Episode 59

Xavier Gorriz Torner on Enhancing Recruitment by Leveraging AI in the Hiring Process

TIM: Xavier Welcome to the Alooba Objective Hiring Show. Thank you so much for joining us.

XAVIER: Thank you for inviting me.

TIM: Oh, it's absolutely our pleasure. Xavier and I would love to start our conversation today with probably one of everyone's favorite hot button topics at the moment: AI, those two little letters that seem to be redefining everything about the world as we know it, and I'd love to get your thoughts on the use of AI in the hiring process. Perhaps what you might've seen from candidates so far, and then have you started to dabble yourself in using AI, or could you imagine any particular bits of the hiring process that you'd like to use AI in in the future?

XAVIER: Yeah, to be honest, in Deriv we use a lot of AI in general to improve the productivity of the teams, and this goes also for the recruitment team for the recruitment process. Then, of course, we use it not for everything; it's more a tool that helps us in the process, and what we try is to use it for screening candidates in general to just remove the ones that don't make sense at all. It's not about removing good candidates or candidates that are in the middle; it's more about, let's get out from the thousands of candidates and let's remove the one that doesn't make sense at all. This is one of the first steps. also candidates that maybe they are cheating in our tech in our tech test that we do and all the other cases, for example, in my case, I use it for helping me. I have a pool of questions that I use for different candidates, and then it helps me, yeah, to be faster and to be more adaptable to the candidates that I'm going to interview.

TIM: So it sounds like you guys have gone pretty all in on AI and hiring already. Based on if I just compare this to other people I've spoken to, you guys are probably at a reasonably advanced level. I wonder if that's because it's just so widely used in the rest of the organization that maybe people in your company are a little bit more open-minded to AI in general. Do you think that could be part of it?

XAVIER: Yeah, exactly. Our CEOs are really bullish on AI, and that's one of the things that comes from all types of places in the company that we want to use AI to improve our processes to improve To automate to that helps you be more productive, that's one of the things that it gains. and then, yeah, I'm also in charge of the AI team, and we have been working a lot on small tools that help each department to do their work faster, and this also implies doing training for different departments for almost everyone. We have been doing competitions on AI and things like that that everyone can use AI. and, of course, recruitment is one of the topics always, and they have done their automations, their use of AI, and so on. Then Yeah, it's something that comes from everyone in the company to be using AI.

TIM: Yeah, that's great, but it's got the support and buy-in for everyone because then I think things can move a lot quicker, and I feel like companies are going to have to move quicker because if not, they'll be left behind pretty quickly. I would have thought I'm interested in one particular use case you mentioned there, which was in deciding which questions to ask the candidate. Could you unpack a little bit more how you're using AI in that situation?

XAVIER: Yeah, as I mentioned before, I created different sets of questions that I say for myself, and I keep updating depending on things that I see online or other people that work with me ask, and so on, but not all the questions are for everyone, and, for example, some candidates are more technical. Background, or they are better on technical things, but maybe on the CV and the SATI I don't know if you know what the SATI is; it is a self-assessment of a great interview document that we ask of all the candidates that gives us more idea of who the candidate is. Then, in this case, with both of them, we are able to identify what the gaps are from the job description to what they show on these documents. Then we can work on it, or it helps me to identify which questions I should be asking depending on what I have seen, and this is one of the top things that I use AI for.

TIM: So you have a philosophy, sorry, of customizing the interview to the candidate as opposed to having the same interview for every candidate. Do you have experience with the pros and cons of that approach of maybe having a standardized approach for every candidate versus tailoring it to each candidate?

XAVIER: Yeah, I think that having the same questions for every candidate is a bit tricky, or it will be a problem over time because not every candidate is the same then. So again I go back: some people will be good to talk to, but maybe they don't have the technical skills, and if you have the same questions, maybe they are able to come around on the answers, but you will not be able to identify other things that you should be asking. Then that's why I think it's good that I can somehow, or it helps me to identify which questions I should be getting from it's not only from the CV or the SATI; it is even from previous interviews I get the feedback from the people that interview with, and then sometimes they will ask me, I'm not sure about this X and Y. and then I will get some questions referring to that to see more in detail what is going on there.

TIM: And you mentioned a couple of times the SATI, so just to unpack this a little bit more, it's like an application form where you're asking specific questions and they're providing more information, like what kinds of extra data points would you get in the SATI?

XAVIER: Yeah, we ask. It's a document that we send to all candidates that they need to fill out. Normally, it is between 15 to 22 pages of information that we get from the candidate. Yeah, it's a lot of work, but also it helps us better understand you, who the candidate is, what they have been working on, and also getting away a bit from just the CV, where nowadays almost all of the CVs look the same. Then it gives more data points for us to understand, and it's a bit of a mix of what your university days were like, what the last three places that you worked were, the information of what you did there in more detail, and also behavioral problems that we put there. When was the last time that you lost your cool? why how you feel in certain scenarios, then you get a bit more of a 360 view of the candidate, and it helps us to better adapt and understand the candidate itself

TIM: And that data set is used I'm assuming in the screening decision of Once you get all the applications, who are the ones who are going to get to that first interview? I assume that's where it's mainly used.

XAVIER: Yes, it's used there, and it's used for me because normally I don't do the first interview; normally I do the second one, and also it's for me to use with all the feedback that I get from before to exactly what I was mentioning to see which from the sati and the things that maybe are not 100% a match or we have some doubts. From the candidate that it allows me to find which questions I should be focusing on and not only questions in which areas I should be focusing more

TIM: And so then to tie it back to AI, you then, I guess, feed in some of this information as an input to a chat GPT, and it helps you then select the questions based on the gaps. Is that how you're using it?

XAVIER: Yes, I put all the feedback that I get from previous interviews from the SATI, from the CV, the job description, and then from all these plus the questions I get exactly what I should be asking.

TIM: nice That's an interesting use case, not one that I've heard yet. What about on the candidate side of things? You mentioned in parsing. So you've got a technical test, and you mentioned, I think, something about candidates potentially cheating on the test. Would you view a candidate using ChatGPT or Claude or something to take a technical test in your view? Is that cheating, or do you think that's fair game? What are your thoughts on that?

XAVIER: Okay, it depends on what you ask for in our case. For example, when we are asking for the tech interview or the tech test, we ask them to not move from the website of the technical test itself. Okay, then if you are using ChatGPT, yes, you are cheating because we are asking you exactly not to do that. But when working on things like that, I'm fine; they should be using ChatGPT. But at the end, the test is about you. It's not a really difficult test; it's just a test to know that you have the basics of the technology. We don't go into a lot of detail and things like that; it's just an easy SQL test, for example. Depending on the level, we need to know that at least some SQL is required; we cannot have candidates that just use ChatGPT for everything. You need to have some technical background, at least a bit of it.

TIM: Yeah, it's such a tricky current situation we're in. I think because my view of it is that because the large language models have improved so quickly and in such a staggering way, now the typical older hiring processes are in a bit of an awkward situation where some of them have been partly broken or completely broken by the fact that now candidates can use a large language model to take a test to change their CV to apply for master different roles. and I feel like we're not quite sure how to combat that at the moment because on one hand I could see how some companies might say we want our candidates to use AI in the job as much as possible because they're going to be as efficient as possible; therefore you can use it in the hiring process as much as you like. but then, as you point out, you run into that scenario where a candidate is using AI for SQL that they don't know at all, and so they're not going to be in a position in the job to really scrutinize and still own their work if they have to rely on something as basic as SQL. Using a large language model So it's such a tricky balancing act, I suspect where we'll end up is some kind of new way of doing hiring in a year that allows candidates to use AI as much as they want because they're going to do it anyway. I don't know, but where do you feel like we're going to get to in the next couple of years in terms of hiring?

XAVIER: Yeah, it's interesting. I'm not sure what the future will be; that's also the magic of the AI at the moment, that we don't know what we can expect for sure in the future, and what we are now is the worst that we will be, and it's already pretty good for some use cases. I'm sure that, yeah, for sure, things will change also. If we reach a point that the gen AI is as good as possible, as needed for creating anything or coding, and coding is not needed anymore, Then, yeah, then I'm sure that we will be looking for other things at the moment. I imagine it will be a mix of both; that's why we are trying to ensure that you have the basics, but we don't care if you are an expert. also depends on the position and things like that, but it's not that important to be an expert. And more on how you adapt and how you are able to learn and use Gen AI or new technologies It's not just that AI is new technology; it's that new technologies are appearing all the time. It's just that now everything comes with AI, and that's why it's also important that you know how to adapt to that.

TIM: Adaptability, willingness, and speed of learning are yes, something that I'm hearing a lot of at the moment, which makes sense if technology is changing so quickly. If you don't have that mindset and that ability, your skillset is going to become outdated pretty quickly. Is that something you try to actively search for in the hiring process? Is that something you're trying to evaluate in the interview or any other step of the process?

XAVIER: Yeah, of course. Yeah, I think that our company is a fast-paced company. We have a new project thing moving really fast, and it's just not for the sake of saying things change every week. There are new things, and adaptability is one of the things that we look for more in the candidates. That's why, for example, yeah, we have a tech stack that we use, but we are not asking for the candidate to use exactly what we have. If we use BigQuery but you come from another tech stack, we are fine with that. It's not about the tech stack; it's about how you show that you are able to adapt and the way that you think. That's why the questions that we try to bring to the table are more about how you think or how you will fix things. how you will work on something on a specific use case and see the way the train of thought And work on that more than, yeah, you will be using these and that technology.

TIM: Are you trying to evaluate or get them to show evidence of their ability to learn new things? Is that part of it, that they have to really show through their experience that they are actually adaptable? Like, how do you differentiate an unadaptable from an adaptable candidate? Do you have any sense of that?

XAVIER: Okay, it comes to different things again. It's with the SATI; it helps us to see the way that they work, and also you have some questions that you can use to see if we ask questions like if you have two teams. One is really fast-paced, but it changed a lot of requests and things to do. and another one is more structured, but you can do whatever you want, and then you see which one they prefer and why and things like that. This is just one example; we have many of them. And you see how they think again, how adaptable they are; there are always some points that you can get from all the conversation, plus everything that we get from the previous feedback. Of course it's not perfect. We hope that recruitment was perfect; we know it's not, but we try to bring as many things and learn from your past mistakes, and whatever is working, we keep doing it.

TIM: Yeah, that makes sense, and if anyone ever invents a way to perfectly predict how quickly and what someone can learn, that would be worth a huge amount of money. I think the closest we came to something like that was because we had the same thought, which we thought was that the number one thing we're really looking for, especially in our engineers, was willingness and ability to learn new things. We weren't so much interested in them already having skills in their exact tech stack as you say, and we also knew that we were going to change what we were going to do pretty rapidly and that they would have to just deal with that, and we just wanted people who could solve new problems they'd never seen before using new technologies. So we thought that it's hard to assess that in an interview, I think, because a lot of interviews are around asking the candidate to talk about a time they've done X or they've demonstrated Y, but talk is cheap; a really well-organized, coached candidate could probably think ahead of time. Oh, I know they're going to ask me about these 10 different themes, so I'll make sure I have my stories ready. and so we tried to say if we could get them to demonstrate their behavior somehow as part of the hiring process, that would be more valuable than just talking, so the best we could come up with was in our take-home test we got our software engineers to create a little algorithm—nothing too complicated but using the language R, like the statistical language, because no software engineers would know that. and we were doing it as a way to see, like, how would they even react to that? Like, some engineers might go, I'm not going to do that. Why would I want to learn? Ah, that's a pointless language. I'm an engineer, so that would be for us a bit of a red flag because it would show that they're not really willing to learn. If they also couldn't learn a new language, they just couldn't figure it out, like they couldn't install it on their machine, or they just couldn't do it; that would also be a problem, and so we found that this seemed to be a reasonable indicator of whether they're willing and able to learn something new as part of the actual process itself rather than just asking them interview questions. but of course then that's quite complicated because you have to come up with a project they have to do a take-home thing like it's a lot of effort, and it's only evaluating one thing, but realistically there's probably more things that we were interested in discussing with the cabinet, but I always find that if there's a way to get them to demonstrate through the actions some kind of behavior that carries a bit more weight in our view

XAVIER: Yeah, I think that was good work. Yeah, I never thought about it, and I think that's a good way to test the candidate to see how willing he is to be out of his comfort zone, for example, and how much he really wants to go ahead and be with you or with the company and see these kinds of things.

TIM: Yeah, I think you're right. It also selects for just engagement and how interested they are in the role in the company you're right. One thing we touched on earlier, and you mentioned, was that the CVs are starting to look more and more like each other. That's something I've heard from several people in the last couple of months: that for any given role, they're getting a higher volume of applications. and the CVs are looking more like each other and tend to be looking more like the job ads, so the CVs are looking better than they were maybe a couple of years ago. How are we going to deal with this problem of just getting flooded with all these CVs? Is this where the SATY, for example, becomes helpful then? because you have those extra data points, like what are your thoughts on that screening stage?

XAVIER: Yeah, I think that in general, something that we cannot avoid is here to stay. Somehow I can see everywhere every job post that we have, yeah, 200 or 300 candidates in one or two days, and it is inevitable now to get that. Yeah, what we do is the technical test; we have a technical test. For everyone, at least it will already remove people that don't have the minimal technical knowledge that we need for the position. And second, in our case, we have the SATI that gives us also people that don't want to expend because it takes time to create the SATI. And to do the SATI, at least I think that when I started in there, it took me four or five days to do the SATI myself, and it gives a lot of information, and then these kinds of things help to remove all these people that, yeah, he's just using AI to get the nice GDI CV close to the JD.

TIM: Yeah, I think that's really the only way to do it, is to have that slightly more validated screening step, as you said, with the technical test, and then with the Sati. What I imagine might then happen, if I just think back to the last five years, is if the market conditions change again, like maybe they would in six to twelve months, where suddenly it becomes a bit more candidate favored. At the moment, it's very employer favored; like every company is getting lots of applications that could switch again. Do you think you would then view that SATI, or that initial step, as quite a lot of work? Would that be something that maybe you might move further down the process because then it's a case of, Well, actually, we need to attract and keep the candidates warm rather than having a filter to screen out the bad ones? It's almost like sometimes the mentality has to shift depending on the market, but what could you imagine happening?

XAVIER: I don't think that in our case we will change that because the sati is part of all culture in the company, and it gives us also more idea of the candidate, and I don't think that it will change much, to be honest. Maybe we will. It will be harder, and we will need to approach more the candidates and explain why we already do that but explain more why we need the sati to understand them better and what is the value of it and why we do it in the process and why it is good that everyone in the company from any level has done this and shows that, yeah, we are a similar cultural company on this regard, and this helps to be on that point.

TIM: So a bit of marketing and a bit of sales of the role in the company should then maybe encourage candidates to go through that first step.

XAVIER: Yeah, and at the end is something that we already do because even if we have a lot of candidates, the candidates that you want, you need to show them that the company is something that is worth fighting for, to say in a way You

TIM: What about then thinking about it from the current candidates perspective? So if I were a candidate right now, I'd be going and looking for jobs on LinkedIn. I'd be seeing the number of applicants and thinking, Oh my God, that is a lot of competition that I'm up against, and I feel like this is probably also what's partly driving the behavior of applying for many more roles because they're seeing there's so much competition for one role. So instead of maybe three years ago, they might say, Oh, I've got a one in 50 chance; now I've got like a one in 500 chance. Maybe I have to apply to 10 times as many roles to have the same probability of getting an offer, and so do you have any general recommendations to candidates for how they might break into data roles or get hired when they're facing these current market conditions?

XAVIER: Yeah, it's a tricky question because it's a tricky moment, but I think that one of the best things is if you think an architect has their projects that they can show. I think it's similar with the data roles nowadays; you should have your own portfolio of projects that you have done and that you can show. more if you are in the starting point that shows the willingness of what you have been working on, the projects that you have been doing, then you have something tangible to show to the people, and these will already make you in the top 5 percent compared to the other people because the other people will have nothing. They just send CVs to send CVs, and you have projects that you can show, and people will be interested to see, and yeah, it shows that you have been working on your skills and have something tangible for the people or even for yourself.

TIM: Yeah, I think that's a great shot for any young grads who've just finished a degree or maybe have done some online courses. There's a very big difference between the study and the work, and so even doing a little project yourself, like, Oh, I don't know, I'm going to download my Netflix or Spotify data and create a dashboard or visualization to show how much I'm using it and connect it to some other data set, not bad, like that's better than nothing because you'll face the real-world challenges of shit data and trying to connect to an API that is erroring, and you know, connecting different data sets together, so yeah, I think there's a lot you could do through something like that. I can even think of an engineer actually we hired who was a fresh grad with no experience. But they built a mobile phone app and had it shipped onto the app store, and I could use it, and it worked. So they can't fake that, and this is pre-Chachapiti. So they replicate themselves, and yeah, that can is something in my hands that I can just see and evaluate and ask them questions. And it shows a level of passion and curiosity and interest that I thought was really impressive. And yeah, it speaks to me. So much louder than just a CV, isn't it? if there's something tangible as you say

XAVIER: Yeah, exactly. The CVS is something that everyone has; the projects are something that only a few will be able to show.

TIM: What about then for not necessarily grads but let's say middle professionals, like a senior data analyst or senior data scientist, someone who's got—who's in the middle of their career—how would you approach the job search at the moment if you were facing these market conditions with such stiff competition? Would you still be applying through job sites and careers pages, or would you take a different approach?

XAVIER: I think Yeah, it's again, it's a difficult moment. I think that going normally, it was better if you had a recruitment agency that can help you because normally a recruitment agency will be always closer to the employees and the employer, sorry, and it will help you to get closer. But again, you need to show something different than other candidates, and yeah, it's difficult times. Then I will go, Yeah, recruitment is for us. For example, we get a good candidate from recruitment. Also, keep your LinkedIn updated with the latest information; make it different than others. That also shows how active you are or what you are doing and that you are not just another AI profile. Also, how active you are on networking blogs or posts is something that will also show the employee that you are a bit different and how willing you are to learn to do things to show what you have been able to do, etc.

TIM: One thing I was thinking about recently was inbound versus outbound, and I think if I were applying for jobs now, I would probably not bother as much with job boards just because I feel like it's so easily going to be lost in the noise no matter how well you craft a CV; even with the projects, sometimes it's not an opportunity to share them in those initial stages. Once you get to the interview, I think it's really valuable, but it's hard in those screening stages. I would probably try to use my networks, and I'd probably approach it like a salesperson would, trying to go through the back door, trying to get a meeting somehow through someone I knew at the company or someone I knew who knew someone at the company, or cold email or cold LinkedIn or something. But I know also hiring managers probably don't want to get bombarded with shitty LinkedIn messages with a generic CV, so from your perspective, is there like a right way to do that direct outreach, something that would resonate with you versus maybe something that's going to go straight to your spam folder?

XAVIER: Yeah, I think it's an issue now because even sending messages is something that everyone is doing. I don't know how many messages I get every week. I cannot answer, and at the same time, the majority of them is something that, yeah, I cannot do anything about. I just—I don't know the candidate, or I don't know the person, and then I cannot do anything on that. The best thing is that even before or while you are working on your position and you are finding your position, you create this networking of people that you can reach when the moment comes. That is not something that you just feel for at that precise moment that you need them. Because then it's difficult that it was like that. It should be more over time, and getting this small, yeah, small or big depends on everyone. Network of people that can help you in the future when you need it

TIM: Yeah, I think that's a great bit of advice, especially for younger candidates who maybe haven't had an opportunity yet to build up this network. It would be tempting to view networking as very transactional. It's Oh, I'm going to go to this event, and I'm going to try to get an interview with the person who's there. Ah, I don't know, not really. As you say, it's more of this long-term investment where you meet people and you build a relationship; you offer value over sometimes years, and then at some point maybe you could help them; they could help you, but I feel like maybe people get into it in the wrong mindset of just thinking, What can I get now? which is not the right way to go, I don't think

XAVIER: Yeah, and also if you don't have this relationship with the other person, it just, yeah, it comes like something cold, and just because you need one or the other, and that's not the way to go. It's better to generate a relation between each other over time, and then whenever he's needed, then he or she will be able to help you. and it will be easy to get help

TIM: Have you ever tried to build some of these networks online, or is this always like an in-person kind of thing for you?

XAVIER: It could be a mix, but normally in person is always a bit better, at least to meet at least once, and then even if you can call them, they can call you from time to see how things are; it should be okay online. It's also possible, but yeah, it is never the same. And again, it can depend on the type of person that you are or the other person is, but always when you meet someone face to face, it's always a bit better.

TIM: Yeah, I think so. We had a few years of COVID, and everything was online, but thankfully we can get out into the real world again. Now we should appreciate that. One thing I wanted to touch on today was something that I've always found curious and been thinking about over the past few years, which is the way we have Recruitment and hiring set up in the world is that we have these kinds of specialized teams who would be the ones responsible for at least the early stages of the hiring process, so the HR, the talent acquisition team, and an external recruitment agency, but a lot of the time, or pretty much all the time, recruiters aren't really experts in the domains that they hire for. So it would be very rare that someone recruiting for software engineering roles or data analyst roles would have been a software engineer or a data analyst themselves, so the more like specialist recruiters as opposed to domain specialists, and the feedback I've gotten from so many hiring managers over the years, is this causes a bit of a problem because there's always this lack of really nuanced understanding where, for example, let's say you're searching for a data analyst. The difference for you between a data analyst and a data scientist is very obvious, but to someone who's not an expert in data, it's probably subtle, and this has caused just so many issues. Especially in the screening stages, the kind of lack of data or technical skills in talent teams is that an issue? Is that a bridge we can maybe start to bridge with AI, like the way that they could almost enable them to have stronger skills? Is this a problem? What are your thoughts on this?

XAVIER: Yeah, I think that nuances will always be there. I also feel that they are getting smaller over time, as you mentioned. Now they can go to any AI model and ask questions themselves, which will help them to get this gap and make it lower. I also think that over time, people that work on these teams get more and more knowledgeable of the specifics of the data and the different data positions. What we do a lot is we sit with them, we do screening together with CVs, and then we explain what we expect in one CV, what is good or bad in the CVs, and then they learn also and understand what we look for. Also, what we do is after some interviews that may be more on the ones that didn't work well, also with the ones that work well, we highlight or we sit and discuss, for example. what is needed or why this candidate was not good enough or what this candidate had that was good Because then they can identify these kinds of things on the CVs or on the SAT for us.

TIM: Okay, so you have this almost iterative approach where there's a feedback loop, and as long as the talent team stays around and they specialize in their domain, then they will naturally learn more about it through time through doing that. Makes sense.

XAVIER: Yeah, exactly, and also we have the technical test that will also help them in case they are not sure. We always say, Okay, send the test; at least we will know if technically they are good. You And after that we can check further because if they pass the test, at least we know that the basics are there, and then we can check further what is going on there.

TIM: I'm interested also in these different stages of the hiring process that you have and which bits are measurable or immeasurable and how you think about that, so let's say for the test, I assume at the end of the day you have some kind of score out of a hundred, and the test bit is, let's say, reasonably objective, maybe not perfectly, but reasonably. What about these other stages? You mentioned the SAT and CV, and then you've got some interviews. Do you also try to score candidates in those other steps? Does it come out as a number, or does it come out as more like qualitative feedback? How does that work?

XAVIER: Yeah, it's a mix of both. The SAT has a score from the recruitment team depending on the behavioral things that they saw, but also they have qualitative bars that they show, for example. We have different parts of the feedback that we say, for example, talk well about himself, talk well about others, and things like that. and then we have some points that we get from all the study or from all the candidates that it's a mix of both; we have an overall score of the study, but also we can see some points that it's interesting to see further or to understand further why they talk about that or that way. In the interview, it's similar. we try It's again in the interview stage; it will be more about what you feel and the answers that they get, but what we try to do, or I try to do in my case, is I get the different questions and the different answers that I get, and I try to keep track of it in my notebooks. And also once we finish all the interviews, we try to meet with the interviewers and discuss what we saw and have a more objective point of view and also different points of view. and then we have feedback that we create with the points of why yes or no and what we saw that was good and bad that we can feed to the candidate and/or the next steps in the process.

TIM: Okay, so you have a round table where you discuss the candidate, and is that an opportunity to almost dig a bit deeper into why you thought a certain way? Oh, I felt the candidate didn't do X. I didn't do Y well, and you almost having to justify that because someone else might disagree And then eventually you'll come to some kind of consensus on who to move forward with. Is that kind of how the process works?

XAVIER: Yeah, it's exactly that. We discuss what we thought about the candidate from different perspectives because we have people from different teams, and we are now bringing what we call the bar raisers—that is, people that interview from not your department—and then you have two or three different ways of seeing it. One is from your department itself that you are hiring; another one is from HR's perspective, and the other one is a third-party user, and then you discuss what each one saw, and then we get to a consensus of if the candidate makes sense or not.

TIM: Use the phrase bar raiser. Can you just elaborate on that a little bit?

XAVIER: Yeah, what we call bar raisers is people that are not directly from your department, but they are good interviewers, and they are good at finding these good candidates that are able to adapt to do their job and things like that, and they are good at doing these questions and identifying these kinds of things. And then we have a pool of bar raisers in the company, and then the idea is that they are involved in other departments hiring and helping in finding these candidates, these good candidates.

TIM: And I'm assuming, by the fact they call bar raises, ultimately they're there to make sure that you hire better people than you have currently. Is that the general concept?

XAVIER: Yeah, exactly. The CVS is something that everyone has; the projects are something that only a few will be able to show. It's not about better people because if you keep getting better and better, you will just get the best one, but it's about finding these eight players or people that have the potential to be an eight player in your team or in the company in general.

TIM: Having these different perspectives, so you've got, as you say, the bar raises, so people from another team, your team, and then HR are all coming at it from very different perspectives; some are domain experts, some aren't. Does that always add value, or is there ever some noise, do you think, in having those wide varieties of perspectives where ultimately the person is going to be, let's say, in your team, and you're the domain expert? I love it. What are your thoughts on blends of perspectives?

XAVIER: I can see it being an issue for some companies; in our case, I think it's really good. Why? Because they come from a different perspective, and they give you a different point of view. Sometimes, as you are focusing on what you need, you don't see it, and they will give you the feedback. In our company, a good thing is that the final answer will be coming from me or the hiring manager. not the recruitment team, not the HR, and you will have more to say only if it's just something that, yeah, it's no, but normally it doesn't happen. It will also be stopped by recruitment. I didn't see it in my case anytime, but that's the thing: I have always the last word to say. but I get all the feedback, and then I can understand better. Yeah, if you are on the fence about some candidates, it will give you the yes or no idea because of the way that they see the candidate itself.

TIM: And in part of the evaluation, sorry, in these interviews, is part of it cultural fit, and if so, how do you define that? Is that measured at all? Is it a purely gut-feel kind of evaluation? Like, how are you guys looking at that?

XAVIER: Of course cultural fit is important; you need to think that As a team or as a company, you work in a way that you need to find people that don't break these dynamics or these things that are working well, or people that we saw sometimes they start on the company and they last one month and they go away because, yeah, they are not used to the way that we work and things like that. Of course sometimes you can use cultural fit, or I'm sure that cultural fit is used just as an excuse of, You know, I don't know why, but I don't want him or her in my company. I'm sure that this happens because, again, it's a bit of the way that you feel with the candidate, but also it's important to know that the people need to match the company culture and the team dynamics. If you don't see that working, you cannot bring someone that will be more issues than problems to the team that a solution you're looking for solutions when you bring new candidates.

TIM: Framing it that way, then, you're looking for someone to fit in with the company culture, the team culture, etc. Is there any risk that that is at odds with hiring for diversity, and I don't mean diversity of people like diversity of thought and different ways of approaching the same problem? If you're looking for people who fit into a certain culture, is there a bit of a trade-off there, do you think?

XAVIER: We always look for people that think differently. I think it's needed because if not, you will not go away from what you are doing now, and the things that don't work will not work because there is no one thinking differently. That's why it's good to bring different ways of thinking, but these don't have to be or are related to the way that you behave with others or the way that you will integrate into the dynamics of the team and the company. I think it's a bit two different things, and you need to bring this diversity and people thinking differently; it's needed to grow as a company, as a group, as a team, etc.

TIM: Someone I was speaking to recently described it to me as thinking that a bit more diversity means you might make decisions slower because there are different perspectives and people might disagree a bit more, but ultimately you'll end up in the right location in the right scenario, and you'll make the right or better decisions because you have that diversity of thought. Is that also what you've seen?

XAVIER: Yeah, I think it's again a trade-off. You cannot go on, and that's also why the cultural feed—you cannot have someone that he thinks away, and it needs to happen this way because then, yeah, it's different ways of thinking, but it's a colliding way of thinking, and then you cannot have that. but it's good that you challenge What is there I need This goes with everything I always ask my team to do when we get requests for a dashboard report: to ask why, then, if the same, it's something that you need to do internally. Why we are doing this makes sense to do it this way; you need to know what the end point is and then work with different ideas on how to get there.

TIM: Xavier If you had a magic wand, the proverbial magic wand, this could be AI; this could be anything, and you could click your fingers and fix the hiring process. How would you do that? What would that be?

XAVIER: Okay, if I can go with the really magic one, the one that does anything, is just get the perfect candidate that will accept the offer, and we'll start working tomorrow, and that's the one, but if I go to a more realistic magic one, I will go for a way that streamlines the process for the candidate and for the recruitment team for both things that things happen fast. The feedback is faster, and the full process is just faster and more clear and more transparent for both parties. And that, yeah, and that this happens as soon as possible for everyone with as much feedback from both sides

TIM: Yeah, I feel like that's a realistic outcome in the next couple of years because I feel like the breakthrough we've had in large language models should facilitate a lot of this quite easily. Like, I can easily imagine each candidate now getting customized feedback at an application stage after an interview because now it can transcribe interviews and score people automatically. So I think if we can just remove enough of the manual bullshit in hiring, we can make it a lot better for everyone, and that might be more unlocked than making it more transparent and fair for everyone.

XAVIER: Yeah, I think that's exactly what will happen. The AI will make things faster. To summarize, to get the feedback and so on, and I'm sure that, yeah, at the end of every interview we will get the points that have been discussed and even some of the feedback already from what has been discussed. And then, yeah, get everything more streamlined, faster, and automated.

TIM: Yep, I am looking forward to that future, which I think is just around the corner. Xavier, it's been a great conversation today. I've really enjoyed it. Thank you so much for sharing all of your insights and thoughts with our audience today.

XAVIER: Thank you to your team for all the great questions and for inviting me to be part of your podcast.