, on which there had always been a very clear link: the more a candidate's letter was tailored to a particular employer's vacancy, the more often — in fact twice as often — the employer replied and responded in some way to that candidate's résumé. Well, some time ago AI appeared on that platform and started helping candidates write exactly these kinds of letters, and that link weakened by more than half.
So essentially, for employers today the cover letter has stopped having any real value, stopped being a meaningful criterion they rely on. Actually, even before, someone could help a person write a cover letter in exactly the same way. A person could write one by doing some research and tailoring it to a specific vacancy. But that took quite a lot of effort. Now it takes much less. And so this element lost even its residual value. The second point, which was and still is the same for me, is credentials in a résumé in general, and the huge attention and weight given to the résumé itself.
For me, personal contact was always extremely important. It always mattered to understand how ready I was to move forward with this person, how they communicate, what soft skills I saw and drew out while talking with them. Of course, if I was hiring for a particular manager or for a company, I looked first and foremost at how capable a given person was of carrying out the tasks, or of helping that manager or that company reach the goals set for the role or for the business as a whole.
But overall, that contact always mattered a lot to me — understanding whether I'm ready to do, with this particular person, the work I'm inviting them to do. On top of all that, one thought always stayed with me: that I can never fully assess a person. An interview lasts an hour, two, maybe even five hours. And there are interviews like that, believe me, quite a few — it's not that rare. So, can you assess a person in that time? People live with each other for years, and there are countless examples where, under certain circumstances, a husband has no idea how his wife might behave, or a wife how her partner might show up in certain behavioral traits. We had an example like that in my family.
My husband and a fairly close friend of his started a business together, and in the end the business didn't work out, and in the end the friendship fell apart too, because as friends they knew each other very well, but as partners they didn't. And the change in circumstances, the change in the shape of that relationship, led them both to realize that they had known each other only within a very small area. So this is something extremely important to me when making decisions in interviews.
That in an interview there is always this right to make a mistake. And here, you know, a mistake doesn't carry a negative connotation; it's actually meant in a fairly positive sense. I think in one of our Sunday episodes — the last one, actually — Alexander even asked this question: so what's the success rate for hiring a sales specialist? Like, 60/40, let's say. And I think the probability of a successful hire is always 50/50. It may sound a bit like reading tea leaves, but again we come back to the fact that getting to know a person in the course of an interview is practically impossible.
The next thing I always take into account is that a success a person had at one company won't necessarily repeat at another company. I've given this example more than once, including to different managers in my own company: Harvard Business School studied star analysts at investment banks. And when they moved, changed jobs and went to other banks, they only started showing exceptional results, and successful results in general, after five years. Because success is always also a team, a particular environment, particular circumstances, and what came together within one company — the experience a candidate brings — is of course extremely important, but that doesn't mean it will repeat with exactly the same result at your company.
All these parameters remain unchanged for me, both in interviews before the AI era and in interviews in the AI era. That certainly doesn't mean everything is completely out of our control. But it sets up a framework that makes interviews more alive and lets you look at them more calmly. And, as I see it, on both sides. The candidate, the person coming to the interview, understands that a rejection isn't necessarily an assessment of their current personal skills, of them as a person, and so on.
In exactly the same way, the interviewer should always understand that at that moment they're not placing their very last bet. So, let's move on. So then what can AI actually give us, if it seems nothing changes, and it seems neither side can influence things much? Well, the whole foundation remains, and I think it's crucial to look at interviews truly broadly. But together with that foundation, there's also preparation for the interview — the candidate's preparation and the interviewer's preparation.
And the whole difference is that today, of course, we can make it way better — higher quality, more efficient and much more effective. While preparing for this podcast, I looked at various studies. To be honest, there are just a million of them. What percentage of interviews, according to researchers at some business school, university or consulting firm — interviews with AI and without AI. But how many job interviews take place in the world each day, nobody knows. I know a figure that, roughly, today four hundred and eighty million people are actively looking for work, according to one study, but there are also a lot of new parameters there.
In the US alone, companies hire more than five million people a month. That means each of them goes through at least one interview, maybe several. So it's extremely hard to rely on any single study and share interesting figures here. What matters more is understanding that any interview study will see one slice, one area, its own audience — it will analyze one platform or one profession. So it can never… It can certainly set a trend for us and give us food for thought, but there's no analysis that would precisely let us — that would give us a formula: what do you need to do to assess someone with 100% accuracy? We're back to this again — how relative the process is.
And to land a given role with 100% certainty. So again, I treat all studies as signals, not as law, but I'll definitely share some interesting figures with you today. So, we're going to an interview. And what matters? First, the résumé. Today we're not talking about résumés. A résumé is just one stage in the overall process of job hunting or hiring someone.
Today we're talking about the interview stage. But still, very often we get invited to an interview based on our résumé. Today, of course, we can tailor it as closely as possible to the position, the role, the company we want to get. And ChatGPT, Claude, any model really, handles this perfectly. You can ask it to edit your résumé for the right role, tasks, company, environment, manager. All the fluff and descriptive filler will be removed. The résumé will show your results to date, concrete actions.
ChatGPT or Claude will tell you what you're missing, what your strengths and weaknesses are. So when it comes to putting together a complete résumé today, models really help us a ton, so don't forget to use them. Second: we know some company is inviting us to an interview. And here, of course, AI helps gather information about it very professionally and to a very high standard. That was possible before, too, but back then, to prepare for an interview, a person simply had to dig through a huge number of different sources.
Not always… Well, honestly, I don't know how many candidates — I didn't find a study like that, though to be fair I didn't look for one — how many candidates read anything more than, say, a page of the corporate website. Just like today, in fact, people come to interviews who don't even know what the company does. So that principle doesn't change either. Of course you need to learn about the company you're interviewing with, and now that's very easy to do, because, first, you can see what state the company is in.
If it matters to you and you're curious what its market capitalization is, what its revenue was last year — if that's public information, if it's a public company. What the trend was, maybe growth, what the headcount is, where its offices are, who represents and who runs the company, what its strategy is, what the parameters of that strategy are, what the plan might be, who the manager interviewing you is — gather information about them, maybe some articles, some press releases about them.
Companies actually tend to publish a lot of information, often even on their own website, but it still takes resources to collect, process and find it all. All facts and figures, of course, should always be checked against the original source. That's a rule of working with AI, but even so, all the models today will be a great help in putting together a very smart, professional dossier. Let's move on. This dossier is a really important background, but the main thing is largely to draw certain conclusions for yourself from it, on the one hand, and on the other, to prepare for the questions you'll be asked, based on the specifics of the company you're going to, and also to prepare questions yourself.
I don't think I'll surprise anyone — this is a very well-known study in the recruiting world, one that's been repeated from different angles a million times — that the more often and the more questions a person asks in an interview, and the more clarifying questions, the more the other person likes us. A very simple life hack. Just ask. People often don't ask, but it's a really smart way to make a good impression, to build that trust, show your engagement, interest and connection, and show your preparation overall.
So the questions, by and large, should be and can be based on the dossier you've put together. And very often, for example, it can even reveal changes — say, that the manager for the position you've come to try for today was appointed only six months ago. That's a very good way to clarify some details and ask what they expect from the new candidate. Or lots of other scenarios. Let's not go down that path, but definitely use this method and this approach. Next: rehearsal. I asked around a bit here, both ChatGPT and Claude. I kept looking — is there anything special?
Something a recruiter or interviewer still doesn't know today about how a candidate might prepare, or how an interviewer might prepare? Of course, these are all fairly basic steps, just stages, which today we can seriously strengthen. And right now we're simply pulling together all this experience of strengthening them. So when asked to find something special, new and interesting, basically all of them — probably the single key approach that all the chatbots we use recommend today is rehearsal.
And I fully support that, because you can always put together the vacancy, your résumé and what you know about the company, and ask ChatGPT to prepare the 10 hardest questions, the 10 most non-obvious questions, the 10 most, you know, I don't know, unconventional tips, maybe, on how to get through this interview, play out different rehearsal scenarios, see what you might be asked, then look at how strong you sounded, how weak you sounded, where there's a weak spot. Give it a try.
I think it's simply an incredible tool today, one that just makes us stronger and makes us — well, and very often lets us prepare well for such an important step, or for an important task in our lives in general. The next one, and probably one of the final preparation points, is the terms. And with models you can now always find out the salary terms in a given region for similar vacancies, which — compare them with your expectations, and understand what figures you can share as your expectations.
But here's a small caveat: be sure to double-check AI at this point. So, as part of preparation — in the European Union there's a directive that was adopted some time ago and, well, is now being rolled out and implemented, with different countries having their own deadlines: the EU Pay Transparency Directive. And in connection with this directive I studied quite a lot of material specifically on pay transparency and, in particular, its connection to AI, and I came across some interesting information.
In Germany, language models were asked the same question: what salary should I ask for in an interview? The position was the same, the level and role were the same, the city was the same, the region the same. Only one thing changed: in one case a woman was asking the AI, in the other a man. And the models advised women to name a lower figure. That's the kind of interesting study they ran in Germany. Maybe today, and depending on how you ask, you'll get a different answer, but as always, we don't rely on it blindly — we check, and look at everything broadly and critically.
Besides being able to find a certain financial, economic footing to rely on before you go to an interview, be sure, of course, to find out all the legal specifics related to hiring, the types of contracts, the types of employment agreements in general that might be of interest to you and that apply in that country, or between your country and the potential employer's country. Because these days we look for work all over the world, and this need really is becoming very relevant.
I often see people come to interviews who don't understand how much, for example, the laws and employment agreements in, say, Poland differ from Germany, or how Poland differs from Belarus, from Russia, from other countries. Today we have the opportunity to clear up all these technical, operational questions too. And probably the last thing. If you have some kind of practical task — and often, well, I don't know, that's a separate story, of course, probably, how to treat practical tasks at all now that AI can be used — but always check with the employer whether using AI in preparing it is acceptable or not.
Because there really may be certain criteria built in that you'll later be assessed on. And if there's no guidance, you can always clarify and ask. Essentially, nothing I've listed replaces in any way the skills and competencies of the person who comes to the interview. And it in no way means that AI does something for them. It's that today we now have the opportunity to show ourselves very well, in a very concentrated way, as fully as possible, in probably our very best light, and to have a better chance of landing your dream job, for example.
Please share: have you had experience going through an interview with AI specifically? Was that experience special? How would you rate it — easier or harder than with a real person? I'm really curious. To be honest, I haven't had that experience, so it would be great if our listeners and viewers could share and tell us about it. In the interview itself, there may be a live person sitting in front of you, like now, or there may be AI. We talked about this in our last Sunday episode too.
How ready or not we are to go through a job interview with AI. To a large extent, of course, it depends on what zone of the job search we're in. Or whether we're not ready to consider certain offers, and there really has to be that personal contact that sells us on a vacancy and draws us to the company. But right now we're talking about something a bit different. These days, of course, a very large number of interviews are conducted with AI, among other formats. It's a very difficult area for, probably, legal regulation in the European Union.
The whole… AI and hiring, AI and human resources in general, fall into a very high-risk category. But nevertheless, according to Greenhouse, a platform that makes software for recruiters and HR professionals, over the past year about half — well, 50% of candidates — had an interview with AI. What's most interesting is that around 70% of those same candidates who had an interview with AI weren't warned in advance about this different form of communicating with the company and didn't know about it at all.
So don't forget to ask and clarify this these days. As we can see, they don't always warn you. I read that candidates in AI interviews often put on an act — well, they try as hard as possible to strip out all emotion, they don't smile, they answer, they look for the most structured answers possible and try to come across fairly dry and analytical. So, first, it's interesting in itself that being interviewed by AI provokes this kind of counter-reaction from people. And second, they do it because there's an assumption that this way the AI will assess them more precisely, more correctly, more objectively.
But in fact there's no evidence that this improves the result in any way. At the same time, there's another interesting study. Journalists from a German media platform tested a program that assessed a candidate's personality from a short video. And a woman candidate, each time she came to this short simulated interview, would put on some new hat, or a coat in a different color — I don't know, yellow, black — put on a headscarf or take it off. And even from such outward things, the score changed noticeably.
So, however trite this may sound, if the score changes with the brightness of the picture, it may also change with what emotions we show, how we speak, the timbre of our voice and so on. Of course, it all depends on how the systems assessing us at the moment are set up. And I think they'll only keep developing and getting more sophisticated, coming closer and closer to human-like resources and assessments. One more thing candidates use during interviews today — and this isn't a recommendation or a non-recommendation, more of an observation — is using AI during the interview.
Honestly, it's very hard to tell whether a person is using it or not, especially if the interview is online, on the other side of the screen. It can be done so skillfully and so cleverly now. First, there are various dynamic assistants that will feed you prompts in real time. Or you can simply set up your environment very cleverly so that you can communicate while also relying on recommendations that, I don't know, say, ChatGPT is giving you. I also tried to find some figures, roughly, to rely on and give as examples.
They do exist, give or take — lots of studies — but they all say that in 70% of cases you can't tell for sure. So it's all a bit, you know, within the margin of error. You can guess that a candidate is using AI, or you can guess that they're not. My position is that if, as a rule, a certain framework has been spelled out in advance — a request not to use AI in the process — then we don't use AI in the process. But if — how should I put it? on the other side of the screen there's a candidate who uses AI so skillfully that you can't even tell they're using some kind of helper or assistant — well, great! That means this person will work just as skillfully, and maybe this is exactly the person you need.
This is probably more my personal, life position than a recommendation based on professional judgment. That it may well help you get the job, potentially, but not always to be successful in it. And you can't fool yourself. So here, too, try to assess and balance that line between where you're who you are, however trite that sounds, right? and where you need to strengthen yourself in this social world, including with the tools that AI gives us today.
And the next block — essentially the last one, right? So: preparation, the interview process, and after the interview. After the interview, I just wanted to highlight maybe two ways we can use AI. One is to go over with AI, while you still remember, how you answered all the interviewer's questions, and ask what went well, what may have been weak, where you gave too much descriptive detail. If the interview was recorded with mutual consent, you could perhaps ask for the transcript of the recording and upload it to, I don't know, ChatGPT.
And then, of course, thanks to this, ChatGPT, Claude, any other LLM today will help you prepare for the next interview, and you'll feel even more confident. You know, it's a tremendous training tool, and I kind of think that career consulting in the form it used to exist has already died, or is dying. For 20 years I've received requests like that from my circle, from candidates, from a number of other colleagues, for help with résumés, preparing for interviews, post-interview assessment, help preparing for the next interview.
And today, if any of my acquaintances are listening to this podcast, they know for sure: I send everyone to the models first. And only if, after working with ChatGPT, a person still has questions about how exactly to edit a résumé or a LinkedIn profile, or prepare for an interview, then of course I give my own recommendations in each specific case. But overall, I believe that today anyone, using this whole arsenal, can become a real supercandidate. At the same time, not everyone prepares, not everyone uses it. And as I already said today, many people come to a company not knowing what company they've come to meet with.
So strengthen yourselves. And the last, most important thing: don't forget to ask the interviewer about the next step. This isn't about AI anymore, of course, but let's close the loop. Because according to the same Greenhouse I mentioned today, practically half — I think it was 48% of candidates, the figure was around that — complain that the employer just disappears, and not only doesn't indicate the next step but vanishes without a reply altogether. And very often people are left without any feedback on the meeting they had.
That's completely unacceptable — in all the teams where I worked, managed and set up recruitment, it was probably one of the first rules: always respond to the person, always respond to the person personally by any available means: a call, a text — well, depending on the context, depending on the communication style — an email, and so on. But you should always be told what the next step is, and you should always be told what to expect and how long to wait. To wrap up this candidate journey, to finish, I think it's very important to use
AI to strengthen your competencies, to strengthen your views, not to replace them. You know, it's even a bit like preparing for a podcast. Today you can gather, and you have a tremendous opportunity to read and collect in concentrated form the materials, studies and papers you need, and see what's happening around the world. But AI won't replace your words, your own line of thought, today. So my, probably, most important recommendation — one that runs through all our episodes — is: use all the current technological possibilities to make yourself even more professional, even better, and to really show who you are, especially with such power at your disposal.