It is my pleasure to host André Rogaczewski, the CEO and Co-Founder of Netcompany, to discuss how AI is affecting the IT services and broader tech industry. André, thank you so much for taking the time today. Thank you, Aditya. Great to be here. It's a pleasure, as I said. Before I start, this is an interactive session, so we will open the call for Q&A towards the end, but we will start off with a few introductory remarks from André, and Q&A from myself. Without any further ado, I'll turn the floor over to you, André. Thank you so much, Aditya. My name is André Rogaczewski. I'm the CEO and Co-Founder of Netcompany. My background, originally, I'm very technical. I studied math and computer science back in the 80s and 90s and actually looked at AI even in those days, but we didn't have the computing power. Started Netcompany in 2000 and 25 years down the road, we are almost 10,000 people. We are a system integrator, but bringing our own software and platforms with us, and we are definitely utilizing all the possibilities for agentic AI as possible, as much as we can. Let me just give you a short intro and then maybe we do some more Q&As as well on this subject that I think in all industries, but specifically also in the IT industry, is of great interest. Now, it's not that agentic AI is a complete surprise for the industry. It is not. I think most of it expected it to come, but it's like with all technologies, when they actually arrive, it's still a mind-boggling how much they can actually do and what they're used for. Now, the big difference between the AI you know from probably using ChatGPT and other algorithms is that agentic AI is just one step further down the road. You have agents that can plan tasks and execute on tasks and actually even solve larger issues or problems. When you think about AI being utilized by, for instance, journalists in a newspaper, journalists write text, right? Code is also text. What we're seeing at the moment is a lot of algorithms being used for writing code and it makes the life for many programmers much easier. This is definitely a very interesting development, and it is also something that will change the business. People in the IT industry saying that this is not disrupting the industry or it's not that interesting, they're wrong. This is disrupting the industry, and it's gonna disrupt basic programming, it's gonna disrupt smaller software companies making simpler applications, and it's gonna disrupt the companies who are primarily selling resources to do programming in simple areas. It will also change the career model of programmers and technical programmers. We will still need to learn how to program and understand programming, but so many of us have to do orchestration of programming instead and think orchestration more than actually programming. It will change the career models as well. I'll go into further details on that later on when we do the questioning. To be quite frank about this, I mean, looking into what the models can do at the moment and probably the next one to two years, this will affect basic programming, and it will affect smaller software products that are peripheral or isolated from the kernel or the transactional kernel of enterprises. What protects you against this disruption is if you're working primarily with talking to customers about their requirements and adjusting your system landscape inside their company. If you spend 80% of your time not programming, you're probably doing the right thing. If you're spending time in a regulated industry where there's a lot of legislation, high complexity, hundreds of integrations, high demands to technical performance and security, and a sovereignty issue about where data is, and if you're able to host yourself and you know about data ownership and you're in some verticalized, industry where you have a platform or product, you're probably fine, and maybe this is good for you. If you're not, if you're selling hours, if you do have a lot of Java programmers just programming, doing the tasks that they're set to almost on a daily basis or weekly basis, then you probably need to change your business model. If you do have a product that is quite simple in its nature, that product will probably be able to be rebuilt by an algorithm or your product will become an algorithm. That goes for simple CRM systems, ticket management systems, could be BI systems where the algorithm will now read basic information and raw data and come up with the functionality you need instead of you having to build the interface towards the customer. This will affect both software and consulting companies. I think a good way to think about it is also to think about sea where you have different fish swimming around. This has a lot of impact on productivity. If you have 10 very good programmers, they can probably program as much as 100 people can do today in, say, three or six months. If you have 100 people, they can probably program or build what maybe 500,000 people can do today. That's the way you should look at it. The competition will not come from someone above you probably, but it would come from some smaller fish. That fish would be able to build faster than yourself if what you do is basically programming. To call the winners in this industry, because most, maybe half of the companies, they would just see a declining revenue over time, and they will have to fight for more customers, and the prices will go down in this area. To figure out who is gonna be the winners in this with this disruption going on, you need to look for growth and everything at the same time coming up. Winners are gonna be the companies who are able to deliver fixed price, outcome-based domain knowledge systems and platforms with a lot of embedded AI hosting inside of it. I think there's a whole new market appearing, because if you look at a company like ours, we are what you probably would call in global terms mid-sized. We're 10,000 people. We do a lot of very domain-specific systems. We deliver regulated systems for airports, for governments, for tax and customs. For us, this is great because we can now compete with some of the biggest companies in the business, the ones who've had these systems for 30 or 40 years, where it's difficult to go in to their domain because they built the systems, they own the systems. If you have enough domain knowledge and you have enough skills on the particular subjects and the technical skills, with this technology, you can go in and rebuild some of the older systems. It will take time, but instead of it takes like four or five years to replace a legacy system, now you can make it. Maybe do it in nine months or 12 months, and you can do it with less risk if you embed AI into your delivery model. The winners in our industry are the ones who will change the career model, so they encompass the competencies I'm talking about, change their delivery model to become AI delivered, still with a great focus on domain, integrations, sovereignty, legislation. The ones who can do that using AI and convince the customers that what they're getting is just as good as what they had before. To some extent, we're gonna see a shift. I think you're gonna see the C level shift. Some vendors will go up and become bigger in the terms of what they can deliver, not necessarily in terms of how many they are, but they will definitely grow in revenue, and they will also because they're so good at delivering fixed price in these domains, they will become very successful in getting the right margins as well. The ones who do not shift are gonna stay where they are, and slowly but surely over the next one to two years, you will see them have a declining revenue and even struggling with profitability and having a lot of layoffs. It is a disruptive force, but it's also as always with new technology, a great opportunity for many IT players. I think that's my introduction. Let's have a lot of questions and let's have a dialogue about this because the devil lies in the detail, and I think there's a lot of different types of both business models and software products and vendors out there we can discuss. Great. Thank you, André. There's a lot we can interpret there. Maybe to take a step back and set the scene, where are we in the agentic AI adoption cycle? 2026 has been called the year of agentic AI, the year where you will see a mainstreaming of AI usage for knowledge workers. Is that something that you are seeing right now already with your customers? If so, which areas are you seeing that in? There's no doubt this has been taken down. People see, especially the last three months, even the last two months, agentic AI has proven to be, you know, it's just better and better every version we get out, and this is happening. I think within the next six months, most programmers in the world will know and understand this. You don't understand it before you tried it. I have spoken with very experienced both architects and programmers, and remember, I'm a programmer myself originally. You don't understand the power of this before you actually sit there and do it yourself. I think in three to six months, most of us on this call will be able to build our own apps to our telephones and even build our own websites without programming one single line. For all these small applications, workflow applications, things that you need, I mean, even, user interfaces, GUI and stuff like that will already by the next 3-6 months be done by algorithms. This is happening. Understood. To follow up on that, when we think about the different functions that IT services vendors are doing, whether that's in pure system integration or ERP migrations, engineering R&D, BPO, et cetera, which parts of the value chain do you think are actually going to see a meaningful shift over the next, as you said, 6 months or maybe even 12, 24 months? Which types of work do you think will be more defensible? If you could maybe expand a bit more on that. I mean, you're already starting to see it. If you have a, say, large programming outsourcing centers where people are working, you know, I mean, where the methodology and this system is set up like you get a ticket for performing a certain task or even a set of tasks that you have to do in some system that you are maintaining, and then if you, say, you find another job and another programmer takes over from you, which happens all the time in those centers, these centers will not be having that many people around anymore. You still need someone to control them and run them, but the programmers are gonna be substituted by algorithms, and it's already happening. Outsourcing centers where you work almost by the hour or by the day, programming Java or.NET or whatever you're programming, and you're just one piece of a bigger machine where programming is the thing. I mean, that's gonna be disrupted over the next three to six to nine months. Many of these centers are typically placed in cheap labor countries. The same thing goes for if you're working on a user interface, say, you have some websites or you have a CRM system or you have a ticket management system or you have a BI application and it's not the core of your business, but still you need it. You might hire, say, three or five people to change it here and there. That's gonna be done by algorithms or people having algorithms with them, right? It's not that people are gonna be replaced by computers. People are gonna be replaced by fewer people using algorithms in those areas. That's happening already. You will not see any effect on, I think, on the IT vendors doing more complex stuff. In fact, you might see them produce what they do faster and hence get a better margin because they are all running on agreements or contracts that were put in place over the last one, two years. In the beginning, you might even see some of them, you know, slowly but surely get rid of some of their programmers and then their margin will go up. Over time, if they don't change entirely their delivery model, they will also get disrupted. That's gonna happen in maybe one, two, three years. If you are delivering more complex application in a legislative world with a lot of integrations, you need to start, you actually need to do that now, change your delivery model, so machines and people will work side by side, delivering those complex models. You will still have people, you know, doing a lot of the work, making sure that legislation is followed, making sure that all the integration is in place, orchestrating the tests, but you're not gonna need as many people on the ground as before. Customers buying these systems will be very curious as to whether they could even get rid of their old systems using a vendor who knows how to use these tools, because the reason why we're not getting rid of so many legacy systems is because the effort and the risk is too big. With these tools and the right vendors, some of the old incumbents owning systems that are 30 and 40 years old could be in danger. That's something I really think is interesting, of course, being in a company. Got it. You just highlighted a very interesting dynamic here. On one hand, the productivity of a developer or engineer will go up meaningfully as a result of using AI tools. On the other hand, you think clients might actually decide to modernize the legacy systems even more. How are clients responding to this increased productivity? Are they using those savings to, you know, ask for better terms on pricing, faster delivery times, or actually are they reallocating some of those to other higher value add digital transformation work, modernizing some of the systems which they haven't got to so far? Which perspective are they taking? Look, I mean, our customers, all customers are also looking at this, and they will be using AI or agentic AI inside their business processes, right? The reason why they can't use AI and get all those benefits is because their IT systems are old. They use AI on the side, like we all do sometimes, get our answers, and we try to put it into our old systems, and we try to do the work that AI should actually be doing. I think the intelligent IT vendor would take the short-term productivity gains and invest them into platforms and software products that has embedded AI in it. You can give customers an offering which is much more compelling than it is today, right? Imagine most systems will help users throughout the entire business process, and some of the user interfaces will even disappear. Some of the screens will disappear, some of the graphs will disappear because you can just ask the algorithm to draw up a graph. It's a question of getting the data in a legible way and making sure that you can manage the data correctly. For that, you need the platforms and the products that are able to do so. If you have the data there, and you're living up to all the things I mentioned before, then you can ask AI to help you doing a lot of stuff that is done by humans today. I think the intelligent IT vendor should modernize their platforms and use those platforms to replace existing old systems in order to set their customers free. The ones who are only doing programming or sitting on the side, doing peripheral systems, the ones only selling hours or consultancy because they're good programmers or technicians, they will have a harder time for sure. I think that business model is under scrutiny. One question that the IT services industry faces as a result of this productivity gains is the impact on pricing models. Are you seeing customers already moving from a time and material pricing to more of an outcome or use-based pricing? I know for you it's a bit different, but in general, do you think that will be the shift you see in the industry as a result of AI over the mid to long term? Yes. I mean, if many customers, I mean, this happens also in customers' IT departments, and it's really interesting to see this dynamic because if you have a company with, let's say, a big IT department, maybe you're used to having that department there with hundreds and hundreds of employees, and sometimes you might think, you know, "What am I actually getting out of this? I'm paying a lot to just have the lights on," right? Operating these systems and having all these IT. "But what are these guys really doing?" And there's. If you have a CIO that's not protecting his IT department, but using agentic AI, he or she will actually be able to show you a better throughput than before with less people. That will of course also influence your procurement of consultants and how much external help you need. Overall, yes, the short answer is, an hour of programming or an hour of Java expertise or an hour of, you know, some programming expertise is definitely not gonna be worth the same. On the other hand, if you can take on risk, if you can take on total delivery responsibility, fixed price, and you can actually get rid of some of the older systems, this is a great market for you. If you have the tools and you have the platforms, and the products or software that you bring in is AI embedded already or prepared for AI, this is a great market. Because most customers, most enterprises are looking at AI as a very huge accelerator for their business processes. You can probably save up to 20%, 30%, 40% of the time spent, even get accuracy up, become much more closer to your customers if you use AI the right way. I think most CEOs, COOs, board of directors, they know this now. They didn't two years ago, but now they know, and they see the IT department as an obstacle to get there if they don't follow suit. I mean, this is a time of change, but the cost of pure hourly programming is, of course, something that will go down. I mean, that is. I mean, people are saying something different. I find it very difficult to understand that. It's all about now. Another very interesting thing about that is actually, if you have a lot of bad code or your system landscape is very fragmented and you have what popularly is called spaghetti bolognese, there's a big mess in many of your architectural layers and the code that is there. To bring in AI and think that that's gonna solve your problem, no, it's not. We already see that in some cases, using agentic AI just produces enormous amounts of code, but it doesn't solve your problem. The winners here are the ones who systematically have platforms, delivery methodology, you know, a very structured model. For them, AI is gonna be a huge enabler to deliver even faster, replacing the old spaghetti. For the ones who are in the middle of this old code, AI is probably not the solution. They have to replace most of what they have with AI, and when they've done so, they will be able to. I mean, it's gonna be a completely different world. Looking at, you know, the IT space, simple products will be replaced by algorithms. Simpler companies selling consultancy hours will be replaced by smaller companies selling code by. Made by algorithms. However, the big complex solutions, the enterprise solutions, the solutions for governments, solutions for regulated industries, they are gonna be not affected to start with, and then slowly but surely, you see a maybe a different vendor space coming up and biting the existing old ones a little bit in the, you know, coming from smaller companies. Smaller. I mean, they're still big, but compared to the global huge ones, I think they have to change very fast in order to follow this exciting development. Very, very interesting. Given what you said earlier about the career model of a developer changing, and the skill set changing as well, how do you think about the talent mix for IT service companies evolving as AI becomes more embedded, in how you deliver your offering? Does the industry going forward need fewer engineers or maybe with different skills, or is it just an upskilling that will happen with the existing talent pool? Yeah. Same thing goes in any other industry, right? People say, "Well, we don't need programmers then. We only need architects," right? If you look at the way people are changing, even in I mean you see it already in the U.S. and some European countries already that people are changing their titles. Instead of saying, you know, lead architect or lead programming or some whatever, it says AI orchestration architect or something like that. Everyone is trying to move away from the code and to become more an orchestrator or someone who's in charge of optimizing code and running digital agents. I think a person who knows about AI and has discipline and works in the right company can become extremely productive with the several digital assistants already after a year or two years. We don't need to have 10 years of experience to become a great orchestrator in this area. You can be one of the best if you have 10 years of experience, but we are all human beings. To my knowledge and experience, I have to say, some people who've been in the industry for 15 years, they're too stubborn and too religious about what they've done. They think programming is art. They see themselves as artists. You find them in many IT departments too. They will not. I mean, these are not the people you want. On the other hand, you can find some really great experienced people who, with this tool set and the right attitude, could even, you know, could deliver as much as 20, 30, 40 x more than what they deliver today. I think you have to look at this as a new breed of IT people. The companies who have a career model. I mean, we've changed ours, right? If you look at Netcompany career model, we hire about 800,000 people a year. If you look at who we hired two years ago, they spent maybe the first six, 12 months being on some programming. At least part of what they did was a lot of programming. That's gonna change, right? It's already changed. Of course, they have to know about programming. They have to know about what it is, and they have to have some basic skills. Very, very, very soon they'll be set into a position where they're gonna be both doing design, programming, and test using agents, digital agents. There's nothing wrong with having a young, intelligent, skillful person to do that. I mean, they become productive extremely fast. I know there's a lot of people saying it's only gonna be the old engineers sitting there orchestrating all these agents, and all the young ones will not have a chance, and you as a junior will not, it's gonna be difficult for you to find a job. Well, it's I don't buy entirely into that. That might be a reaction, like the first three to six months, nine months. At the end of the day, the company that survives and thrives is the one that also creates talent and has the talent. That's not only old programmers or architects, it's also a good mixture of both. You need to change your delivery model. They need to know. They need to have discipline, very important. Everyone needs to know what who is doing what. It's not that AI's gonna solve that. A clear methodology. Where does data reside? What does legislation say? What does the customer want? How many interfaces do we have? What are our technical performance requirements? What are our functional performance requirements? What about sovereignty? Who controls the algorithm? Who controls the solution in the very end? When we wanna change it and maintain it, how are we gonna do that without, you know, messing it up and rebuilding everything again? Can we just do it in a closed loop, having AI helping us? If I change something because of a functional or business requirement, can it both change the code and the documentation, and can I explain that to the customer and deliver that within, say, weeks, maybe even days, instead of months? The way you're gonna judge companies. 'Cause there's gonna be winners and losers here. I mean, let's be honest, could we have won Netcompany? Obviously, we are very well-known in Scandinavia and also because we came from Denmark, we are one of the biggest, if not the biggest IT company in Denmark. Could we have won Heathrow Airport or the TSS thing in the U.K. where we deliver a custom solution for the U.K. of that criticality, so societal criticality, like an airport or custom solutions without having the platforms, without having embedded AI, without having this methodology and being able to deliver it, say, in a year's time instead of what would normally be three years? No, then we wouldn't have won it. That's how you look at it. If you can win things that before took three years to do, you can deliver the same thing in one year, take the risk, deliver at fixed price, and guarantees the quality, you have a good. We are in the right spot because you're spending most of your time discussing with the customer how it works, and not so much time on actually coding or programming it but rather spending the time in putting it in there and integrating it with what is there already. That's truly exciting. It's gonna change the industry for sure. It's gonna be a better world 'cause we're gonna build systems much faster than before and also the systems that before were almost impossible to renew or replace or even change. We can spend our funding and our investments into new and exciting stuff than rather than just maintaining old stuff, which I find really interesting. No, that is interesting indeed. I do want to get into a bit more on Netcompany's approach to all of this. Before that, maybe on a slightly related note, where do you stand on the entire debate around the death of software and SaaS 'cause of all the concerns around that that you're seeing today? Do you think enterprises will still use their you know software vendors who can maybe continue to innovate and sustain their presence, or do you think you'll see more of more enterprises going towards you know tools from the foundational model providers who then you know who are all coming up with their own plugins for various functions? I think the one-size-fits-all software era is gone. It's gonna disappear, right? There was a time, I think it was really irritating as well, but there was a time where everyone said this, you know, standard, it has to be a standard system. I think that terminology is misunderstood today, and I think it's gonna change. Yes, you need a foundation, you need a platform, and you definitely need to standardize what is where you don't differentiate. The differentiation comes from where your business processes are different from others, and that can be digitized really fast now. I think SaaS products that are simple, very simple, and one size fits all will and can be substituted by algorithms. Give you example. Let's say you have 50 salespeople working for you, and they have to report to you on, you know, progress with their customers. Normally you would buy a small CRM system, probably in the cloud or whatever, and they have to report on 10 metrics. Now, you would just have an algorithm listening to what they tell you after their meetings or whenever they tell that algorithm that they had a meeting with this customer, and the algorithm will ask them for the information it needs if they don't bring it, if they don't give it to the algorithm. You as the sales leader or the sales director will ask the algorithm, "How did we do today? How did we do last week? Can you draw me a curve of opportunities, you know?" You can make your own views. You don't have to have people programming those views because the algorithm will draw it up for you. You will have an algorithm talking to your salespeople, getting the information it needs to report to you, and you will tell the algorithm how you want that reporting done, and you can change that reporting on the fly, right? In this particular isolated example, in this, I would say, peripheral usage of IT, and if this system is a standalone system, and it's a one size fit all because it has a lead list, it has an opportunity list, and it has a graph of backlog and a pipeline, for instance, if you can get an algorithm doing what I just said, and it's cheaper and simpler, maybe even better for you, why not, right? Whereas if you are working, If this system had 100 integration points and it was crucial for your transactions, and it was the very heart of your business, you probably wanted to renew that system for it to be AI-enabled, so you could ask all the questions I'm just saying right now, and maybe even in real time. Most companies are operating not near real time, but on a monthly basis or even a quarterly basis. It's all been about reporting and administration. Most of the systems we put in place over the last 30, 40 years had to live up to a lot of legislation and, of course, but also a lot of history in the sense, in the way we've been administering the world with pen and paper. Now for the first time in human history, the interface between computers and people is becoming humanized. That is such a big change, fundamental change. For it to work in your core systems, you need to change your core systems. They're too old. They can't do it. There's gonna be at least five or 10 years of work moving those old systems into newer landscapes where you can start utilizing the fact that you have intelligence, that the gap between computers and people is becoming smaller and smaller and smaller. Hence, you can become that sales director that says, that sits there and you know, asks all these different types of questions about his sales team. Imagine if you could do that for your entire enterprise across all your systems, knowing exactly what's going on now or yesterday or in the last three hours or the last two days and compare it to the same period last year and you can even have systems reacting when they see things. I think I'm not gonna get too philosophical here, but for the simple systems, the simple SaaS models, one size fits all solutions that you just buy off the shelf for maybe even reasonable price, they're all gonna be replaced by algorithms very, very soon. Simple apps, companies building websites. I mean, all of you guys are gonna build apps and websites in one to two months from now without being programmers. Go ahead. Thanks, André. Maybe briefly, do you wanna just expand on how Netcompany is in position in this context, and what actually maybe you spend most of your time on today, as a business? Yeah. It's interesting because I know I said I had a technical background, and I have a technical background, but the last maybe five years, I've been spending most of my time building the platforms and products that we needed in order to replace older systems. We launched a product called Feniks AI nine months ago to use AI to figure out how the old system works and how we can replace it with a new platform. I've always been interested in exactly this, and I've been spending, of course, most of my time driving the business. I'm still the CEO of a 10,000 people company. I must say the last six months, I've been spending two hours every day just trying to understand all that's going on within the agentic AI specifically, because right now the development is really fast, and how to embrace that into our entire delivery model and our sales model, but also on everything. I think if you don't do that right now, you're really gonna miss a big opportunity. You can build a very compelling demo in a few days. You can have things put into place, working in production, if you have your hosting center, if you have all the competencies that you need, you can build something in nine to 12 months today that normally would have taken three or four, five years. That for me is truly interesting. For us, it's an opportunity, and that's also why we're so I think we're very fortunate and that we have our platforms and that we have our industrial focus, building solutions like the one for Heathrow, building tax and customs duty for the government system, the whole transportation business, the energy business, the utility business that we are in. I think these are the sectors where if you come with the vertical AI-embedded solution to replace all the old stuff there, you are really looking into a very, very interesting market. Actually, I mean, yeah, I spend most of my time two, three hours a day just looking into what's possible and how we can become even more powerful with what we do. Yeah, I think if you are a software consultancy company without platforms or products, you might need to change your ways. Because to get the benefits out of this, it's not enough just to use AI for, say, for instance, producing code. I mean, everyone will do that. It's not gonna differentiate you, and then you're gonna end up in this Red Sea, right? I think in more traditional times, you had independent software vendors producing software, enterprise software, and then you had system integrators. Then the system integrator allied with the software product, and then they went together to the customer and the project was, you know, just the analysis phase was like six,12 months, and you had to write many reports, and you had to plan, and you did this journey three and five years and this and that. Today, if you have this mixture of being a system integrator and a platform vendor, you can really go far. 'Cause then you have the skill sets and the technology where you can replace this heavy process of delivering regulatory systems. I mean, every time we replace a tax system or we replace a Heathrow, like an airport system or a transportation hub, or you go in and you replace the system for a huge utility or an energy company, or government systems, I mean, it takes years, right? That's not necessarily the case anymore. You might be able to do it in half or even a third of the time and still get a very good price for it and have a good margin on that, and that's disrupting the existing incumbents for sure, especially if their livelihood is selling hours. You can have a look at your invoices. If your invoice is saying that many hours for this person, that many hours this person were coding or... You might need to rethink your business model. I don't think customer are gonna– I mean, six, 12 months from now, they rather see outcomes and you taking more risk. If you do, they will reward you for it because they are paying so much for existing old IT systems, you can't believe it. Many companies, enterprise companies, have a huge IT budget where they're spending most of it, up to 80%, 90%, just paying for old systems running, doing basically what they did for 10, 20, 30 years. This is the market space that we are addressing. Got it. Maybe we can take one question from the audience. Operator, do you want to maybe open up the line? If you would like to ask a question, please use the raise hand feature at the bottom of your screen. If you are dialed in by phone, please use star nine to raise your hand and star six to unmute your line. I'll pause a moment to allow a queue to form. There are no raised hands on the webinar at present. All right. Well, André... Okay, looks like there's one hand that's been raised, actually. Maybe you can just take this question. Warren, d o you want to go ahead and unmute yourself? Warren, please ask your question. Sorry, can you hear me now? Yes. Okay. Perfect. Yes. Thanks for doing this. This is kind of a three-part question. I'm just picking any company just as a way of reference. We're just gonna pick Accenture, for example. The one thing I'm trying to understand is the first part is like when you think about a lot of your clients out there and a lot of companies that want to adopt AI, like, how good is their current data l ike, are we at a process where, like, their data's in good enough form inside, like in the correct format that they can use AI, or are we still in the very initial stages still? Where, like, you got a lot of disparate data, and they're gonna need the help of consulting firms, whether it's Accenture, whoever they may be, to get their data ready to really leverage AI. That's my first question. Then I'll let you answer, then I'll follow up with the second question. Yes. I think the question is. Yeah, I understand the question. You don't need a large analysis phase or to think again about your data because I think most companies actually did. What you need is probably some platform that can go in and layer on top of your data, even be built in parallel to what you have. You'll be looking for vendors who can do that instead of trying yet another time to clean up your data. I would look into building in parallel a system that takes over parts or whole parts of your legacy system. That's the way to go. Remember that to build a system now is easier, will be easier and is easier than it was just a few years ago. If you wanna replace the old system with a new one, you can build a new one pretty fast. Then, of course, you need to figure out which data you wanna bring over to the new system, which one you wanna keep in the old. I'm not sure it's beneficial for you to stay too long in this analysis phase that you've probably been through many times before. Okay. Perfect. Does it make sense? Yeah. That makes sense. Yeah. In terms of if we think about, like, BPO separately from consulting, and obviously, you know when you think about, like, the disruption on the BPO part of companies, businesses, whoever that may be, versus consulting, like, where in your mind is, like, most at risk short term versus long term, you know, when you think about what companies are looking to tackle, sooner rather than later in terms of reducing their costs? Well, I mean, you already see it, right? All the simple tasks, including programming, but also call centers or stuff like that, you mean you will pay less for that. As I started up with, any large programming centers or BPO centers where the tasks are very systematized and. Yeah Well-described, they will be using AI very, very soon, if not already. The price for that should go down. Okay. That's helpful. Thank you very much. We have no further questions on the webinar. Great. Thank you. We are out of time. Thank you so much, André, for taking the time to share your insights, and thank you all for joining. Hope you all have a good day, and if you have any questions, feel free to reach out to us.