Ep. 29 | A Human-Centric Path To AI In Healthcare with Partenamut's CDO Bart Vandenreijt

We explore how AI can make healthcare feel more human, not less. Bart Vandenreijt shares concrete wins in service operations, a culture playbook for change, and a 2030 shift from insurance to assurance at Partenamut.

• psychology background powering data-driven decisions
• lateral thinking as a driver for innovation
• telecom, retail, healthcare differences and constraints
• human-centric values: simplify with empathy, digitize to humanize
• Dennis persona and broken care journeys
• building FOCA: voice transcripts to knowledge
• email triage, knowledge base, faster onboarding
• computer vision for claims automation
• legacy tech debt and stepwise scaling
• team growth from 6 to 70 with clear purpose
• 2030 vision: smart health data ecosystems

If you're interested in knowing more about Movify, don't hesitate to visit our website on movify.com. Stay tuned for the next episode and don't forget to follow us on LinkedIn and Instagram

Chapters

• 0:00 Intro
• 2:04 From Psychology To Data And AI
• 4:26 Relationship between people and data
• 8:32 How do you compare Healthcare, Telecom, Retail industries?
• 11:31 How is AI aligned with Partenamut's values
• 14:05 Partenamut's vision. Peer question #1: Giovanni Briganti (Professor - Psychiatry, Computational Medicine, Mons)
• 17:35 Partenamut's positioning compared to the competition
• 18:38 Peer Question #2 - Kristof Meganck (CDO of BNP Paribas Fortis)
• 20:30 Success story: FOKIA
• 24:00 FOKIA: Voice-To-Knowledge Breakthrough
• 26:35 Dealing with complex automation such as KYC, Fraud, AML, etc
• 30:20 Partenamut's culture and team changes
• 34:52 Main driver that pushes Partenamut forward
• 40:16 The Future of healthcare in 2030
• 42:24 Fire drill questions
• 44:56 Closing notes from Louis & Isabelle
• 45:33 Outro

Transcript

Isabelle: 0:00

Hi everyone and welcome to a brand new episode of Flux. This podcast is hosted by myself, Isabel Bamber, marketing manager, partner and designer at MoviFi.

Louis: 0:10

And myself, Louis Cornet. I'm the founder and CEO of MoviFi. Today we're joined by someone who lives and breathes digital transformation, Bart Van der Eijt. You're the chief digital and AI transformation officer at Barthélamut with a career that spans healthcare, retail and telecom. Always at the crossroads of people, data and innovation. Bacht isn't just a strategist or technologist, he's also an entrepreneur who builds team, reshapes business models, and turns the disruption of AI into real human-centered impact.

Movify: 0:49

You're listening to Flux, a podcast about design and development.

Bart: 1:04

Thanks, thanks for having me.

Isabelle: 1:06

To kick off this podcast, I would love to jump straight in and ask you what fuels your fire and drives you out of bed every morning.

Bart: 1:14

Well, uh, that's an easy one. Passion, passion, and people. If I'm not passionate about something, I will not get up in the morning. And people is like I said, if I can talk with somebody who's passionate and proud on sharing his experiences, even if it's just postal stamps, I really would like to understand what drives them. But of course, I'm not working with people that are collecting postal stamps, but I'm working with data scientists with PhDs that present to me things that I don't have a clue. But if they start talking about it and explain to me what it is that where they're proud of, that that's what makes me tick. So actually, my best day, my best moment in the day is when I get to go to my team, say hello, and then they show me what they're doing and and are very proud of it.

Isabelle: 1:56

That's human connectivity.

Bart: 1:57

Yeah, yeah, yeah. It's people and passion. That's uh and I'm also very passionate about what we're doing. I so uh making a difference.

Isabelle: 2:04

Now that we know your your passion and drive, can you tell us a little bit more about yourself, your background? Is there anything else we need to know about you? Something quirky, maybe anything?

Bart: 2:14

Quirky, definitely a geek, and I'm proud about it. We rule the world behind our computers, but nobody knows it, except now Elon Musk and I started to get a little bit too much here. But no, in in general, I'm uh I'm a proud father of two kids, Lucas and Mateo, 10 and 12. So they're pre-adolescent, so it's yeah, yeah, it's it's intensive, but I I love it. I mean, I'm married, so I'm not with a family, I'm not working. What I love doing is actually uh I'm I started peddling. Paddle, you say pedal, yeah. What's your professional background?

Louis: 2:46

What did you do before becoming chief AI and data?

Bart: 2:50

I'm I'm having I'm a little bit atyp, what they call I'm I'm a psychologist and I'm passionate about psychology. I I loved it. I studied for seven years. I'm specialized in clinical sports psychology, organizational occupational psychology, shopper psychology, and and what drove me and also throughout my my career is uh the realization at a certain point. Well, psychology is interviews, like we're doing here. It's been a long time I had some interviews doing myself. But what you have is that you realize that people are not always saying what they're doing. Yeah. And so I discovered when I started working for I for the big three uh telecom operators where I started testing mobile data on mobile phones, pilot. I discovered that people are actually not saying what they're doing. And so when you started interviewing them on those new technology, well, we discovered that they were on porn sites all the time. But they actually said during the interview that the that the uh that the technology was rubbish. Yeah, there's something that that's that's not adding up, so you cannot confront them with them. Of course, it was before privacy, etc. We're talking about well, I was well, almost well, 25 years ago. But I said, okay, there's something in there, there's something with data you can actually retrieve. I understand way better what's driving people, what are doing people, then then just talk to them. And so that's the reason why I started becoming more and more interested in data and big data. And yeah, I was lucky to develop a whole career on it for 25, 27 years now, starting with MobiStar, Orange today, and uh then retail and afterwards now Partinamit.

Louis: 4:26

And how did your psychology background help you better understand people or the relationship between data and people?

Bart: 4:32

I like the book of Christians, he's talked about disruptive innovation, but the book goes a little bit about how that organizations manage disruptive innovation. But if you translate it from a person point of view, we're not talking about AI agents, it's going to change everything. I would now we have a computer in front of us. Tomorrow we just have a series that's really going to be intelligent or okay, or an hive Google. So we're no longer going to have screens like we have today. It's going to change completely. But how do you cope with that? That's the psychology behind it. So each time if you have something new that you're doing, that we're doing either with data or with AI, trying to find what's the value, what's the the the the diff the small thing that makes people say, okay, well, this I'm going to use. There is something I'm I'm I'm willing to to change a little bit my my code, my my mindset, and also what I'm used to be doing. And I'm a little bit open to that. And try to discover that. That's what I love. And that's that's definitely the psychology that's still behind it. And then, of course, all the technology that gets there. Because if you look at it today, you're afraid. But then if you say, Okay, actually it's going to help me a lot. Oh, shut damn then. Those meeting transcripts, I don't need to do them myself anymore. I've got my smart recorder now, it's doing it for me. And by the way, I don't need to repeat myself three times or four times. It's doing that. Oh, well, for me, I have everything at my hand.

Isabelle: 5:51

Yeah. So you're talking about a bit about the psychology in a forecasting way, like seeing how the future is going on.

Bart: 5:58

But also today, also today is that if you're confronted with I one of the things that also what you see is that also the the the the what do you call what you have when you're young, when when you're 15 or adolescent, if you have a new uh what do you call new operating system for your computer or your or you have a new Android or new AOS or an Apple, you're excited because you want to have the new stuff more. From the moment that you start to get older, you say, Well, what again? I was used to do it, and they already changed it. And what you see is that technology is accelerating at a at a pace that's that's that's incredible. Even even uh us data geeks or starting geeks are starting to get some difficulties on it because it goes so fast. How do you manage that? That is today. That's not tomorrow. If you talk about tomorrow, you make people afraid, but you need to have a vision, you need to tell them where we're heading. Well, what does it mean for me today? And how can it make it change for you today? And what are the expectations that are realistic that you can have? Not something that they're promising, not something that they're selling. No, what is it that you can have today? And that's that's something that makes me tick. And that's the psychology, probably that's still there. Yeah. And the other thing is I think what's also helpful, but then again, you need to interview somebody of my team for that, is that well, I have a different view than than than most positive audio cult scientists that are very linear. I'm a lateral thinker, so I challenge them quite a lot on other ways that they could attack things. But I love it.

Isabelle: 7:26

But can you just explain what lateral thinking in is for uh linear thinking is very simple.

Bart: 7:31

I simple. I'm not saying that it's about it's for explaining this, it's linear thinking. I'm always saying it's one, two, three, four, five, six, seven, eight, nine, ten.

Isabelle: 7:38

Yeah.

Bart: 7:38

If you talk to me, well, we'll talk I will start with one, two, three, and then I will say yellow, brown, orange, and then I will go to alpha, beta, gamma, hop up. We all go in an order, but we go to the same direction, but you will have you'll see that I have different, what do you call it, different ways together? It's always to the same vision, it's always the same flagpole that you have, but it's not one line. Yeah, and that's that's a little bit the difference. So if you have, for example, what you see is typically architects in in IT or enterprise architects that have a holistic transversal view, they also have a lateral thinking because they start connecting everything that that as a linear thinking, you're not always connecting.

Louis: 8:15

Is lateral thinking a technique that you use, or that's I use it all the time. Is it in you? Is it it's part of your DNA or is something you do you do on purpose?

Bart: 8:24

Uh no, uh, it's part of my DNA, it drives my wife crazy.

Louis: 8:34

Now that you we know uh yourself a bit better, who you are as a person, your passions, your driver. Uh it would be interesting to understand your industry uh a bit better. So you said you've worked in telecom, you mentioned Mobistar in retail as well, before healthcare. How do you compare these industries with the healthcare industry?

Bart: 8:53

The three industries don't have anything to do with each other besides the fact that they're big data and they're qualified at the uh in the past as big data uh provider, but the the industries are completely different. I'm what you have is that Mobistar, I was lucky to start with Mobistar when they were at the glory time, so telecom where you have huge margins, and we're actually Mobistar is also uh a great story. It's actually a startup that became a huge organization, and I was lucky to get there at the at this really at uh decode when there was at the top. So that means you also had a lot of possibilities in terms of experimentation, innovation, also in terms of marketing. They were very advanced, and so and you had the means. So of course it's profit, of course it's it's sales oriented, but but uh that that combination that makes that that makes that you can go ahead and you can start things that were not possible before. And so that's what we did with with the mobile start. We started experimenting with those geolocation data and went a little bit further on that, and but also started developing a whole uh data environment to start exploiting data, so it was possible, and you didn't have that pressure yet because, well, if you have margins that are huge, that's never been seen before. Well, you you you have the time to experiment and to go for if you look at at uh retail within the retail what's what's there, you have very low margins and high volumes, it's a whole different ball game, it's completely different ballgame. So your return on investment needs to be almost immediately. Yeah, but try to build something up and and and building a new IT infrastructure or data architecture, uh new uh omni-channel uh environment. If your return on investment needs to be almost in three months' time, yeah. And so it's a completely different approach. Partainment is also non-profit, that means that you because it's human-centric, you don't have that pressure, you don't have any commercial pressure. So that means that you can take your time and really start. If we're convinced that it makes a difference for our members, there is the value and not uh how much they're going to sell. That's that's retail and telecom. Well, then then we're willing to invest in that. And that's a huge difference. So you have the time to invest, and we have and we need it because we have a huge debt. It's a completely different dynamics.

Louis: 11:08

I guess there's also a strong difference in terms of a sense of purpose that you find in the healthcare industry. To what extent does do you live that in the day-to-day and does that really make an impact?

Bart: 11:20

If you look at just that what our values are, I it's centered around to our members. For example, what one of the values which I love, there are actually there are multiple values I love, but what I like is what we simplify with empathy. If we would have been in another sector, well, simplify means okay, we make more automation, we're gonna be more efficient, and well, it's well that can be at the cost of we're not uh how do you call the fintech where we say okay, we're gonna automate everything. Well, what we do is we simplify with empathy, and that's what what I love about this. So, what we're going to do is actually we s we simplify process, we simplify automate elements just to ensure that we have a better human connection so that our team members or employees that are in connection with our uh with our members, well, they have more mental, how do you call mental space to start talking with them and listening, really listening to them. And that's that's one of the values. Another one is digitize to humanize again, is that you start digitizing processes that's better for our members so that that they actually respond to their needs. Again, we're non-profit, so and and that gives a lot of sense to people. So I always use an example on that when I have somebody who's new in the department that starts. I always explain to them a little bit the story of Dennis, which is uh a regular, what do you call member of Partenham? Is your typical persona? Well, actually Maria's our typical persona, but but I took Dennis six years ago when we didn't have Maria yet. So you can you can we can we can call him Maria if you wish. So I can talk about hours for Dennis, and I think then you're going to throw me out, and probably uh well you don't have enough time uh to uh to be up. But make a long story short is that what you have is that you have on the one hand you have somebody who has ambition who wants to do something from his health, but he comes by accident, by something that he didn't oversee, gets injured one way or another, and he gets into a a care traject care journey where he gets lost, okay, where he gets multiple opinions. So, what's the role of Barthenamut? Is to make sure that you, Dennis, throughout the ter uh the the the journey, well, he stays human, stays listened to, supported to it in a proactive way, and that we're there for him to help him to make the right choices. We're not going to do the diagnosis for him, but if he's looking for an expert and he wants to make an appointment, well, we're there to help him. And that's the the vision on how we put our members central to our uh to our activities, and that makes a lot of I I hope it makes a lot of sense also for you because there's always somewhere a little bit Dennis inside of it's really interesting to hear about Dennis and and actually you've created our your own segue into our next question, which is we would love to know about the vision of Partenamut.

Isabelle: 14:18

And to kick this subject off, we actually have a question from one of your peers. His name is Professor Giovanni Briganti. I hope I'm saying that correct. Mr. Giovanni Briganti is the head of department of computational medicine from University of Moss. And his question to you is what is the humanist ambition for the adoption of AI in the service of patients by health insurers like Bartin Amut.

Bart: 14:46

So I can go back to uh to the story of Dennis. So what does our vision is actually to keep the the the whole journey humanly? Yeah so that means that uh either our vision is that we need to evolve from what's uh what you as uh what you what we call as a as an insurer to an assurer. We need to make sure we assure you that well if there you have a problem, if you have a need, well you know what, you can call us. We're there to help, to listen. That's that's that's the humanistic approach. And of course, you can only do that. What's the reality today? Let's say let's let's just say it like it is today. Well, you need to fill in a form. If you don't have fill in the form properly, you will not be reimbursed. It's a transactional way of working. But that doesn't help you. Again, if Dennis he wants to be a marathon and he's completely lost and he sees that he will never be able to run again. He's not interested if his form is completed enough. Of course he's interested, he wants to be reimbursed. But he wants to be listened to. Well, if he wants to be listened to, then you need to make sure that that form is already pre-treated, that everything is filled out, and you can just ask for a validation. And so that's a little bit if you'll ask for an humanistic approach, that is well, trying to be there during the journey and try to proactively answer the needs by on the one hand having somebody physically that's listening to you and that is also mentally able to do it. I mean, have the uh the space to do it and not just ask you to appropriate form. And the other hand, also to provide them the right information at the right time. That's that's the way that we what do you call it implement AI. So it's a a human-centric approach in terms of AI. We put the human and the needs of our members but also our team members central and make sure that how we can uplift them and what they and where they can make a difference for our members and which would and the things that it doesn't have any added value. Well, we try to eliminate it as much as possible. That's a little bit the the vision. So it's kind of like a reassurance to uh people's fear of it's uh reassurance, having the the right information on the right time. It's it can be having access because like mom was explaining about tennis or having access, uh try to get an appointment to an expert.

Isabelle: 17:05

Yeah.

Bart: 17:06

I it's it's not easy. I and especially if you then if you're hurt and it already takes, I'm not saying whatever, uh one month to have an appointment to an expert. You won't have a second opinion, but you're hurt. You already need you still need to wait another month. Well that's that that started to become problematic. So also helping there to get more access, I've and more easy access to have to know which expert is maybe available, having access to an agenda or something like that, that's something where we would would like to go into the future, would like to work on.

Louis: 17:35

Would you say that the positioning of Partinamut in its industry is similar to other players, or would would you say that Partinamut is fundamentally different? And if so, how the challenge is the same for all health insurers.

Bart: 17:51

So there is no difference. But there's one huge difference is that we're an independent health insurer. Uh so that means that we're not we don't have any political uh links, we don't have any associations in one way or another. What we do are completely focused to where we can make the difference for a member. So we don't have any other interest than that. And that's that's the huge difference for us. So we're not linked to uh to to all that. There's less politics involved. There is no politics, there's no religion, there are no uh economic affiliations or it's and if you look at other uh health insurance that are affiliated to uh political instance, religious instance, or even economic uh instances. So that's not the case with us.

Louis: 18:40

We've discussed the vision of Parthenamut, the ambition, the values of Parthenamut. I would like that we now delve into something more tangible and concrete aspects of uh your role, and that we talk about your mandate and the real some real business cases at Parthenamut. And to start with that, we have a question from another peer uh peer of yours. And uh this question is from Christophe Megank, he's the chief data officer of BNP Paribas Fortis. Let's listen to his question.

Christophe Meganck: 19:12

Hi Bart. So Christoph Mehanck, I'm chief data officer at the BNP Paribas Fortis, and I have the following question uh for you. As a Partinamut, you certainly have quite a lot of processes. I suppose you have also done some epochs on AI, traditional and solely also on generative AI. And my question is, I'm wondering how did you succeed in bringing these POCs not only into production but also in covering your a broader range of processes within the Party Namets by AI, traditional and or generative AI.

Bart: 19:49

Just to be very clear and and meet expectations, so the full scaling in terms of process and automations were not there yet at all. What you have is that what Christophe is referring to is that. You have actually two elements where you need to work on. On the one hand, you have a huge stat in terms of technology. You need to know that we're working on technologies that date from the 80s, 90s. We're still working on mainframe uh systems from IBM, uh, a little bit like the the banks, also. So, what you need to do is you need to start building things around it. So that makes a lot of time. And you also need to make sure to find the right balance. Another thing that also uh a success factor is can you get the business with you? Can they see what's the business value? The first success story that we had is we called it FOCI. What you need to know is that we started with that three years. So what we have is we had an intelligent system that knows a vocabulary. The model is trained based on all those conversations. So and so we would what we started to do was the next use case. Okay, well, actually, for some topics, what we're going to do, it's very simple. We're not yet on the handling automatically. We're not there yet. But actually, just by identifying actually, that's this category, that's actually this team, that specialized team that should be said that should be handling it. So what you have is that you have the email, you have our NLP module that starts reading the mail and starts redistributing it to the proper teams. Before it was somebody who was reading those mails one by one. I know team of males, well, today they're responsible for the quality of of the distribution. Again, it's always step by step, keep it simple. And I the KISS principle, so that people can follow. And that's a little bit the success. I what we that were the success stories that we have. Does that mean that everything goes very well and that everything is fantastic, etc. Far from well, circumventing or trying to circle around a system that's 40 years old and not made for that, that's not an easy one. But but adapting to something that makes a difference for employees or for team members and building little by little knowledge was one. For example, what we're going to do, and that's starting to become a little bit more innovative. For example, we have the same model that we use that know the the uh that knows understands the language of our members. Well, the next step is how are you going to link that to a knowledge base? Because actually you're calling, Louis, you're calling to a service center agent, and the service center agent, well, he's going to reply to you. It's based on knowledge. Well you need to know that we have a huge challenge within Partenamut is that to become a service center agent that's independent, senior, that can solve first line autonomously your questions, it takes at least two years. Because you need to know we have a Brussels government which has quite a legislation, we have a Wallon, you have the federal, it's huge, it's complex, and and every two weeks there's something that changes. So it's a knowledge base that's usually we are training. So what we're doing now is now what we're and we're going to launch that on two subdomains, which takes actually two years for people to get expert on it. A knowledge base that allows people that imagine the wish we will start in a service center not tomorrow in a couple of months. You'll have a question. Well, you can interrupt immediately. Allah ChatGPT. Okay, well, actually, this is the question. Well, if I was a senior service center agent, this would be my response. And you're going to say, Well, Bart, this is not sexy, this is actually ChatGPT. Why don't you just plug ChatGPT in it? Well, what I would like to say is that well, we tested ChatGPT and uh and the model from Microsoft and we put it in competition of ours. Ours was in 90% of the cases was uh was the right answer with them, uh with with the GPT in traditional, it was 30% of the cases. Because ours was trained based on the real questions that people were asking already three years ago with the knowledge base that were adapted, and so that's that that's one of the ways that we manage our AI use cases.

Louis: 23:58

Nice, it's really uh a beautiful case you're sharing here. So let me try to rephrase what I understand from this concrete use case. So we're you're we're talking here client service. You analyzed voice conversation first, based on that, you trained the model. You reuse that knowledge also to analyze on other channels like emails, analyze unstructured data, and act on those emails, for example, dispatch them to the right teams. And you've mentioned at least three benefits, concrete benefits of this use case. You've mentioned that you've reduced in a team of 200 FT in the customer's time service, you've reduced a uh a part of the of the team.

Bart: 24:45

So no, the workload. Of the workload. That's the difference because again, we're nonprofits, so the the thing is that we reduce the workload of those people, okay, allowing to get them more time to listen to our member. That's a really important one. So we track in terms of timing. So that means it can today translates into having more calls. Tomorrow is going to be translated to listen to you as you would have a dentist at a certain point.

Louis: 25:08

Well, so you kept the same workforce, but they are more efficient, they can do more. So one gain is efficiency. Another gain you mentioned is actually client satisfaction has improved. What what what else other benefits I'm missing?

Bart: 25:23

The thing is that we're having we're setting up building blocks that can be uh continuously being used. That's so what I'm trying to explain a little bit is that we have uh a vision. You start with something simple, which is a use case, but you have a vision that goes further. What I mean by that, the step from having an email redistributor to an intelligent knowledge base. Well, the next step is that you're going to have an intelligence chat, but based on your because if you're now answering emails, are you going to answer a chat? It's going to be the same. And so what we're doing is we have how you call it an iterative approach. Each time we set one step further on how that we can make a make a difference. And I think and what that's what's what's the other benefit is that we learn from that. We see we we have how do you say it's each element we have it under control and constant and we're able to constantly optimize it. So that's that's I think that's also one of the advantages.

Louis: 26:19

Yeah. You also mentioned that the training of or of new people on boarding is uh long, takes up to two years, and that thanks to that probably you can be more efficient as well in this process.

Movify: 26:32

Yeah.

Isabelle: 26:38

We're quite interested to know how you deal with element stock will take a lot of time. For example, in the banking industry, these are KYC, AML, and fraud. Like how do you deal with this kind of automation?

Bart: 26:51

So if you compare it a little bit with us, so what you have we have still uh an a sector that's heavily driven by farms and and the small piece of paper that you receive still from uh medicine or from the kinesis therapist of manual therapist, etc. And so I talked to you about everything that's voice, natural language processing, everything that's text. And so the other one is computer vision. And so uh there we have the ambition. And so to develop an asset that's able to uh out of the 520 different forms that there are, that we can are able independently for which form, that we're able to uh extract the the proper information. Today we're able to do it already for uh I'm having a look, I'm not knowing it by heart, but I think it's for five types of documents, and that we are able to extract, of course, these are the documents that are most sent. Yeah, to extract the right information just to make sure that it's already integrated into the into the system because what happens today is you you send your paper, yeah. It comes into us, it's scanned, and then the scan has been looked at, and then somebody put it manually into the system. That's actually what's happening. And so what we're doing is that having this computer vision asset that's also evolving all the time, well, the ID is that from uh tomorrow all the documents that we have will be extracted automatically, corrected if necessary, and so that in the near future that everything that terms of well um that manual part that well which doesn't have a lot of added value can be replaced by actually for the the the the real complex cases that demands more seniority expertise where you really need to take some time that we can take more time for that one and that we can improve our services. And today I uh we're not that far if I if I uh as I would have wanted to be, because you dare we really always say with the voice, I could it's other systems, and it's a little bit a more modern system where we have our our service center on driving on, so that's easy, but the claim handling, it's it there you have that mainframe, and and then it becomes difficult, and then you need to start working with a combination of different tools, and you need to make sure it's not overloaded, etc. etc. So though we I I wish compared to what we've already have developed and where we see that the minimal viable projects are well, and the way that we were able to implement it, it's it's not at the scale of I wished it was. Yeah, today we're we're have having the equivalent of 50,000 man hour more or less that we're uh that we're compensating with our automations. Yeah, but it's it's yeah, but it's it's if you look at what's possible, imagine that there would not be a mainframe. Well, that would be uh that would be everybody at the same stage as you.

Isabelle: 29:50

No one's happy with the results.

Bart: 29:53

Yeah, but it's a bit frustrating because if you see what it can do and and how it's possible, and if you are able to uh extract and import the data, even though it's uh in batch mode, you say, okay, well, actually you should find something to overcome them. But uh that's the debt in terms of technology is that uh is that big that that well there are other types of investments needed to to uh to s to overcome that one.

Isabelle: 30:22

So you've spoken a lot about the product side of things, the technical side of things, and now we would love to talk a bit more about the culture side, the culture change since the the arrival of um AI and or especially your team at Barthenamut. You said you created a team from scratch, is that correct? Yeah at Partenamut, yeah?

Bart: 30:43

Um it's the same also at Carrefour. So at Partenamut we start with six.

Isabelle: 30:47

Okay.

Bart: 30:48

Today we're a department heading towards 70. We truly believe uh in terms of Partenamut that that that we can make a difference with with the data that we have and truly believe in data and AI-driven transformation. So there's a reason also why we invest uh uh we have invested uh largely in that. The way that we evolved is the same way that I explained a little bit before with each time with small use cases showing each time what are the benefits that you can generate, and then identifying what are the interdependencies. The team has evolved from six people. We started just building dashboards, corporate dashboarding, to now a team that goes from platform product managers to platform engineers to data engineers to data and AI business partners to uh data security, uh, of course, governance, uh but also uh data and AI governance, ethical users. So we really developed that. And then you need to start hiring people. And one of the things is that again, I couldn't tell the whole story about Dennis because you didn't have enough time on here. But I the thing is that, and maybe it could make you laugh, but I explained that story about Dennis to explain to them afterwards, independent of who's in my organization, if you're a business analyst, you're a data analyst, or you're a data engineer, how that you can contribute to that vision of Dennis. And I always say to laugh, but I'm not I'm not that kidding that much. I'm going to come back to you and I'm going to ask you the question: what is it that you did to change the life of Dennis? You as a data governance man, what is it that you did for Dennis? And that's the vision, and so we repeat it all the time. And and having that vision that allows us to make a little bit of difference gives also purpose and sense to the team. And and that's the biggest change I think that that has happened from a dashboarding team that's actually just executing in what you call uh do you call the post box where you say, okay, I have a request, okay, well, I'll take it and I'll deliver you something. To okay, well, actually, well, no, we have a vision. Proactivity, this is what needs to be done. Those use cases that I'm telling you, they're coming from people within the team that are intrinsically interested in what a business is doing, they know what the technology can do, they know where we're heading, and they start sparring with the business. But if you're going to ask the business now, okay, what are we going to do about AI agents? Nobody's going to have an answer because they're not aware about the technology. And that's something that try to install. So everybody within the team knows how they're contributing to the success of partenaments and to the success of the department.

Louis: 33:30

That's a much more centric, uh customer-centric uh approach. I guess six years ago, when everyone was building dashboard in the team, it was internal focus, or it was your users who were, I don't know, internal management probably. But now the focus is on Dennis. So actually, you've worked on the customer centricity uh a lot.

Bart: 33:49

Yeah, I try to make sense about what is it that we can do that makes a difference for the customers. At Carfour it was differently. How can we where's the money? Yeah, and how can we do upsell at cross-sell? And so we build a whole team around that one. So can we make a difference within the scope that's been used with our laptops? What how can we make the difference? And then afterwards you start building. May you give sense to people. That also makes and that this because you know where you're heading, it's lateral thinking, but uh the I can tell I can tell I could show you the data strategy of six and a half years ago, I can tell you're still heading in the same way. A lot of things are changed, but the poll is still the same. And I think giving that that that vision where you're heading and concretizing that and remembering people on a regular basis on that one uh that gives sense to them, that makes a little bit of difference, that helps you build teams and keep people also within uh within the organization.

Louis: 34:54

What's the main driver today? Is it more are new initiatives more technology based? Meaning okay, you and your team realize there's potential to deliver something extra with the help of AI, or is it more customer-based, meaning focused on Dennis and what we can do for him, or even you serving the business, business business focus, and and asking your colleagues from the business what can we do to help you and make your life more performant?

Bart: 35:27

So I'm going to give the political correct answer, all three of them. Uh, you don't know. I am uh some of the colleagues will be listening. Of course, we need to continue working on the thing that the business is asking because, well, of course, the strategy in terms of business is decided by business. On the one hand, on the other hand, uh it's our role because we're a technological, uh techno transformation department. We're not an IT department, so it's IT is you deliver what's been asked. We're transformation in terms of technology, so we have that part in terms of geeks. So from the moment on, there is something new. We're going to try it out, we're going to start experimenting with it. But very, very quickly, if we see that it makes sense, we're not going to experiment on something that's completely out of uh the business, but from the moment that it can make sense, we have our data and AI business partners, we have our product owners that need to start building a vision around it. Okay, what does that mean? The the what we call the Vokchia case, that was the from Audio Tob. It was because we were just messing around, Libutin, Amazon Web Services, and we said, okay, this is interesting. Hey, we're still geeks, but it became more and more concrete because well, we started to understand where could be the business value in that. And so what our ambition is not only to help the business, but also to partner up and and help them shape the the business of tomorrow with with technology. And that's I think that's another approach. And that's yeah, that means the tree, but we'll continue experimenting with new stuff.

Isabelle: 37:01

Uh referring back to what you said at the beginning of the podcast on about your psychological background. Do you use it in your day-to-day job in terms of motivating your team?

Bart: 37:11

Try to find what is it that makes people trick? What gives them energy? What does what is uh what drives them, and what are the the elements that really don't give them energy? And and that's something I I really like doing. It's a little bit the psychology part, and then identify, okay, where am I going to put the right people that have the right drive on which type of challenge? So you can have a a job title, but within a job title, you have quite a lot of responsibility. There's something that you really want like you ask me, what does make you tick? What is it that makes you drive? Well, and you put people on those types of projects and you make a stretch, you try to stretch them, and that's a little bit what I also like to do is okay, what you do good. Well, now we're going to challenge you really and see let's let's see, let's make it a challenge, but it's something positive. And then next to that is see, okay, what is it that we need as an organization where we're heading to, but what you want to develop, and you're not feeling comfortable with okay. Well, let's coach you on that one, let's put you on a project, let's help you, let's do it together. And that's that's why it's maybe it's basic in terms of people management, that's actually what I really like doing, and then what really gives me energy to come back to what first, if you then say, Okay, well, I had some idea of what Lewis, for example, where you were capable of doing, and while you mastered completely your vision, and you come with something that's extraordinary with I didn't have thought about it because you're proud of it. We say, Yeah, you didn't thought about that part, but look at that, look at what we see here, and then then my day is great. I get goosebumps just by saying it. This is this is this that makes my day, and I'm lucky. I um you also asked me what's what's the difference with Partenamut and and the other um organizations I work for because we have we're as we are an organization by nature we have a purpose. I'm lucky to be uh having a team. It's a huge department of people that are really cult. I it's for me in terms of quality, in terms of competence, expertise. I think I have uh people that know they're really a level higher than than the ones that I had within the profit sector because they're there because they want to make a difference. They're not there just to have their own the salary, they're not there. No, they want to make a difference. So the level is uh is it's way higher. So I'm really how do you call it uh in French? We say en tournée, on circles with with people that are uh that are really smart. Yeah, don't tell them that this is the third time that I'm building teams, it's incredible, and I really love it. So they're challenging me. So I'm it's two or three years ago I I came to the consultation. Actually, I'm holding up this department. So guys, go for it. Yeah, let's uh take over and and and and and and and tell me you strategy is clear, vision is clear. It's your time, so tell me how we're going to do it. And then I went on holiday for two months.

Louis: 40:06

Nice, nice to hear there's so much uh commitment and engagement in the team and in the industry uh overall. We're going to close off this podcast with one last question before we jump into our final quick fire drill questions. Here's CAD. Okay. One question first. Where do you see the future of healthcare in 2030?

Bart: 40:38

I already mentioned a little bit this the biggest challenge is going to be from uh it's how do you call it health insurance? So health assurance, uh manifest is a book, it's a small book that's been written uh during the corona. That goes where we need to move from health insurance, meaning okay, well you have uh you've fallen and we need to reimburse you and it's a transactional part to health insurance. We need to make sure that the next time you don't fall anymore. And it's going to be a huge shift. So this is where we need to go. But in order to get there, I uh I hope well I'm I'm I'm I'm a bit pessimistic, but I hope that that in order to get there, well there's a and there's an urgent need to have a smart health data ecosystem where we start exchanging data. And for me, I hope in 20 well and and and thirty, I hope we will really move on to that and that we're actually having a landscape that's very dispersed, that one way or another we're starting to get by something that we're it's interchangeable and that we really can make a difference for for a member that from time to time get lost because if you go from one specialist to another and another hospital to all your data is not falling. Yeah, but yeah, but we cannot all rely on you. So I so if we want to move from insurance to assurance, we need to open up what we have today in terms of transactional data to to other data sources. And I hope that the rest of the sector would also evolve to that. And you see there are a lot of initiatives, but I hope that that in 2030 we have something that is concrete and where we can really make a difference for uh for our members.

Isabelle: 42:27

Now to really end the podcast, we have a fire drill set of five questions. So, first question prevention or cure?

Bart: 42:37

It's easy prevention.

Isabelle: 42:39

Human intuition or data-driven insights?

Bart: 42:42

Oh, that's a difficult one. No data-driven insights. But yeah, it's difficult because uh um human intuition is still uh more um powerful than data insights because data has its limits.

Isabelle: 42:57

Yeah. Interesting. Agentic AI or generative AI?

Bart: 43:02

No, agentic, of course.

Isabelle: 43:04

Obviously, just the trick question there. Uh doctor or algorithm, who would you trust first?

Bart: 43:11

I uh uh for me it's human contact. Uh it's also um that try to explain a little bit um human um human-centric AI at the end, and I know that there are quite a lot of discussions, there are quite a lot of articles that say yeah, but uh uh an AI has more patience, etc. Me as a human today, and probably also something to do with my age, I prefer talking with people. But please, please make sure that my form is already complete and I don't need to bother about that, but just let me talk to somebody that understands me and listens to me and and has empathy. Empathy, yeah.

Isabelle: 43:52

Yeah, that's a strong uh word of the day, I think. And last question speed of innovation or perfection of government governance.

Bart: 44:01

For me, speed of uh innovation because um I'm I'm convinced you need to iterate a lot, but you need to take into account that there are some frameworks that you need to take into account from the start. So if you already start doing something that you know you will never be able to do because you cannot govern it because it's illegal, blah blah blah, you don't start with it. But you need to make sure that your framework is in line and the guidelines that you are are in line with the cycle in which you are within your innovation process. If you're just about IDs and proof of concepts, it's another um framework, a legal framework that you apply if you are going to start to deploy it for 1.3 million members. So you need to make sure that there is a right the right balance between them. So not at the end, so it's not either one or the other, yeah, but from the start, stop thinking about it. But don't let it um how do you call it um limit your innovation?

Isabelle: 44:59

Thank you so much for your answers and your clarification on each point. Okay, um, it was it was really nice to have you here. So thank you very much for coming.

Bart: 45:07

Oh, thanks for having me.

Louis: 45:09

Thank you very much, Bert. It was a pleasure, it was insightful, and uh congratulations for what you've built so far. I mean, in six years building that team from six to seventy people and very mature already with concrete use cases. I think it's it's nice what you guys are doing out there. So congrats.

Bart: 45:27

Thank you.

Louis: 45:27

It's the team.

Bart: 45:28

Uh I was just I'm I'm just selling it.

Isabelle: 45:40

If you are interested in knowing more about Moviefy, don't hesitate to visit our website on movify.com. Stay tuned for the next episode and don't forget to follow us on LinkedIn and Instagram.

Movify: 45:55

Thank you for listening to Flux. If you like this episode, subscribe and rate us on your favorite podcast platform. See you on the next one.