#396 The Intersection of AI, Security, and UX: Hubert Behaghel’s Insights on Digital Trust

[00:00:00]

Mehmet: Hello and welcome back to a new episode of the CTO show with Mehmet. Today I'm very pleased joining me Hubert. Hubert, thank you very much for joining me today on the show. So the way I love to do it is I keep it to my guests to introduce themselves. So tell us a little bit more about [00:01:00] you, your journey and what you're currently up to, and then we can take it from there.

Hubert: Very good. Thank you for having me. It's a great pleasure. So I'm Hubert Behaghel and I'm French. I actually started my career as a consultant doing lots of missions coding, and then we go to quiet and I, um, decided with some of my colleagues to start a new company. And then I met my wife and it wasn't a good time to found the company.

Hubert: So I went with a customer. But that entrepreneurship adventure that started and had to stop early, stuck in my mind. And then I, I did actually what sometime we call in the industry, the engineering pendulum, you know, have been pushed into, uh, management roles, et cetera. And so now, um, I'm, uh, I'm a, I'm a CTO at Verif.

Hubert: Verif is, um, is actually the company that's, Has a mission to make the Internet safer. I can elaborate a little bit on that a bit later. But the reason why Verif made sense to me and why I feel like I'm at the right place is because [00:02:00] Throughout my career, actually, after I stopped doing the pendulum, I, I actually, um, work with very big companies like Amazon, Sky, Marks and Spencer, and a few more.

Hubert: And then actually I decided that scale up was really an interesting space for me because not only I understand the mature model that we need to achieve, uh, but also I need to, um, I need to feel like I'm shaping a technology. I'm shaping a team and, and scale up does it great. And, and very few in this regard is perfect because it brings exactly this combination of challenge to have a mission and to be able to, to, to bring something new.

Hubert: That will really solve problems.

Mehmet: Great. And again, thank you very much for being here today. We have a lot of things to touch on, but I would love to start with, you know, your current role and your current company verified. So, um, so you are into a space, which is, you know, I would say it's top of mind for [00:03:00] everyone, whether you are an individual, you are a business because you are solving something related to identity verification, right?

Mehmet: So what I want to understand from you Hubert first is if we can just, you know, go over the main challenges that, you know, people face when it comes to identity verification and then relate that to, you know, maybe the main core, uh, I would say solutions that Verif offers today. So If you can like enlighten us on on that aspect And

Hubert: can I ask you do you care a lot about the end user experience or the companies that's trying to manage?

Hubert: Uh, let's talk

Mehmet: about let's talk about both and feel free like to to dissect it. No, very

Hubert: good. No, no, that's great So look when um I guess the biggest challenge we have is, is that it's actually surprisingly hard for you to prove who [00:04:00] you are online. And then the other thing that's very interesting is it's very easy online to tamper, uh, and to, uh, fake, uh, And creates this type of identity fraud, impersonation frauds.

Hubert: Um, so, because if we think about what has happened is the internet has, has boomed, uh, and, and keeps, you know, uh, bringing more and more of our lifestyle into the digital space. Um, and so we need to really create a lot of trust because any economy and the digital community, no exception is, is. built on trust, but we didn't have actually the basic trust infrastructure, uh, that's digitalized, right?

Hubert: So what Verif did actually, uh, at the first move is you could almost consider it like we breached them and use the trust infrastructure from the real world, uh, with paper doc, plastic document and made it scalable and useful for, for companies to be able to verify, uh, their end users and, and, and [00:05:00] that's where create trust.

Hubert: And that's, um, that's actually a very hard problem, um, because you need to have expertise around, uh, well, documents first, uh, but also, uh, biometrics, because unfortunately, uh, online, it's, you know, so interesting with the documents we use, the passport, et cetera, they are very sophisticated documents, and it's actually, you know, we could write books about the level of security they bring, but when you upload a photo of it, Like a lot of it is lost right and then suddenly to just swap place on such document You need photoshop or even something much cheaper and you can start being a broadster, right?

Hubert: Um, so so that's the main challenge is how do you create strong definition of identity online? When all you have actually is mechanisms that are fairly easy to tamper. Um, and that's where also, um, one big, uh, weapon you could say that we have to, to, to fight that is AI, of course. [00:06:00] And then what has been so interesting in the recent two years is, uh, we've seen that AI can also be the threat, uh, and then deep fakes, uh, has become even more of a challenge.

Hubert: So now if you step back for a company and you could take, for instance, um, Um, Monzo that we are helping with the KYC in the UK, uh, know your customer, um, their vested interest is to get a highly accurate verification mechanism because what it means for them is that not only they create an environment of safety, but of course they also allow themselves to not incur massive costs on fraud.

Hubert: Which, uh, we know now is growing by 20, 25 percent year over year, right? And the rate of fraud is growing 20 percent year over year. And also when you look at the actual cost of fraud, it's roughly 6 percent of the revenue that are made online. So there's a big, you know, trust tax that is carried by the good actors, um, to, to, to, to make it for this.

Hubert: And so [00:07:00] I think when we put ourselves in Monzo's shoes now, it's very clear. And that there is high value in having early in the onboarding funnel this ability to verify, but then there is a ton of smaller interaction that you also want to create a higher accuracy, a verification on so you know, we get delivered more and more stuff all of us.

Hubert: And, uh, if you have, for instance, age restricted deliveries. Well, here you probably don't want to pull your full passport and get a full verification. What you want is just something around the age which can be done. Well, you can still flash a document and that's something that we support, but also for more easier use cases, you can just, you know, use biometric and create an age estimation that's good enough.

Hubert: And so here, uh, we work with a really cool company called Starship, and they send boats to your place. You can't open the container until you have proven your age, etc. So there is just, and then, you know, the gig economy is so interesting as well, right? Because now [00:08:00] you want to, and that's an amazing opportunity that the digital world gives us.

Hubert: Like, we can have jobs that are either more temporary or just more, um, adapted to our lifestyle. But here again, like trust is important and suddenly it's not just the onboarding that you need to prove. You need to prove as well that the person that's showing up to do the job is the person that you were expecting, right?

Hubert: And suddenly again, like your face is used every day, uh, or, or, or the, or the mechanism, by the way. But right now, the best we have is actually this, right? This, uh, either go full blown with documents and selfie or get with something that's more matrix centric, which is smooth. But also, um, requires new technologies like likeness detection to be able to, to be done, uh, without the problem of deep fake, for instance.

Mehmet: That's very interesting, uh, Hubbard. And you know, like, I have like a couple of very quick follow up questions on this. You mentioned AI. How this made things more challenging, I would say, uh, [00:09:00] for any anti fraud technology in, in, in, in general, and then, you know, coming to the specific use case, which is mainly in KYC and, uh, these kinds of use cases.

Mehmet: So like, did it make it like more complex and, you know, It kind of led us to think again, the technology that has been in place for some time, like what was the effect of the AI, deepfakes mainly, of course.

Hubert: Yeah, no, it's a very good question. And it's interesting because this is a question we get from our customers a lot.

Hubert: And I have to say, initially, my answers was not super convincing because fundamentally, AI is really, deepfakes are really good at, um, you know, Tramping the human eye or even the computer vision, but you can't fight fraud with just one layer. Right? So the whole richness of a solution like verifies it has several layers.

Hubert: And then, um, if you think about GEN AI, I, In a way [00:10:00] for us, it hasn't been new, like being able to craft a very convincing faces or even documents online is something we've faced from the beginning. So one thing, one thing you could say is that it has lowered the barrier to entry to create sophisticated and high quality fakes.

Hubert: Um, but then if you, if you don't understand how you should distribute your attack, if you don't understand the concept of device fingerprinting and all the other layers that we use, Then all that you were doing is actually, yes, we could see that. You could, uh, you know, um, go through one of our layers, which is a computer vision.

Hubert: But because you were caught by the other layers, then actually for us, it was a great opportunity to learn without having any impact on our customers or on the use case. Uh, but it's true that this is generalizing. And now we see more and more sophisticated, um, attack that are not just using the Gen AI to create an image, but also.

Hubert: Cool. So, uh, it's not using AI that much, but the distribution and all the other layers can [00:11:00] be actually managed by specific actors. And that for us, what that means is we, so on one side, you know, we need to be good at either generating or finding data that allow us to refine our, our, our training on AI. So that's clearly has been a very clear demand on our side, but on the other side, I would say.

Hubert: It has also amplified the need for us to have a two, two pronged response, right? One is the ability to quickly, um, I would say operate on, on a band aid if we see that really the attack is going through. Uh, and in seconds, sometimes minutes, verifier is, is going to take some actions. A bit too coercive. And then that allows us also to understand that, okay, now we've, we have, uh, something that can trip us and that creates a roadmap for the, for the, um, for the product team and the data science team to productize these and remove these band aids, uh, that, that works, but may not scale the way [00:12:00] we want or be a bit too coercive again.

Hubert: And, and, and, and so this combination of having to acquire data. And learning super fast from new and new. What it means for me is I need to invest on, you could say on the velocity of very, not necessarily about a specific feature of specific technique. Uh, it's really about how can we create a feedback loop.

Hubert: That allows minimal impact on our customers, even though we know there will always be, you know, the process is always looking and there is always at some point a breach or something, but, but it's this ability to be highly responsive, both at the operational level and at the product level. And it's true data is, is a very interesting, um, Challenge because, uh, right now, so we are an European company, so we work under GDPR and we tend to be data processor, right?

Hubert: But that means we can't keep the data for very long and we can't use it for, for long term training. So there is a very interesting, you could say battle for us to kind of maintain or [00:13:00] generate, uh, high quality data that allows us to always validate we have the top class accuracy that we, we promise to our customers.

Mehmet: The other question I wanted to ask you about, you know, a little bit, it's a kind of, uh, you know, from, from your perspective as a CTO and someone with, with a long experience. So you are in a space where you need to work on protecting the client and at the same time, Of course, and the business, like it's like two, two side, uh, uh, picture here at the same time from UX perspective, you know, like what usually even I come from a technical background when I have like this friction in verifications and all this.

Mehmet: So I feel like, you know, So what have you seen, like, really something interesting that, um, can be done [00:14:00] to keep, you know, the security aspect, you know, the identity protection, you know, the fraud protection perspective while keeping the things smooth, you know, what, I know, like, it's like kind of stuffed question, but like, is there a secret sauce for, you know, combining these two together?

Hubert: It's a very good question. And, um, it's a tricky thing because you're right on one side. You could say the thing we have to do is to increase maximum friction for the bad actors. But ideally with no friction at all, right? The ideal solution with no friction at all for the good actor, right? So very uses, uh, I mean, it's true that what we call our end user flow is, uh, is, um, is a space of a lot of attention.

Hubert: Um, One thing we do, I think we do really well is for instance, we minimize the amount of steps. So one thing that's a bit special about Verif is we, for instance, we don't ask you to tell us, is it a driving license? Is it a passport? You know, our [00:15:00] AI is going to actually figure that out for us as well. So on one side, just trying to get to the super minimum, which is like, you take a selfie, you upload your whatever document.

Hubert: And then we just, you know, have a bit of a throttle and we tell you, that's it, you passed and that that's what, uh, that's what we try to maintain in every scenario, but it's not always possible. For instance, when you work for the mobility industry, well, then they need sometimes several documents because it's not just the identity.

Hubert: It's a driving license. And the other thing is, um, there is a, uh, A concept of, um, uh, that's, that's a concept we're working on. That's a bit new is that you can also try to understand what is, I would say the motivation of the user for going through, um, and, and take that into account in how you structure the flow, because.

Hubert: I don't know if it's, by instance, we, we work also a lot with the, the HR, uh, industry. And if you onboard it because you're going to get a job or [00:16:00] going to get an interview with a company, then you, your motivation is quite high. So here, depending on how much confidence you try to build, like we can structure a flow where it's okay to take a bit more steps because the motivation is high and the quality will be higher down the line.

Hubert: But if it's a scenario where actually the motivation is really low. Then you need to accept that here there is a trade off you need to make consciously. So that could be failing in favor of the customer, for instance, or that could be just a, so, you know, We have a scoring mechanism and we have some labels we share with our customers on our decisions And then it could be also calibrating the threshold With them, etc.

Hubert: But this tension between conversion And, uh, you know, miss fraud rate, which is the gold standard of the North Star of any fraud system. It's always to be renegotiated. And more and more, we believe the concept of use case needs to come to the fore. It's not even the industry. It's not even the customer.

Hubert: But the customer is important as well, right? For instance, in this flow, for instance, if you see very flow go [00:17:00] and you weren't prepared for it, you Like it's a problem as well, like where I am now, who I am talking to, right? So there is a lot of detail that this flow looks for super simple, and I hope it is for most of our customers.

Hubert: But you're right to say, like, there is a density of problem solving here and the balance that's again, always to be renegotiated.

Mehmet: Absolutely. And now this brings me to, you know, to another, I would say, topic, which I want to discuss with you, because Hubert, what you're describing is exactly, you know, what, uh, Solving a Problem from technology aspect.

Mehmet: And I want to ask you about, you know, this, uh, concept of, you know, the long term business vision versus, you know, the goals of the business. And I'm asking you from a CTO perspective here, because, you know, taking, you know, like just, What you solve today at Verif is maybe one aspect of it, right? So, so from, from a customer perspective, but what I'm, I'm, I'm, you know, like [00:18:00] curious to know from you is what have you seen the biggest challenges when it comes to aligning, you know, this business vision with, with the, maybe the, uh, the, the, Goals that we set for for our teams.

Mehmet: I mean from technology perspective

Hubert: It's a really good question and actually I have to say verif has given me the best playground to take that to the next level because Actually, we are not the only one at Verif to have this dream, but one goal post, if you want, of identity online is, is going to be reusable identity, which means in effect, you don't treat, so Verif right now is a highly transient platform, right?

Hubert: We don't retain anything like you come, you do your session and we forget about you again for data privacy consideration, which are very valid. But we, the, the, the, the longterm vision is to be a partner with our customers of a, of a higher level where we could be data controller. And what that [00:19:00] allow us to do is then we can remember, uh, you, and we can remember a different level of data.

Hubert: That means suddenly, I don't know, if you open a bank account with this partner, then you go on this, uh, social network, and that also works with both. We can be much, actually, we can even do, you know, one of the grail would be a bit like, are you a human, you know, the, the Google recapture thing that you could just get on the page, it would load a bit.

Hubert: So, well, actually, I know you are recognized as your device, et cetera. So just go ahead. Um, now, so that's, that's long term, as I told you, like there is a technical challenge. If I'm honest, it's not the biggest, but there is a level of partnership, uh, how we can position both. Now, what's so interesting is to achieve this grail.

Hubert: We need a lot of foundations. Um, and, and so verif just launched, uh, it's biometric authentication solution this summer, biometric authentication, I think everyone understands what it's meant to be doing, but really what's, [00:20:00] uh, what's important is, um, is the concept of. Of, uh, of liveness of, of biometrics itself and how you can plug it into different solutions.

Hubert: And, and so we have three portfolios right now, the IDV, as you said, the biometric solutions, and then, uh, what we call data product, product expansion, uh, which is trying to answer the question, like, can you be trusted? Right? It's not just, are you, who you say you are is, can you be trusted? And, um, and, and, and the, um, what's so interesting is all of these Needs to play their part to build, uh, toward the vision.

Hubert: So they all have their own lab. They all have the intermediate feature and product, but the real thing they are trying to achieve together is, uh, the ability to deliver a reasonable identity platform for, for the internet. And, and then underpinning it, uh, which is also very interesting. Is this concept of product platform?

Hubert: Uh, and this is where now I'm going to be, uh, talking much more like, um, hardcore [00:21:00] CTO. Um, when, when you, when you solve the problem that Draper has solved and you realize that the internet needs, you know, more sophistication and be able to plug solution in different part of the life cycle, you, you realize quickly you need to go for a multi product, um, uh, strategy.

Hubert: And then suddenly when you want to become a multi product company, it's It's very easy to miss the turn in a way. And so we have formed last year, um, an area in my org, which is called product platform and product platform in a way you could say is here to breach the proper platform engineering with, uh, the cloud settlement, the infrastructure, the CICD, all this very foundational stuff for doing engineering.

Hubert: The product platform is kind of sitting on top of it. And he's, he's here to amplify this platform strategy, which, you know, has things to say about velocity efficiency, making sure that the passive list resistance for engineers and product teams [00:22:00] is also the part that maximize, um, doing the right thing in terms of customer experience in terms of reliability, et cetera.

Hubert: But here with the product platform, we don't treat the. engineer are the, as the customer, the persona, like in platform engineering, which we, the entire product team, right. And, and so we have several product teams and this product platform is here to. really understand what is, uh, the right level, what is the right environment that's going to allow them to focus, to, to, to, to play this velocity game I mentioned before.

Hubert: And so one, one thing that's emerging from the strategy of the product platform is a concept of orchestration or workflow, which actually is going to become a product for Verve. Um, but right now we actually don't even focus that much about the end user, even though we have a vision for it. We're trying to say, look, if all these portfolios have all parts that needs to be assembled into this reusable identity solution, then, you know, let's, [00:23:00] let's already think about the foundation and let's make sure that not only we get fast there by having the right focus, but also we get this ability to reassemble.

Hubert: Part of our solutions into new solutions. I don't know if it's helpful, but that's that's really a the way we attack that.

Mehmet: No, absolutely. Now, a couple of, you know, things that also I know that can happen, especially when you need this velocity that you talked about. And, you know, especially we mentioned maybe AI.

Mehmet: And of course, AI is not new by itself. But I mean, you know, the Fast development of AI also was very fast. So from your experience and, you know, from from your current position also, how do you prioritize or decide, for example, that, okay, there is this cool technology, there's like maybe technologies also, like, More than one, uh, or maybe it's a practice.

Mehmet: So how do you decide like, okay, [00:24:00] this is a good one to adapt, right. And we should implement it, but at the same time you have to think, okay, about again, the, the goals and you know, like, uh, the needs that the company needs, right? So how do you make this a balance in, in deciding which emerging technologies or practices to get on board?

Hubert: Yeah, that's a very good question. So there is, um, There is first, you know, being very deliberate on where you want to be at the forefront of the technology. So for brief, it's about biometrics. It's about documents. Including digital documents. So we know very precisely what expertise we want to drive forward.

Hubert: Um, and, and then, um, there is, um, so when a technology emerged, um, we actually don't do it exactly the way I want yet, but, uh, there is this concept of radar, right, uh, [00:25:00] that, uh, we have, so there is, uh, some documentation when we can record, or even we just, uh, have, uh, an injuring form on a weekly basis where, you know, we can showcase.

Hubert: Things that are actually popping on anyone's radar and trying to first, uh, create awareness. Um, now when it's time to decide, okay, now we actually want to, to try it, that can be very quick in the sense of if we have, for instance, you know, a new technology on document authenticity, right? And, and we know we, we sometimes, uh, see challenges and we think they could help us with some problems that are not systematic, but could become a problem.

Hubert: Then you, we don't hesitate. And there is a bit of bandwidth that, uh, any team is maintaining to just jump on it. But if it's a bigger, um, uh, decision and a bigger investment upfront, then I'd very, actually, I've, I've, uh, championed, uh, a betting framework. That's not just for new technologies used for each time we want to place a bet.

Hubert: Uh, I call it [00:26:00] TRAP, uh, TRAP, uh, stands for, so there is, uh, time, results, and then the A is appetite and the P is potential. So the three, these four, uh, constructs, uh, what they are trying to say is, look, uh, we use it a lot more by the way on, uh, you know, one thing that's so interesting at Verif is the SaaS enterprise.

Hubert: So we work with Fortune 100 and big customers that tend to expect from us a lot of engagement. And at some point I had an issue where, uh, it was hard for me to understand where to draw a line of, you know, how much, how many calories can I put on this customer versus the bigger picture, et cetera. Right.

Hubert: And so framework also helped us with this, which is to say, okay, You have an idea, uh, and the P of trap is going to capture the potential. So we can agree, say in one year or two years, this could be, I don't know, 2 million, uh, uh, run rate, uh, for life. So, okay. So that position already, the initiative, same for the technology.

Hubert: It doesn't have to be [00:27:00] revenue. It can be, you know, if we have a concept of balanced scorecard, which are like the 16 metrics that the business is optimizing for. So if, if you can justify a potential for optimizing moving the needle by that much on that metric. That as well can, uh, can, uh, be translated into a score and then, you know, you have your potential.

Hubert: But then the problem with the potential is, uh, it creates a long term view that's very appealing, but you don't get a chance to understand how you're going to fail fast, for instance, right? And you don't, don't even understand again, how many calories you want to put in. So the A of trap appetite is the How much we're happy to waste even on this like it may not pan out, but we want to give it a go What is our appetite and actually I found that for engineers Uh, if you actually tell them this is a prime, this is a requirement, or, but also this is how much we think we should spend on it, I guess a super useful input to kind of calibrate the sophistication, the, the amount of, um, of efforts we want to put in it, and then the, the, the, the [00:28:00] architecture of the solution that should come with it.

Hubert: And then, you know, okay, so now we know how much we are happy to waste and we, we know how much we could make. We still need a line in the sand that tell us how to fail fast, right? So time is, and results, I hate to say by that time. And usually we want a horizon that's less than six months. We want to have materialized that much result.

Hubert: It could be on the potential, the same metric as potential. It could be something much more of a leading indicator. And so once you have packaged, you know, your trap, your, your bet like this. Then, you know, based on historical data or prioritization framework, we can decide, look, this is worth going ahead or no, sorry, potential is not big enough.

Hubert: Um, and then, you know, this, this also helps me if I look at all the traps we have in flight, it also helps me manage actually these calories again of how much I want to put on risky bets versus how much I want to put on on safe outcomes. Right? But that's one thing that has helped me. Helped us a lot to kind of allow us to be [00:29:00] deliberate on what you were saying, like when to try something new and to have a betting framework and be clear on the expertise you're trying to lead forward.

Mehmet: Absolutely. Yeah. I agree with you, uh, Hubert. Now you mentioned engineers and you mentioned, you know, like fail fast. Um, so let's shift to, to, to a little bit, uh, leadership perspective. And when we talk leadership usually, so we talk about building culture, right now from What you have seen successful at Verif and maybe from previous experience also like how do you build a culture where people can have this, you know, ability to fail fast to keep trying at the same time, you know, foster the continuous learning, um, and, you know, be eager for innovation within the engineering team like, um, And especially, you know, like now, I'm not sure how is it at Verif, but you know, as I spoke to a lot of, uh, [00:30:00] um, CTOs and CPOs and VP of engineering and all of them, like now they're telling me they are either like a hybrid mode, you know, so You know, people can work remotely.

Mehmet: Even sometimes they have people shifted across, you know, multiple time zones. So how you build this culture, um, Hubert?

Hubert: Yeah, that's a good question. For me, that the, the, the foundation of a leadership that enabled that in a fine in injuring has to be built on authenticity, which I like even more adverse because Authenticity is exactly what the internet is kind of removing from our life.

Hubert: Like I can't be too sure that I'm talking to you right now. Maybe you're just a, uh, engineered video. Um, that's a really good at, um, on the spot improvising text. Um, so, so authenticity, why also, I don't have a great explanation, but the best I could come up with is, Engineers are the people who are building this modern [00:31:00] world, right?

Hubert: They know how it works, so they can, and sorry for the word, but they can smell bullshit like 10 miles away, and you can't afford to kind of play with that with them because that's already a, you know, that's shut down trust. And what you get out of this authenticity in your leadership It then actually challenges, but also ideas are going to be to, to get a voice, right?

Hubert: And, and that's, that's exactly what you want. Um, there is this first element of how to create the environment that makes it safe and build the trust for us to actually start challenging each other. We own better solutions or alternative way that could turn out to be better. There is something that, um, Is also, um, you it's very interesting that you mentioned the remote or hybrid environment, which is actually something that birth has been through as a transformation.

Hubert: And I think it's going to be a even more of a challenge of future of the company growth. So, to keep this mindset of [00:32:00] any eagerness to innovate, I find it's also another thing that's so interesting is when you speak to engineers or product engineering teams more generally. I think, you know, one, one prop is to talk about change.

Hubert: Um, you know, for instance, when we do a real, we talk about change when there is a new priority and we talk about change. And then we talk about change fatigue, which is a bit normal in, um, in a scale up like verif. But if you think about it, it's actually not okay because people who join verif and companies like verif, they actually want this change.

Hubert: They want this agility, right? So when you put them into a place where it's changed fatigue, It's probably because you're doing something wrong at the leadership level because these guys are actually here for that So and it you could say it's a communication Program and most likely it is what I find a good way of engaging here is to talk about Problems and problem more interesting problem.

Hubert: We have identified or problems that are going to help more of our customers, etc [00:33:00] Because then that that create a mindset where actually these teams are well, I joined for problems and that that's that's what I do Uh, so it's it's small but actually it changes a lot and when I was vp or even director before It was my product counterpart that was always clear that when we I want to be at the forefront because I know how to sell it and it makes a huge difference.

Hubert: Uh, the last thing I would like to say, um, and actually I'm still learning a lot about this, um, is of course to create space. So, you know, the, the, the hackathon, uh, culture is very important. And I'd very, if we have, uh, you know, some work to do, but in previous companies, I've seen companies doing two weeks of hackathon every quarter.

Hubert: And that actually was with kind of creating a great product out of it somehow. So, you know, you need to decide a bit how much you want to index on it and how much it's relevant. I think the more the, the, it depends on a lot of things, but, um, the other thing that I'm learning, and I think is very important is that for instance, [00:34:00] data science doesn't have the same pace and doesn't have the same relationship towards.

Hubert: innovation than more classic product engineering, say, for web app, right? Um, and I think it's also important to respect this space because, um, if you try to put everyone on the same, you know, approach every practice on the same page. Space, et cetera, in relation to innovation, you're going to not get the outcome.

Hubert: Right. And so I think also understanding the space and understanding how to create the space and write timeline for it is super important. And we find that at Verif that data science. It needs several iterations, it needs to go into its bubble, then it starts having a view and then only you can ask for a timeline and an execution plan.

Hubert: But it's much longer at the beginning of the cycle.

Mehmet: Absolutely right. Now, again, something, um, you should touch base a little bit on, Hubert, like, So one of the things, because I like to cover this topic, [00:35:00] I'm a more, uh, you know, um, I would say leaning towards startups and, you know, trying to get someone like yourself to give them some hints.

Mehmet: So every startup and you, because you were mentioning scaling and you were mentioning like, you know, The company starts to grow. What are the things for fellow, maybe CTOs or maybe even a VP of engineering that you can give on this, uh, road to, to scale? Because the road to scale is not like today. They are, for example, a startup and tomorrow they are a scale up, right?

Mehmet: So it's, it's not like just a switch, right? So there's something that happens, but from technology perspective, and maybe even from leadership perspective, there are like some, some, you know, Some stuff they need to avoid. So, because what they have to do, maybe we will need like episodes about it, but like major things, major thing that they need to keep in mind while doing this journey for going out from being just that small startup that maybe no one [00:36:00] still have heard about them going mainstream and becoming, you know, that company, that enduring company, which is now growing, getting access to more customers and maybe geographies.

Mehmet: So. What should they avoid? I would say,

Hubert: what should they avoid? Um, great question. Um, I actually tried to focus on, um, on one specific challenge that I found fascinating at Verif. So when you start in the early days of, of the startup, you have a concept of accountability. That's a very bottom up. And, um, and I, and I, and I, and I think on on one side, uh, What you need to realize it's, it's that that needs to be natural.

Hubert: That's exactly what it should be. Of course, you need to create a top down direction that create clarity, but the truth is it's only going to, to, to work and to have the right level of energy. If you have the right accountabilities at bottom up, but then [00:37:00] what you should avoid doing is assume that this concept of accountability is going to stay that much bottom up.

Hubert: And that people as well, and maybe even it's actually probably the harshest part of a startup to scale up I find is that the same people who enjoy a very bottom up accountability are going to be able to embrace a concept which is not top down at what I find we need to shoot for is a concept of accountability where Now it's, it's a team that's more taking the time to do that.

Hubert: And for instance, we have a set of priorities and we all stick to these priorities, right? There is no one that comes and say, well, I care a lot about this. All the price is not on the list. So I'm going to progress that, right? And so I think the, the, if you don't actually manage as part of the story, well, And, and understand that at some point, you know, we either need to pull more resource and be a bit, uh, you know, like we had with the last two years, like a tighter ship, et cetera.

Hubert: So we need to be more joined up on what we do. If you can't [00:38:00] help people take on this journey, that now the concept of accountability and the concept of team are going to take very different shape. Then the problem is you're going to have your top performers are going to become Even an anomaly in your organization and I I want to believe that most of the time actually we can take them on journey Some some people are really startupers and we should respect that as well And so that means the way we take decisions, the way we bring a concept of prioritization, the way we also refresh our goals, et cetera, they, they need to illustrate and even tell the story very frankly, like again, this is a synthesis, like, you know, we're changing a bit the way we do things and we want, All of us to be still as accountable as a total, but you know The way we are running this topic or that topic now needs to be In coordination with this wider team, etc So that's the one thing I would like to say is like Don't be naive about the fact that accountability takes a very different shape along the journey

Mehmet: Yeah, I agree with you ever because I think [00:39:00] uh, you kept Mentioning, uh, authenticity, which is, you know, something I, I speak about it a lot because, you know, like this is, and this is a race to communication, I believe.

Mehmet: Right. So, um, when, when you communicate everything in a transparent way and, you know, you bring everyone up to speed, so things becomes much easy because the whole team understand what you're trying to do. So they don't get. You know, because I've seen examples and I've read about like some use cases where the team, they were left blind about what's happening.

Mehmet: And, you know, when, when they get to know, Oh, like the company now is shifting to this direction, they get upset. And you know what happens? Like people start to leave and, you know, Problems start to happen. So, so yeah, this is resonated with a lot of stories that I've read about. And even I've heard it from, from a couple of people of founders and they were complaining.

Mehmet: I said, yeah, like have exact, I asked the question exactly the same way where you were explaining, [00:40:00] like, have you been transparent and, you know, authentic enough with your team and said, no, you know what, but like we needed to hide things and you were like there, but I've seen some people, they do, they think Let's do it first, you know, and then we'll let the team know about it and they fall in the trap that maybe something will go wrong so they don't want to upset the team and you know, but actually they They they they make the whole thing fails from from from the beginning But nevertheless now final thing final question for you, Hubert Because we are about to, to, to wrap this up.

Mehmet: How was the experience from shifting industries? I would say for you, because you know, you mentioned in your intro, like you work different places. So I know like, like you were also like, uh, going up in the ladder of, of your career at the same time. At the end of the day, again, that's a shift. So from Amazon to Sky to now Verif.

Mehmet: So this, like how much it enriched also your, your perspective as a CTO, [00:41:00] because you need to understand the business. And of course, you know, uh, take the company as a, with a vision for the technology, how, how was that, uh, experience with, with shifting industries?

Hubert: Yeah, that's a great question. And I'll be honest, initially, I, I just felt if I'm honest that what we do as software engineer is we have our technical skills, but we always apply them to a domain.

Hubert: And the truth is, then that means you need to become good at. you know, getting under the skin of that domain. And now whether it's private banking, like I started, or, you know, the media industry with Sky or the retail and e commerce industry with Marks and Spencer, what was interesting is each of these industries, when they saw me coming, there was always something to tell me, like, you, you're not from the industry.

Hubert: Right. So I think it was the strongest with retail. Like I wasn't a retailer, so maybe I wouldn't understand a lot of their challenges. And I never felt that actually it was a struggle for me to get under the skin of these industries [00:42:00] on to the point where even if I if I hope it's okay to say, but my experience is that it also doesn't really matter that you're in a very big company or small company.

Hubert: in terms of what challenges you have, right? Um, surprisingly, uh, velocity and injuring was a challenge, uh, uh, for, for Marks and Spencer. It's, it was a challenge for Variff. And it's the way you attack it. I would say the thing that has been the biggest and the most fascinating type of challenge is shifting companies or industries.

Hubert: Has mostly be around culture, right? For me, I had my big moment when I joined Amazon, first American company I was working for. Uh, and then actually, to be honest, Verif was also quite kind of cool because, um, it's a, it's an Estonian, uh, Baltic culture. And initially, if I'm completely honest with you, I was struggling with the fact that am I engaging enough?

Hubert: Is it a disengaged team? Is it more introverted? And I actually needed the help of a book I have behind me called the culture map to, to help me, uh, So I think culture is, um, is a much more important technologist is a much more [00:43:00] important, uh, dimension to, to acknowledge. Now, as a CTO, what I really enjoy is having been across these countries, across these industries.

Hubert: My ability to empathize and to connect quickly with the CTO on the other side of the customer is, uh, is great. And, uh, I have to say here, it doesn't take me a lot to always find some common ground that allows us to break the ice. And, and then, uh, and usually when I'm engaged, it could be because there is an escalation or something.

Hubert: So you need actually this human connection. And so I think that's where, uh, this diversity of experiences is helping me the most today. Well, otherwise it has just reinforced to me that. You know, the same approach, the same technologist mindset, but really at the service of a domain on a new set of customers is what we do.

Hubert: And I mean, it doesn't matter too much where you apply it.

Mehmet: Absolutely. Yeah. And I love this, because, you know, I used to bring, um, You know when you bring to someone a peer I mean from customer perspective [00:44:00] They like it Right because first they say okay. You're bringing me someone who talks the same language as me Uh, and of course as an executive also as well.

Mehmet: So they feel like they are kind of You know treated well by the company. So it always works like I've always seen it Working, um, you know, like really I enjoyed the conversation Like great insights from you any final things you want to mention and where people can Find more about you and about Zwerf.

Hubert: Thank you. Yeah, I would say The final thing I want to say is take seriously your digital life Uh, you know, you you have certainly people in your family that may be more elderly And, uh, I exposed much more than they realized to risks. And also the way you choose your partners, your bank online, et cetera, like pay attention a bit on how much they care about just sticking the boxes of some regulation or actually foster, uh, security.

Hubert: You can find a lot on brave. com of course. And you can find me, I mean, with my [00:45:00] very complicated, the last name on, on, on Twitter or X, as we say nowadays. Uh, and I have a blog, uh, blog. behagel. com. org, um, um, which is should get a lot more love, but yeah, I think X is the best way to reach out to me or LinkedIn, of course, uh, for every professional interaction.

Hubert: Thank you very much, Mehmet. I really enjoyed it and thank you for the great questions.

Mehmet: Thank you very much, Hubert. So this is how just for the audience, you will find the links in the show notes. So if you are listening, uh, through any of your favorite podcasting application, you can find the, the, in the show notes.

Mehmet: If you are watching this on YouTube, you will see it in the description again. Thank you very much, Hubert. for being with me here today. I really enjoyed the conversation. And this is how I usually just take small close from my side. This is for the audience. If you just discovered this podcast by luck, thank you for passing by.

Mehmet: I hope you enjoyed it. If you did so, please subscribe. As I said, you are available on all podcasting platforms and on YouTube and share it with your friends and colleagues. And if you are one of the [00:46:00] people who keeps coming back and send me their suggestions, comments, please keep doing so. I really enjoy reading all of you.

Mehmet: All of your feedbacks. Thank you very much for tuning in and we'll be again, uh, very soon. Thank you. Bye.