GFT Technologies SE (GFT.DE) Earnings Call Transcript & Summary

Hello, and ladies and gentlemen, welcome to GFT's Capital Markets Day 2025. It's great to have you with us here in Frankfurt or online. Great to see you today. Today, we want to give you an inside look on how we are working with our clients on some of the most forward-leaning projects in the industry from AI-driven modernization to next-generation core systems and large-scale digital transformation. Artificial intelligence and emerging technologies are becoming central to how financial institutions operate and innovate. And GFT is playing a meaningful role in shaping that evolution. Let me start by introduce the leaders who will set the tone for today. Sorry, I need to go there. From our Board, we have Marco Santos, our Global CEO; and Dr. Jochen Ruetz, CFO and Deputy CEO. Both will begin with our 5-year strategy update, sharing where GFT stands today, what we have set in motion and where we see the next stage of growth coming from and why GFT is the artificial intelligence digital transformation challenger. So you'll also hear throughout the afternoon from several of our senior experts, including Ignaci Barry Willadel, our Global Head of AR and Data and Regional Head of Business Development for Western and Continental Europe. We will have Dean Clark, our Head of Technology Office and Chief Technology Officer of GFT U.K. We have Floris van Hist, our -- since January this year, our Senior Vice President for GFT Software Solutions; and Sacha Beck, Executive Director and Financial Service Lead at GFT Germany. Before we move on, I want to extend a warm thank you to my Investor Relations colleagues, [indiscernible] Nicole Schutford and Maren Dellis as well as the event colleagues, MarAllens Sabrina, for their tremendous work in preparing today's event would have been impossible without that. Thank you. We are also delighted to welcome outstanding speakers and partners who bring valuable perspectives from leading institutions across the industry -- and I will introduce them by order of appearance. We have Glaioanico, IT Superintendent from Bradesco Seguros from Brazil. We will have Mike Laws, Managing Director at Deutsche Bank coming from U.K. We have Alexander Graff, Senior Vice President and Managing Director of FICO from the U.S. We have Alexander Listen, General Manager at DKB Code Factory coming from Germany; and Eric Nercu, Vice President, IT, Property and Casualty Insurance and Corporate Services at Geneva from Canada. We truly appreciate your participation, your insights and your partnership. Before we dive deeper into the agenda, let me briefly connect today's insights with our last Q3 performance and the progress of our 5-year strategy. GFT delivered solid results in the first 9 months with double-digit growth in key regions like Brazil, APAC, Colombia and the U.S.A. At the same time, we continue to scale the strategic pillars of our plan, driving AI deployment across markets, expanding Rings as a core modernization engine and strengthening our position as the AI-centric digital transformation challenger. These achievements give us strong momentum for the midterm as we move in the next phase of our strategic execution. And much of that you will hear today reflects exactly that. So let's take a look at our agenda. We start with our 5 years strategy update from Marco and Jochen. We will then move into agentic AI modernization with Wings with Ignasi and Dean. We will move into a coffee break then. Maybe you have -- need to have some break and some coffee after that AI deep dive. Next, and this connects very well with the technical insights from Ignaci. We shift to the client's perspective with the impact of AI on productivity from Bradesco Zegurs, lafio Iuanico and Ignaci. We continue with the crucial topic of AI and anti-financial crime, featuring Michael Laws from Deutsche Bank and Floris from [indiscernible]. We stay with that topic of anti-financial crime with Alexander Graf from FICO, who brings deep expertise on decision intelligence, analytics and risk technology across financial institutions, and we'll talk about FICOs and GFT's global partnership, what we will announce shortly. So this is kind of a small spoiler. We then explored the joint success with DKB and GFT for efficiency, agility and digitization presented by Alexander [indiscernible] and [indiscernible]. And this will follow by future core systems and tech leadership with Eric McCu from Veneva and Marco Santos again on stage. And Marco will then also return for a brief closing message. So this is the agenda. I think we put very interesting things together. And with that, thank you again for joining us. We appreciate your engagement, your partnership and your trust in GFT. We're excited to show you how we are shaping the next chapter of AI-centric digital transformation together with our clients, partners and all of you. And having said that, Marco, I would hand over to you. The stage is yours.

So good afternoon. Welcome, everyone. Thank you very much for joining us here in person, and thank you very much for those who are joining virtually. So let's get started. We have a pretty full agenda. And I would like to start with a video, a video that our marketing team created that is helping us on bringing the positioning of GFT. Let's take a look on this video. [Presentation]

So that's GFT. Let's go beyond the hype. What do you see out there? Too much noise, too much hype. We create impact with artificial intelligence. We believe on that, and we are doing that. We are the artificial intelligence digital transformation challenger. That's our positioning. And I'd like to also share with all of you some personal insights and learnings that I had. So I met 2 brilliant minds on the artificial intelligence industry. One of them is Denis Hasabis, the Google DeepMind CEO and Founder of DeepMind, who takes care of all artificial intelligence projects at Google on a global level. He has 6,000 data science, AI, artificial intelligence engineers at Google. Everything is under him. And I met him in person in the United States, in California this year in the second semester. And it was amazing to see the way that the artificial intelligence is evolving with building blocks of modules like Gemini and other models, large language models, small language models and elements and agents that they are building on top of each other. That is exactly the vision that we have for Wings, and we're going to talk later on about it. I also met in person in Berlin in Germany, actually 2 months ago, Sam Altman, which is the 1 mind and person, individual in the road in artificial intelligence. I met him in the [indiscernible] Springer event awards for a dinner and then for a workshop in the day after. I met -- I talked to him in person. And I could talk hours here about my learnings, right? It was an amazing opportunity for me, for GFT. And one thing that stands out, the artificial intelligence, the progress and evolution of artificial intelligence has been material this year in 2025. Things are moving and evolving so rapidly. And he, he believes that in 3 years from now, there will be so many change, so many things that are difficult to predict. And obviously, he's working 24/7 to be leading on that. And on the same way, I am working 24/7 to also lead there with artificial intelligence for our industry. But that's the good learning. Things are changing so fast and we're going to see the strategy of GFT that is to tackle these transformations. So this is -- for me, it was an incredible learning. And now taking you to what's going on, trying to bring a real understanding about the transformational impact of artificial intelligence into our industry, information technology, IT because there are a lot of people talking. When you talk to investors and you talk to analysts, when you talk to clients, lots of questions and lots of inputs, feedbacks. I'm going to bring you not the hype, tangible results. So let's try to understand this. And to do that, I usually utilize a quite known analogy, the good and bad printing revolution. Before this, before the good and bad printing revolution, we had -- we were writing books with hands for thousands of years, [indiscernible] were writing books by hands. And with the good and bad printing revolution, we created a complete change in the way that we think, that we learn, that we create, that we share ideas, that we scale in mass production, mass product books. And we created an infinite number of jobs. So we scaled it up complete massively. We were able to create universities because of that. We changed the way that education works. We created the renaissance and the modern age arrives from middle age to the modern age. That was the change, and we created lots of jobs. We even created companies to create the machines for that. We created inventory companies, inventory. We created logistics. We created even companies to sell books. And we created the commerce companies to sell books and then we created Amazon.com to sell books. That was the beginning of Amazon. So it's a massive, massive, massive scale up and changing. And on the same way, the same analogy, that's what's going on now. We understand that at GFT. So what we are doing today, we are writing software with hands, hand writing, hand coding. And with artificial intelligence, we are unlocking completely new set of opportunities to obviously produce things faster, be more efficient, time to market, to reimagine the way that we think, learn and create technology, software and naturally laying down the foundation of a new era with artificial intelligence. This is beautiful. And I could -- we have today with Wink and artificial intelligence around 50, 53 case studies of real impact. I'm going to just go to talk about simple things. One, major financial service, car manufacturing company, our clients, We put -- they started a journey to modernize their core banking loans. We put 2 proposals. We are doing that project with AWS. We put 2 proposals, 1 proposal doing the first piece of the project of the core banking modernization, which is the documentation, just the first phase documentation, 1 proposal, 6 people, 1.5 years. 1.5 years, EUR 1 million, 6 people to work on that. We put a proposal. We put a second proposal using Wink, artificial intelligence, generative AI to do the documentation of the same system, 6 people, 6 weeks, EUR 100,000. The clients chosen the artificial intelligence one. We already delivered. It's not hype. This is a reality. We already delivered that client is happy. And by the way, the deliverable that we generated with 6 weeks using artificial intelligence for the documentation of this has a better quality than if I had 6 people working for 1.5 years trying to document millions of lines of codes. This is groundbreaking. This is artificial intelligence creating impact in a simple case. I could go on and on with more complex case like a regulatory change in Brazil that they changed their social security from numbers to numbers and letters, which is a huge transformation in the whole industry, multi-industry in Brazil, and we deployed artificial intelligence to solve that problem. And we simply wiped out all the competition. We put proposals over the last 3 months, and we won all of them in financial services, in banking insurance, in telco, everywhere because they are using artificial intelligence to do something different. And we are doing something new. It was a regulatory request. And I could go on and on and on. So pragmatic, simple, tangible results. We have -- now let's talk about our strategy. I talked to [indiscernible]. [indiscernible] Altman said, I wish that all companies would move to become an AI-centric organization regardless of the industry. So we are more than ever confident and happy to have designed and have communicated that strategy. We want to be the best responsible AI-centric digital transformation company in the world. So that's our vision. And we have the mission to bring AI centricity, digital transformation for software development for technology services for all the industry. And we are positioning ourselves as the challenger. I won a challenge, and I am challenging the IT -- the traditional IT service industry. I am a challenger inside my clients. I want to challenge and help the clients challenge the status quo, the legacy systems that they are there. The coherent ecosystem of partners that are doing those legacy maintenance for 20, 30 years, doing the same thing, super hard to change, complex tool to be -- to deliver new functionalities, costly. We want and we are the challenger on this. And that's why we are positioning as the artificial intelligence digital transformation challenger. Our 5-year strategy. So it's -- we talk about the vision mission position. We have strategic goals. And the strategic goals are: number one, fast learning, adaptive and an innovative organization. As some Aman told us, as Denis, if you are not this, if a company in the IT industry are not a fast learning and adaptive organization, it's challenging because we need GFT, we need to be prepared for the next change that happened in 1 year from now, 2 years from now, 3 years from now, whatever, something new come that replaces OpenAI ChatGPT or replace completely Google, whatever, we want to be positioned on the forefront of that shift of that change. That's our position. So that's our strategic goals. And naturally, an agile company that orchestrates our job -- our work with agentic work naturally global AI-centric market leader and democratized AI as well. And we want -- so that's my commitment. We want to create a company to last on these super changing environments that we are information technology. I want to believe -- to have a company in 3 years from now, 5 years from now that is ready for the next shift, the next leapfrog, the next wave of opportunities. And so vision and mission, strategic goals that are long-term goals to last, and we have our ambitions. We have ambitions, and we have ambitions and bold goals. I want to almost double the company. on our long term. That's our ambitions, and we have that ambition to make our revenues EUR 1.5 billion to improve considerably our margin to 9.5% EBITDA adjusted. And to do so, we need to increase the high value-added service that we have, transform everything that we do in AI with AI, high value-added services and also improve our cost structure with more nearshore and offshore. We have several strategic initiatives. I'm going to -- we are going to -- myself and Jochen are going to touch. We're going to cover the great majority of them, but we are not going to cover today, but just want to mention, I want to, and I strongly believe on that and the team also strongly believe on that, to globalize more the company coming from a federated silos company to a more global organization. So that's a strategic issue that we have that is changing many dimensions of the company. It's moving. Naturally, we're going to transform ourselves as an AI-centric organization on the software development with Winks as our spearhead and also all the other services around it. We are also focused on positioning ourselves into the market, positioning as the AI, artificial intelligence, digital transformation challenger with the next-generation brand design. We launched our new brand and so on and so forth. We also won the generative AI AWS award. We also won many other elements of Gartner positioning of generative AI and so forth. We -- and we naturally want to keep innovating, investing in asset creation as we did with Winks. We want to grow Winks. We want to create more Wings and the other strategic initiatives that we're going to cover right now. So let's talk about strategic initiatives for sustainable revenue growth, areas that can bring large-scale opportunities for us. First, implement our global delivery platform focused on smart shore and startup in India. So that is one important fundamental elements of our strategic for the 5-year strategic plan. And it is simple. Before GFT, we had some delivery centers and the delivery centers were federated, also connected with one market and another one with 1 or 2 markets. Now we are harmonizing everything, bringing that into a big platform of delivery. We have one central demand leader for that, one center supply of all the delivery centers in a global platform. And every proposal that comes across the globe for a project that is large-scale project that needs efficiency, brings differentiation skills, we use that to deliver the proposal and win the deal. And that's what we did successfully this year with [ MAPFRE ], one of the largest insurance companies in the world that we started -- that participate in a strategic process with them, RFP. And we won because we were able to articulate our delivery capability from Colombia, from [indiscernible] solutions, the acquisition that we did from Colombia, and it was a requirement on the RFP, and we were able to deliver that combined with other -- with Brazil and other centers of excellence, and we're able to win a strategic project at [ MAPFRE ] this year. Fantastic result of this strategic initiative. the result of strategic initiative, we also won this year a strategic project in robotics here in Germany. And it was a project that required a large-scale team of data AI and cloud, impossible to do that with the 1 or 2 centers. We had to bring all this platform together, Poland, Spain, Brazil and India to put a proposal to bring such a large-scale [indiscernible], we won it and you deliver that. We have the same -- on the same message, India. We started up India for -- I'd say -- I cannot say the name, but I'm going to tell you one of the top 2 largest retail banks in the United States of America, top 2, one of them, we were able to stay on the game on the core banking implementation that they are doing. And it's their requirements. You need to have a footprint in India, a delivery center in India. And because we had India, we were able to be part of the game, participate and we won, and now we are growing on that client with a next-generation core banking project. Without India, last year, 1 year ago, 0, not possible to compete, 0. Reality, our 5-year strategy. And I tell you more, 2 weeks ago, we signed a new master agreement, service agreement, MSA, with one of our top 4 clients in the world, GFT with India rate cards. So we've been operating with this client for several years. And we have been -- this client look at GFT as the European shop, and European partner. We unlocked India now. We unlocked a new world of possibilities because there are a new set of programs and services that are done with an India platform. And now when we are 2 weeks ago, we signed the master agreement for the first time becoming a European regional company, service provider to a global one. And not to mention about our AI strategy, 12,000 engineers, 1,200-ish are already working with [indiscernible] projects. Most of the team, 96% trained with AI tools and so on and so forth. A second strategic initiative from our 5 years plan is to first move everything that I do, this is soft application development, software engineering to soft engineering AI, the AI-centric thing, number one. And we also have some other, let's say, core elements of our high value-added service offering that we are putting focus. One of them is the next-generation core banking, which stands out for us. We are the #1 in the world, delivering thought machine projects, delivering 10x, delivering [indiscernible], delivering Oracle Flex Cube and now a new partner, which is Visa [indiscernible] that I'm going to mention a little earlier -- later. So that's next-generation core banking. GFT, #1 in the world, very well positioned and moving forward. Also on the front of the ISVs like Salesforce, like Guidewire that we're going to talk later on with a client, [indiscernible] and SAP, a new element of ISV specialization that we brought on our system, on our ecosystem. And by the way, we already sold, and we announced that last week on our Q3 announcement, already a new project with SAP in one of our clients. Hyperscalers, AWS, Azure, Google, Clouds and our products, SmartAC, Ingen, and Winks. And I'm going to put highlights on AI data intelligence and artificial intelligence modernization in a second. Just to complete, we have our products in Smart Act is our anti-money laundering compliance solutions, and I'm going to announce that now. We are going to have Deutsche Bank here formally coming into picture and talk about their decision to move with Smart Act at Deutsche Bank Germany. This was a competition that we participated, and we won that, and Deutsche Bank is going to come and talk in a few minutes about that. This is very good. Last week on the Q3 announcement, I did not mention any of the clients because we still do not have the approval, but now we have the approval to do that. This is fantastic news for our software solutions practice -- unit, sorry. We also have Ingenia and that we also communicated last week that we announced that we closed a 3-year deal with Audi, Volkswagen Germany and Wink that is -- I'm going to -- that is a chapter that we're going to talk soon. And among all those partners, 2 new ones, which is Visa [indiscernible] with the most advanced core payment systems that are coming to play. GFT is already doing some projects with them, and we are creating a global partnership with Visa Pismo and also with FICO that we're going to have Alex Graphi soon talking about FICO and why FICO chose GFT as a strategic partner. So for us, it -- I could not be more proud to have FICO, the largest rating scoring consumer company in the United States of America and U.S.A. and one of the largest anti-money -- anti-fraud company in the world, the whole world that are choosing GFT to be a strategic partner. Moving from the strategic initiative of focus on high value-added service and offerings, let's talk about our global accounts, the strategic initiative that is focused on what is going to move the needle, Tier 1, Tier 2 clients, global accounts. That's -- we have a strategic initiative that is moving all GFT because I want to -- that's our current footprint. That's our current, I would say, wallet share of clients. We have 4 clients today, more than EUR 25 million, 20 clients more than EUR 10 million, and then I have almost 40 here, Tier 3 clients, more than 500 clients Tier 4 that generates less than EUR 2.5 million. And I tell you that in several areas, in several units and markets, we are putting more efforts here than I'm putting on here. So moving the company, moving GFT to focus on what matters. Let's focus on Tier 1 and Tier 2 accounts. Let's focus on those guys here in order to bring my offering, industrialize my offering, my capabilities, by differentiations to move those guys, have more Tier 1, have more Tier 2, move Tier 3 to Tier 2 and be smart to focus on the Tier 4s that can become a Tier 1, but the laggers, those clients that are requiring more energy than other accounts with low profitability. This is definitely -- there's no space with GFT, right? We are going to simply put the right focus. And so that's -- this is simply a glimpse about some of our clients -- and you see the clients that we have, HSBC, Citi, Deutsche Bank, Itaú, the largest bank in Brazil, Latin America, BBVA, we are the one of the key partner on modernization of BBVA, one of the benchmarks of technology in Europe. We need to grow here and many others, right? You see -- and many others that you cannot mention this one of the top 2 largest retail banks in the United States of America and many others here that are non-open logos for us, yes, yet. And what we want to do, I want to focus prioritization, industrialize the process of taking my offerings, my differentiations, my capabilities and bring to them to grow the -- bring up my -- also my capability, my strategic initiative of Smart Shore and my offering of India, bring to them and bring and then create spending a virtuous cycle, engineering growth, engineer growth accelerate growth into my business. And naturally, not only bringing growth to the current Tier 1, Tier 2 clients that I have, take the Tier 3 to become Tier 2, bringing all the capabilities, differentiations that I have into the clients that I have, industrialized process and obviously, engineer profitability because on that process, I'm going to engineer in profitability because I want to get rid of a Tier 4 clients that are clients that cannot take us to more than EUR 1 million, never. And we have low profitability or we have not a nice profitability. So we need to put the right energy on the right things, on the things that are going to scale. And I do believe that this combination of strategic initiatives creates and can create and will create a large-scale growth. You just -- if we get -- if we bring and we make new Tier 1 clients paid year over the next 4 years, you can get 8 clients, you can double it. And so on and so forth, right? So we can really -- so that's our goal to have those -- obviously, those percentages are approximately percentages, but our expectation is to have more of the company around 70%-ish on the Tier 1 and Tier 2 clients on the long run and have less energy, less dissipation of energy and heat here, right, on Tier 4, be smart, be smart here, right? So that's what we are now industrializing, pushing forward, pushing forward, pushing forward, and that's a combination of strategic initiatives. Now let's getting back to the high value-added service offerings and differentiation. I mentioned about -- I touched a little bit on the next-generation core banking on the ICVs, hyperscalers and our products. Let's talk about artificial intelligence modernization. This is immense. This is huge. This is huge. Artificial intelligence modernization. AI is breaking down, steering down the barriers of changing legacy systems across the board. These unlocking opportunities that were not there before. There were no opportunities before to change legacy for some legacy core banks. The barriers of change were super high. They were there for 25, 30, 35 years to change that programs, endless programs, risky, risky, enormous cost. Now with AI, we are making modernization projects feasible, affordable, less risky, impactful and transformative. 62% of the U.S. companies, they rely on legacy modernizations. 75% of the banks on a global level run on legacy, and they are failing to deliver what they need in order to deliver their goals. HSBC estimates 5,000 applications need to go to modernize. 86% of the banks executives say modernization is critical. This is huge. And we are the best company positioned to do the artificial intelligence modernization, this artificial intelligence accelerated journey, artificial intelligence accelerated modernization. We have a complete set of offering capability units with since the strategy to the top-down approach, bottom up and important to highlight Winks, with an agent, a functionality that is our legacy transformer that is moving forward in the market, bringing 40% of efficiency at minimum with the example -- one simple example that I mentioned before. Artificial intelligence modernization, we need to deploy AI. We have our technology, our differentiation. Winks, important to highlight, we're going to talk more about Winks, clients improving this year from 25 to 58. We expanded from 1 country to 8. Bradesco Insurance, we're going to talk here, 40% improvement in terms of productivity with a large-scale team. It's not a toy POC. It's a real 180 team working, large scale, delivering development and modernization and question, does or did Bradesco Insurance ask GFT to reduce the team in 40%? No. We are growing the team now I have more than 200 because capital is smart. Capital flows to where return is higher. Capital flows it. So if I deliver with more productivity, I am going to get projects from the competition. That's what the effect that's happening at Bradesco Insurance, and you can have the clients talking in a little while to all of you. And staggering figures, right, in terms of documentation, of code fixing and et cetera. I could not be more proud of the Winks team, what we have done with the creation of that technology, what we are doing on a global level. I could not be more proud of that positioning. Finally, AI and data. AI, data and intelligence, that's a major practice that we have. We have 224 clients, more than 1,000 projects that we have already delivered, more than 2,200 AI data specialists, 6 global centers of excellence and leader at Gartner 2025. We have a complete unit with enablers offering, cases, accelerators, services, differentiators, including the ADA marketplace. [indiscernible] is going to talk later on. And naturally, what are some logical steps for GFT with this? Some logical new industries where our experience capabilities offering differentiation and even our roots as a German company is in high demand today. So we got it. Robotics. We -- GFT has been selected as a strategic partner for Neo Robotics to develop their software platform powering the next generation of physical AI for [indiscernible], publicly referenced. We did that large-scale team for GFT and Nora. And that's the quote of David Zager, the Founder and CEO at Neo Robotics. With GFT, we have found a strategic partner that shares our vision of bringing cognitive robots into a real-world application with a deep knowledge in AI software engineering, that's our differentiation, and complex and highly regulated industries. And we made it. And obviously, Again, that's a capability that we have, it's in high demand. Warfare is movingfrom hardware to data, to software. So since 1987, we've been delivering complex projects, mission-critical projects, high volumetrics, security sensitive and for highly regulated industries. That's GFT since more than 30 years. And we've been pulling into several discussions right now with defense companies to help them what? To discuss about how to modernize. We need to modernize. We all need to modernize, and we need to help us to accelerate the digital transformation with AI, data software. And GFT plays as -- clearly as a German champion for that, German champion in AI data and modernization. This is -- in this transformation. This is something that, it's coming. We are discussing now a major projects in defense in Poland. It's a pitch. I'm discussing right now 350 deal size, 350 FTE size in a defense company in Germany right now. It means that we're going to win that? No. It means that we're going to win 100%? No. But this is wonderful to have that opportunity. This is wonderful because we have a strong capability, a global capability, international data, AI, cloud, and this is in high demand for Europe and for especially Germany and all Europe, right? And I think that we come into play with a good [indiscernible] all that. So Jochen, I would like to invite you on stage. Otherwise, I keep talking the whole afternoon.

Thank you, Marco. Warm welcome from my side, everybody in the room, all virtual participants. I will talk about 2 more work streams of our 5-year strategy. I will talk about the gravity program and our M&A strategy. It takes about 10 minutes. Gravity program is a self-chosen title. And what does it mean that's the headline, focus on simplifying and optimizing countries, offices and shared services inside the GFT Group. We have some KPIs here. They are going to repeat on the next slide. Let me highlight them there. And let's look at what we have done in the year '25, '26. I'm looking at the bullet points on the left side right now. And so we are doing one thing in all countries, which is we are reviewing the clients, especially the subscale revenue and margin clients and their contribution. Do we want to work with them in future? That's the Tier 4, Marco was already mentioning. We're looking for major accounts, and we have to be critical on all the Tier 4 clients, which are the very, very small clients. Second bullet point, review subscale countries. We're currently in our budgeting process for the next year. We're going to have our reviews end of this week. And there we're going to discuss also our small countries, the subscale countries, the countries which are not generating economies of scale as of now, but we need economies of scale across the GFT Group in future. Third bullet point, optimize G&A cost, cost base and location. Well, of course, we do what everybody does. We simplify, we automize and if it can't be simplified or automize, we move it to cost-optimized locations. And the fourth point, reduce office space without -- with low utilization. GFT is following a very strong work from anywhere, work-from-home approach. But you see this office, it's pretty full when you walk the aisles right now. This is where projects are happening, where senior sales is happening. But a lot of the coding our people are doing must not be done in a GFT office. Therefore, office space, we're critical about and we reduce where we can. And now what are the proof points of the year '25, '26? And you see that on the right side, the U.K., we had to reduce by 50% in workforce in 2025. Well, there's nothing to really be proud about because we were not able to staff our people in projects locally anymore. After COVID, there was a high demand for local resources and the demand diminished again when the work bench from Eastern Europe and India was working well again, demand reduced for local resources, and we had to rightsize our U.K. organization. This year, 50% down, productive people, but also in the back-office teams, we had to rightsize the team. I'll come back to U.K. on the next slide. Software Solutions, our second turnaround topic of the year 2025. We went into the year with a too strong workforce. We brought on board a strong manager. Floris is here with us today. We reduced the workforce by 15%, and now it better fits the revenue structure of the business. So we are set to go into 2026 on a better cost base. We will still invest in that business, but the cost base is now far improved. Third bullet point, we have reduced our global G&A workforce by 6% in 2025, something that is still ongoing today. So this is work in progress, but it's our target to get the right size. And people we can't simplify or automize, we have to move to low-cost locations. That's happening right now. Strict cost control, the fourth bullet point, I will not go deep, but office space, we reduced in Germany, Italy, Poland, Brazil over the last 12 months by 7,500 square meters. This office here is roughly 1/4 of the size of the former office, which was around the corner, while we didn't manage it down. In this case, we had to wait. Sometimes you have to wait. There was a long lease contract, but we had taken the office on board before COVID, and we learned we need less space. So we have reduced globally office space by 7,500 square meters. Now at the bottom, the 5-year strategy about this, the mentioned cost effects on the right side, they will all materialize next year. This year, many of them have been more a cost than a benefit. The benefits will come in years after. And until '29, we believe we can improve our EBIT adjusted with gravity program work by at least 1 percentage point. And let me highlight again the U.K. story. In the U.K., the gravity approaches, we leverage what works best. And again, on the left side of the slide, we are replicating in the U.K. our blueprint from the GFT U.S. And the smaller bullet points say it, we have a very efficient local team in the U.S. today. It's roughly 40 productive people, plus sales, plus administration in the U.S., and they are mostly managing the 95% offshore delivery that we have. This generates a mid-teen margin today, which is comparable to what peers generate. Simply, our U.S. business is only 9% of total GFT. This blueprint, we're bringing to the U.K. right now. We're revising the go-to-market strategy, the leadership and the organization. We're implementing a new country manager. He will be on board somewhere in the first quarter, and the new governance for sales, operations and delivery are underway. So outlook for next year, this year, '25, but also '26 are transition years. '26, we will already see the margin improvement. So we should see a black breakeven for the U.K., and we will return to revenue growth in '27. I promised my team, I will only bring one financial slide today, right, only one because this is all about the business side of GFT. This is the financial slide, and I'm trying to highlight the isolated GFT Group performance of the non-turnaround countries, which is by far the majority of the GFT Group. First, the 2 turnarounds, GFT U.K. and Software Solutions, including the numbers and the decline on revenue and profitability basis. We have a third enemy this year, which is FX. The euro is so strong that we have a lot of headwinds, especially on the revenue side. It's also burdening the profit. But if you take out these 3 effects, GFT in constant currency is growing by 11% this year. This is happening on the back of a very strong Latin American market, of a strong North American market and a somewhat weaker and not growing European market. It's already the blended growth rate of these 3 markets, but excluding U.K. and Software Solutions, 11% growth, and we are at a 9% EBIT adjusted margin. So this is where we can take this. Turnaround programs happening under management will take 1 or 2 years to get closer to the group numbers in those turnaround markets, but we are on track for that. Now M&A strategy. Come on in. Take a seat, please. M&A strategy, we have also refined with the 5-year strategy. And again, the headline is highlighting it. We focus on bolt-on acquisitions of high-value-added services company in existing GFT markets. On the bottom, you do see the history of our M&As. We've done 13 M&As over the last 14 years. So we are quite experienced. We've done them in Europe, North America, South America. But now the strategic focus is a bit changed. It's on the top right. Country-wise, we stay in our existing markets. We don't go for M&A outside of GFT existing markets. Gravity program, focus on markets, maybe even reduce the markets, and we focus M&A on the existing markets. Second, technology, unlock new ISV solutions and business to strengthen the core. Marco was talking about ISV a lot. M&A strategy has to follow that. I get to the example on the following page. And clients, of course, we want access to future Tier 1 and to Tier 2 clients via M&A. Current priority list on the right side, on the top, it is North America. Again, as I said, we have good margins in the U.S., but it's only 9% of the GFT business. It should be 30% of our business if it would be like the market share that the markets globally have from the U.S. North America, high focus for clients, more clients access. We all know Americas is not cheap. So that's a challenge. The second, Latin America. Here, our strategic approach is to buy ISV companies and by that, also get their clients. And number three, a bit more opportunistic in Europe, also looking for ISV clients, so software implementation product companies with their client base. That's the fundamental M&A strategy. And now the one example that we implemented in the year 2025 is the company Megawork in Brazil, an implementation partner for SAP software products. Megawork, now more on the left side, is a EUR 15 million, EUR 16 million a year company right now. We are consolidating since September. That's why we only will show EUR 5 million of revenue this year from Megawork. They do a 20% EBIT adjusted margin. So ISV business is highly attractive margin-wise. And it's roughly a 300 people team. And how does it fit into the strategy? We focus on existing GFT countries. Well, the client -- sorry, the M&A target is in Brazil. So obviously, we're following the strategy. We will later roll it out into other existing GFT Latin American markets, potentially Colombia, Mexico, and maybe even beyond into Europe, but currently focuses on Brazil. The second M&A strategy, expanding client base, cross-selling potential. Well, in this case, we have acquired a company that has a diversified client structure in health, pharma, public sector, utilities, manufacturing, and we want to cross-sell GFT services into them. But also the other way around, we want to cross-sell the SAP expertise into GFT client base. And this has already succeeded in 1 case after roughly 4 weeks after buying. That's, I think, a record for GFT. Number three, focus on high value-adding services. While it is an SAP product implementation business, we're showing the margin over here. So that fits. And last but not least, the GFT AI portfolio that Marco talked about a lot, should support the M&As. While in case of SAP, we still have a bit of work in front of ourselves, but we want our Wings suite to also support SAP services in future, something to work on for the next years. So implementation of the M&A strategy at work. That said, Marco, conclusion is for you.

Thank you, Jochen. So conclusion to go direct to the point, we are executing our 5-year strategy with a mission focus and to be the artificial intelligence digital transformation challenger as I presented, as I mentioned to you. We are engineering growth and profitability while also focused on initiatives such as the acquisition of high value-added service companies in current markets that we have. There is a clear path for large-scale growth opportunities in several fields, as I mentioned to all of you, the artificial intelligence modernization, the AI data and intelligence, the robotics, the defense, our core banking part and ICVs and clouds, but especially on the artificial intelligence modernization with an immense huge opportunity in terms of legacy architectures modernization. And the path ahead is bold, is ambitious and full of opportunities. GFT is ready for this -- to lead on the AI-centric digital transformation, right? So that's what we are positioning for. I think that we've been moving strongly as I presented before. And I would like to thank you for the time and open for some questions. I don't know how we are with time, but go ahead, Andreas.

Okay. That one. Well, thank you very much for your update and almost staying in time, almost. I think before we proceed with our tech expertise, I would allow one question to keep the dialogue a little bit flowing, but then maybe move the other questions to the coffee break. So if there is a question in the room, we have one, maybe 2.

Yes. One question to the recent acquisition and the SAP topic you mentioned. Is there any risk that you get in conflict with SAP's own AI strategy?

On the contrary, one of the reasons of the acquisition is that SAP has a strong AI and AI strategy for their products. And we want to -- we do not want to do the traditional, the vanilla SAP. We want to definitely be the leading on this AI-centric transformation of SAP with the Agent platform that they are doing, number one. And number two, to also use AI to accelerate their migrations from the ECC to SAP/4HANA. So that's for us, we believe that is a strong opportunity. And we already started cross-selling, right.

Are you talking to SAP on that topic?

Yes, absolutely. Absolutely. I met to -- with some -- in the event of Sam Altman, I met one of the Board members of SAP here in Germany at a global level. And we are talking in many countries, not only in Brazil, but in Spain, the Global Head of Partnership, which sits in Spain globally for SAP. So we've been, let's say, active discussing with them. Very happy, very happy with SAP. And by the way, for me, this is gold. We simply brought SAP an ISV. It's a global SAV brought into our ecosystem. And then we created a cross-selling immediately. In less than 1 month, we already sold SAP into one of our clients that we have. So for me, this is -- can really -- our expectation is to do more than the business case that we bought put on top.

Thank you. We have another question in the second row.

Yes. I have got one question with regard to the focus on the Tier 1 clients. I remember that you pursued very successfully this land and expand strategy. So the question for me would be, is that now that? Have you changed the strategy with regard to that? Or how should we think about that?

Yes. I think -- may I, if you want to take it, I believe that the land and expand strategy was a strategy that took GFT from EUR 100 million-ish, right, several years ago to the size that we are today, EUR 885 million. That's our guidance for this year. And I think that was quite, I would say, important and relevant strategy for us. But what's the situation now? The situation now is that I have a quite interesting and strategic client list of Tier 1 and Tier 2 clients, as you saw. I also have a lot of more than 500 clients on Tier 4. And now I need to rebalance the focus of the company because I'm clearly investing the same time and effort of our sales team, marketing, presales, technologies, architects into these long tail Tier 4 clients. I'm investing the same amount of effort in clients that you look at the clients and say, wow, these clients will never give me more than EUR 1 million never because there is no size for that. And by the way, my margin is not that good. So it's no-brainer. Let's invest the time on the ones that matter that can move the needle, Tier 1 and Tier 2 clients. And obviously, let's be smart to keep doing the land and expansion, but smartly land in specific clients that can be a Tier 1 and then have a right plan and measure that every quarter, every 6 months, every year and see the plan evolution to take it from Tier 4 to Tier 3 to Tier 2 to Tier 1. That's the mission, but focus on what matters because if I double my Tier 1 and Tier 2, I can double the company. And they have scalable right and capabilities. Obviously, it's not simple, it's not trivial on that journey, we're going to lose some. We need to win others. But that's our mission.

Perfect. Thank you. I think there will be more questions in the coffee break, and we will have also the opportunity to talk about other questions later on. So thank you very much for that update on our 5-year strategy and showing us how our 5-year strategy is taking shake and for demonstrating how AI is breaking down the barriers of change in legacy systems and unlocking new major opportunities. And we exactly take it from here where we now move from strategy to execution to the technology that is already transforming client projects. Today, please welcome Ignasi Vilardell and later on, Dean Clark, who will take us deeper into the world of agentic AI modernization with Wynxx. Ignasi, it's yours to have.

Thank you. Thank you, Andreas. The first statement that I want to say is that GFT was founded in the '87, right? And more than 10 years, we started doing AI projects. So we haven't jumped into the AI wagon last year so we have more than 10 years working around data and AI projects. One thing that I want to share with you is a set of examples, some of them being spoiled by Marco at the beginning. So I'm going to try to speed up a little bit that part. But I also -- I want to share with you that whenever a company is adopting AI, the impact of that adoption, it is different. We've seen different flavors and colors. And the reason being is because it is impact, it is not uniform. That is the reason why at GST, we have a very, I would say, concrete offer when we are going into the market to help our customers to do this AI adoption in a very measurable and impactful manner. How are we starting this type of projects? The first thing that we need to understand and that we love to demonstrate the value with measurable impact. It comes for individual productivity right there. This is automating individual tasks that you have to do in your daily routine. And as we do that and we are accumulating experience, then we are ready to jump into the next step, which is precisely improving the process-based productivity, which is automating entire workflows across different business units. The importance of having this stage, let's say, approach is because companies require time to get confidence on what is the possibility of AI to also change and adapt the cultural impact that represents the AI adoption and at the same time, to start realizing that this exponential ROI that AI will bring as long as they move into an AI-centric enterprise, it requires this step-by-step approach. So this is what we are currently doing. But of course, at GFT, we are determined to go to the next growth phase, which is this AI-centric enterprise. Every lessons learned, every project, every success that we do, we make sure that we capitalize those and we make it tangible in our set of products, Engenion, Wynxx and Smaragd to continuously helping our customers. If we have to think about AI adoption, the most, I would say, immediate area where we should adopt AI is a very, I would say, mature arena. It is precisely the software development, right? This is the area where AI has been largely adopted over the past years. Quoting Gartner, by 2028, 90% of AI software engineers will be using AI. That is a tremendous jump from 2024 -- sorry, here you go, only 14%. There is a massive jump of AI adoption in that field. And GFT, we are not just, let's say, being watching what's going on. But on the contrary, we are [ actionating ] this trend by helping our customers with Wynxx to generate better documentation, to generate better code, to generate better test and also helping them to maintain their systems. This is clearly a trend that GFT is capitalizing and that we are helping our customers across the board. Now as I know that you would like to see some examples, I'm going to be covering some of them. The first one is in Mexico. It's a Tier 1 multinational bank that you see more or less here the rationale of these cards is we see in blue, let's say, the motto of the project. In that case, it's testing automation at scale. And then we have the pre-AI baseline, so the customer not using GFT services, not using Wynxx, which is our AI-based product to accelerate SDLC. And then on the right-hand side, we have the results after adopting Wynxx, after engaging with GFT. And here, what we can see is that we have decreased the amount of time required to do test execution from 5 days to 2 hours. This is a 98% reduction in the times. Similarly, if we look at the human-based intervention, we see that we decreased from the 75% human intervention in the test execution to only 5% of manual execution. Going to another customer. We're still in LatAm, one of the largest customers and banks that are operating in Colombia. We are seeing more or less the same behavior, right? But in that case, not only in testing, but also in different steps of the software development life cycle. Generating code, we reduced 72% of the time required to do so. Identifying code vulnerabilities, we reduced 88% of the time required to do this check and to fix them, which is the next step. And the same for code review, generating user stories, et cetera, et cetera, et cetera. This is changing completely the landscape for our customers whenever they have to produce code at scale. And the last example in this slide is in North America, another Tier 1 institution, and again, we have a baseline of lack of documentation, very slow process to have coverage on the code, testing, et cetera, et cetera. And using Wynxx, we are able to improve dramatically the entire situation. So in a nutshell, in just these 3 examples, we are seeing how Wynxx and how GFT is improving significantly the times, the quality and the coverage that we are providing to our customers between 70% and 95%. But the most important thing that we need to highlight as well, that's a message that I want to share with you is that the more that the customers are engaging with GFT, the more that Wynxx is used, the better results that we are obtaining, the better efficiencies that we are generating. And if you don't trust me, because we never met, right, later on in the process, Bradesco Seguros themselves are going to be sharing that story with us. I'm going to spoil myself a little bit, but the story is in early 2024, the customer launched the challenge to all their IT providers saying, guys, we need you to deliver more business value with the same budget. And the result of that challenge, the GFT was the winner party, achieving the result of 40% efficiency and productivity and 80% of faster code correction. Again, later, we are going to be -- we are honored to have Bradesco Seguros themselves, one of the largest Latin American insurance companies to share with us this project and these initiatives. So as I said, we started in the software development life cycle. But if we have to think about what's the next step, immediately, we think about business processes. And this is precisely what we've done. So back-office processes, middle office processes, front-office processes. This is where AI could help all these financial institutions and industrial companies, insurance companies, et cetera, et cetera, to automate many of their functions. Some examples could be document management, where we can apply AI and actually, we've done it, credit risk assessment, customer service, et cetera, et cetera, et cetera. I know that you like to see some examples. So let's go for it. The first one, let's think about a mainframe system. This is a very critical infrastructure. And every single minute of downtime, it is actually affecting that the customer is not able to transact, right? It's not able to buy, it's not able to pay, you name it. So definitely, reducing the times required to fix incidents, it's critical. It's key. And this is what we've done at GFT. We take an existing service. We download all the data from the customer, of course, with their permission, how these tickets -- how this incident has been solved over time. And we generated this engine being able to help the operator that whenever new ticket, a new incident is coming to the system, he or she can expedite the resolution of such kind of an incidents. And after just 3 months that we put this system in production, we reduced the average ticket resolution time by 30%. And this is precisely what you can see here, the SLA times going down and down and down and down. And I know that you like also to see some proof -- graphical proof of that because another thing that we do at GFT is not only training and putting together all these AI engines, but also bridging the gap between the technology and those people that are not technology, but they need to use that technology. And here, you have some screenshot of this solution that is already in production, bringing that value for different customers across GFT. This is not just operational efficiency. This is strategic leverage because of that, now whenever we are pricing a new application maintenance for our customers, we can be more competitive. And those existing engagements where we are doing maintenance, we can also open the door for potential margin improvement. So this is what AI is unleashing for us. Time is very sensitive, and I know that I'm actually running out of time already perhaps. But it is even more important when you are submitting and creating credit risk analysis. A credit risk analyst, whenever they have to create this report, what they have to do is some of the -- all the tasks that you see here and beyond that. But this is just, let's say, a summary of some of the tasks that they have to do. And as you can see, they have to do documentation, they have to gather information, they have to fill out some sections. They have to answer some qualitative and quantitative KPIs, et cetera, et cetera, et cetera. Together with Oliver Wyman, we create a pilot with one of our customers where what we said and what we did is creating an agentic approach that some of these steps could be enhanced with the use of AI and therefore, reducing the time required to do this task. On average, we said that we can save between 30% and 40% of the times required by the analysts in conducting these credit risk reports. And for that particular customer that you see here some key figures, how many analysts, how many customers, how many new customers, et cetera, et cetera, because that, of course, influence the business case. It means that in a single year, they can save between EUR 4 million and EUR 5 million. That is in a scenario where the customer is not growing. If the customer is willing to grow, which, of course, they want, that savings go much beyond that number. And since I know that you like an image better than me talking, this is actually the solution that we have implemented that we are pitching and that we are using as an accelerator to help our customer to achieve these goals. So this is not only about being able to deploy AI, but doing it at scale for enterprises. Last but not least, we talk about SDLC, right, AI for SDLC, software development life cycle. We talk about the incursion of AI into business processes, back office, middle office, front office, but we are not going to be stopping here. For us, the next frontier, the next jump, as Marco has said also at the beginning, is the capacity to go towards this new arena, which is a tremendous volumes of opportunities for GFT, which is the large-scale modernization. We've seen over the last, I don't know, 10, 20 years, many of these initiatives repeatedly stalled not because of lack of vision, but because of the prohibitive cost of delivering such kind of a project, very long projects, a lot of risk, et cetera, et cetera. I'm talking about projects such as mainframe modernization, reengineering of machine control systems, you name it. Nowadays, GFT with the use of Wynxx, we transform the forbidden type of projects into something that we are actioning, we are delivering. And again, following the same structure that I've been following across the presentation, some examples. The first one, Tier 1 multinational bank in Spain, mainframe, core banking, more than 40 years of a legacy system. And the important thing, I know it's a little bit tiny here, but more than 428 million lines of code, more than 347,000 programs, more than 4,000 functional groups. All this entangled altogether. Now imagine that you're an engineer and you have to serve in to understand all this complexity and to determine what is the best strategy to migrate this into a new banking architecture. And I'm going to tell you, you cannot use AI at all. Not possible, right, unless you have a very deep pocket and a lot of patience because that will take ages. GFT with the use of wings, we are transforming this opportunity into a reality and not only for one country, but as you can see here, we're doing this for 2 countries at the same time. This example was already shared by Marco. Tier 1 multinational car automaker happen to be German, but in that case, operating in Brazil. similar complexity, right? We're talking about core banking, 2,000 programs, more than 200 databases, more than 3 million lines of code. And again, you have to understand everything with no documentation whatsoever to then establish what is the migration strategy, the modernization strategy. Marco said before some numbers, right? You will see a variance of the Wynxx. And the reason here is because we needed for them to deploy the LLM. This is a more technical aspect of it. But even though this is a 69x reduction of the original project without using GFT and AI for less than 10% of the cost. And you may say, yes, Ignasi, but before you were getting that and now you're getting this, don't get confused. Before we will never get that. That was prohibited. That was not even a possibility for the customer. They didn't start this type of project. Now they do, and they are doing it with GFT. And last but not least, I'm not bringing KPIs because we have started quite recently. But you think about industry, think about German champion in the machinery space. And imagine the amount of lines of legacy code that is running inside these machines. And now I don't know if any of you have quite recently went to China, but the competition being created there around machinery robotics is fierce. So definitely, they have now a motivation to accelerate the innovation to shorten the time to market. And to do so, they have to create new type of software to be run embedded into these machinery systems. This project -- this challenge, even though it's not a core banking system, is very similar to this. Legacy systems bear entangled. You need to understand the business logic behind, you need to determine the migration strategy, and you need to do that in a cost-efficient manner. And this is precisely what we are doing. And this is precisely what we are going to deliver out of this immediate project to them, help them to achieve these objectives over there. Now everything that I've said share the same common goal, right, which is helping our customers to run this enterprise AI adoption while we can measure the value that we are bringing to them, which, for sure, I've been sharing with you. But we are not stopping here. For us, as we said at the beginning, agentic is the next frontier. And for us, as part of our strategy is that AI, it is not something that we see as our customers shoot it up, but something that internally we are adopting. What you see here on the left-hand side is our agentic marketplace. So every single process that we have inside GFT at the moment is being analyzed to see what can we do to improve the process and to bring agents to help us to be more efficient internally. And that is something -- let me see if that works now. Yes, there we go. And that is something that we are creating internally to help our employees to get access to those agents to find a way that we can build agents in a standard manner, et cetera, et cetera, et cetera, et cetera. But this is not everything. On the right-hand side, you have our agentic governance framework, which we have created to ensure that internally and to our customers, whenever we are creating an AI system, whether it's an agent or not, and in that case, it is, these agents are aligned to the business goals. So we want to generate return on the investment. Second, we want to do it in an ethical and responsible manner. We want to make sure that it's compliant. We want to make sure that it's secure. And of course, we are building all the technical underlying architecture to make that a reality. I know that you are thrilled to know a little bit more about what's behind the scenes in terms of the technical ins and outs. And for that reason, I will welcome my colleague, Dean Clark on stage to tell us about it.

Hi, everyone. So welcome to this presentation. So the idea that I'm going to talk about now really focuses on a few of the key messages that Marco was pitching to you in the beginning introduction. So firstly, before I move forward into what GFT are doing in terms of its future road map and where we're moving Wins, where we're moving our AI adoption, what I want to do a little bit is just rewind for a second, just understand where the market is, what we're seeing in the market and how we're going to react to that. So right now, today, you can already see a number of organizations are adopting AI. It's now part of business as usual. Everyone is trying to embed it in their processes. We've seen a large number of use cases where it's being embedded in coding assistance like our Copilot, for example. We want to make sure that we are not competing with those, but we are contributing additional savings. And what you've already seen from some of the existing case studies and success stories with our clients is that we are achieving those things alongside the other coding assistant. And that's a key part that I want to talk about at length later. But in order to try and progress a little bit, right now, we need to focus on how AI is already transforming both the world and how it's transforming the possibilities that large organizes and enterprises have in terms of modernizing their estates, moving forward with their productivity, bringing AI into their products, into their services, into the offerings for their clients. And the way that people are already doing that is that they are already seeing AI doing what is traditionally seen as automation really of the low-hanging fruit. So the manual repetitive tasks, the research, the data processing. These are activities that everybody has already been looking at using AI for years. Generative AI, of course, is making that a lot more easy. Generative AI, of course, focuses on that content creation, content summarization. Now in addition to that, we've also got use cases where people are starting to enhance their products, where they're able to personalize customer journeys. They're able to personalize or hyper-personalize user interfaces, customer journeys, be able to target specific products and processes around people's thoughts, wishes, demands. And finally, the third point around how AI is already embedded in our world today is purely and simply amplifying human potential. And really, for me, this is the fundamental concept of why AI is really being adopted wide scale by some of these larger enterprises. And this isn't about how tools are there to replace humans. This is about how tools are there to amplify their potential. It's about lifting that up to the next level so that we can do more with less or ideally more with the same. Now if it will change, there we go. So in the market today, we've already seen that we are moving on as GFT. Marco mentioned already that we have our legacy Modernizer agent. That is GFT's first agent built within Wynxx. Wynxx as a platform is an exceptionally mature product already that sits alongside the software development life cycle, pointing at successfully improving a number of the processes along that route. Now in order for GFT to grow, we need to take into account all of the market trends. We need to make sure that we aren't following everybody else, but we are leading other people. We understand that today's ecosystem, we have a number of our clients already looking at and indeed implementing agents. But there is a challenge today. Marco mentioned at the beginning, there is a lot of noise around AI. How do we rise above that noise? How do we create that level of impact? Gartner coined about a year or so ago, the term agent washing. This is something where existing large organizations created their new branding around their existing platform. They didn't take into account the true agentic need. It needs to be autonomous. It needs to make decisions and it needs to execute something. So what we have seen in the industry are a lot of applications, a lot of software products be labeled as agentic when they aren't. They just purely use AI. Now one of the things that we are doing, and you've just seen it from our agentic architecture that Ignasi has displayed on the screen or at least the governance model anyway. we are ensuring that every product that we build that is labeled agentic will be truly agentic. It will make decisions. It will be autonomous. And we've created a governance framework. We've created a model. We've created architectures and blueprints that not only are we using internally on Wynxx, but we're also starting to embed on our client projects. Now one of the interesting aspects of this is how can you identify what is truly agentic. Without sifting through it, without using it yourself, it's very difficult to do that. What we want to do is build the trust in the level of execution that we perform with our clients with the products that we build. So GFT's name is synonymous with being that challenger in the industry all around AI and agentic AI. Now in order to move forward, we can't just make our platform agentic, okay? Other people are creating agentic platforms. Why does that make GFT special? It doesn't. We need to focus on moving that forward. We need to look at the evolution of those agentic platforms. To date, GFT has obviously focused its Win product on the software development life cycle. Now Win in itself, as you can see from the slide, has focused on design and planning, development, testing validation, deployment operations holistically, okay? The features that we have sit within the software development life cycle, and they are there to enhance what you can get by buying something like GitHub Copilot or CursA or using Anthropic CludAI. Now those specific tools are aimed at code generation, code completion. They will create a read me file. And I'm sure if you've done any reading on Wynxx or you've heard some of the presentations today, you understand that our product doesn't focus in those areas specifically. It can touch on that. It does some of the fringes. There are a number of articles that you can probably find on the Internet right now by Gartner, by McKinsey, by Bain, and they all state the same thing. A developer spends approximately 20% to 30% of his time engaged in writing code. If you're looking at these coding assistants, which have come a very long line and are very useful, I use them myself, you are going to be focusing those savings on that 20% to 30% of your day. GFT's Wynxx product focuses on the holistic software development life cycle. So we try to get as much of the savings as we can from that other 70% to 80%, which is a far greater share. For that reason alone, without adding in the additional savings from using one of these coding assistants, that's why you're seeing some of these larger numbers in terms of our savings compared to the numbers that you're seeing from those traditional coding assistants. Now I mentioned that we need to move forward. I mentioned that we need to enhance our product beyond agentic. One of the ways that we're going to do that is that we are going to start to focus our product on becoming more of a general intelligence model. We are going to get our platform to think how can I do this? What is my reaction to that? So the way that I or hopefully, many of you in the room cognitively process things based on our surroundings, by our situations, by the things that we have at our disposal like tools or a glass or a piece of paper, we are able to solve complex scenarios. We are able to understand the context of each of these things. We're going to create a central orchestration process within Wynxx as a platform. And that orchestration, you can think of it as a brain. It is there to understand its context. That context might be architecture. It might be code, it might be documentation. It could even be an organization's IT security policy. But we're going to be able to ingest all of these components into that central orchestrator or brain. It is going to then understand contextually the meaning behind all of these things. It is then going to create a plan, okay? Very important thing in agentic AI is to create that plan. The orchestrator is then going to be able to provide that sequence and that communication out to other agents or APIs or micro services to execute on that sequence, giving you an end-to-end transformative process. Now both Marco and Ignasi have talked today about the importance of being able to transform organizations. This transformation is very expensive. It's very time consuming. It's very complex. What we're doing here is we are going to provide a platform that is going to break down some of those barriers. It's going to reduce that time. It's going to simplify to the point where the risk of doing those migrations is reduced. Now how are we going to do this is probably the question you're going to ask. What I've described right now is probably the holy grail of products, the holy grail of platforms. How do I minimize my cost? How do I make it go faster? How do I reduce the risk and complexity? So to date, I'm hoping that you've all heard of a large language model. The term has been thrown around a few times today. A large language model ultimately is therefore creation, okay? I'm going to ask it to provide something it will based on its understanding or neural network within its large data set, it's going to provide some contextual output. A large action model goes beyond that. So whilst a large language model can create, for example, a plan, the construct that is now a large action model can actually take that plan and execute and communicate. So this is going to be one of the central components of our brain. Large action models are available today. You can find them in open source. So they are starting to become a little bit -- I don't want to say widespread because that would be exaggerating slightly, but we are starting to see bleeding edge technologies that are starting to make use of this technology. GFT also want to be a bleeding edge organization taking advantage of this. This large action model adoption will essentially move the dial of where we are, where typically we would have to write a huge amount of code, a huge amount of logic, we'd have to build learning cycles. We'd have to build a memory. The large action model actually takes care of a huge amount of that. And we can leverage all of GFT's existing learnings, coupled with that simplified technology stack, enabling us to go to market quicker, more effectively to a much better output for our customers. Now what we've heard from Ignasi earlier today, and we've heard from Marco already, these legacy transformations are highly complex. We've heard that they can take several years. They can cost millions and millions of pounds. One of the larger cost components of these types of transformation projects is the validation. If I write a piece of code, someone's got to check it. If I release a piece of code, it needs a review. If I write a document, it needs a 4 eyes check. All of these stages across that process require somebody to put a second pair of eyes on it and effectively sign it off. Now with agile release processes, we've obviously seen DevOps tool chains. We've seen better workflows that have enhanced it. But ultimately, it still comes down to someone needing to check. One of the benefits of, again, the large action model and some of the technology understanding of our architecture around putting inspectors or auditors into the loop is our AI platform is going to validate at each stage a lot of that output. The winner of the race for transformation is going to be the organization that manages to build in as much of the solution validation to their product as possible. If we can reduce that amount of validation that typically can be 60%, 70% of a project's time line and cost, this will be groundbreaking for our clients. This is where GFC can essentially pick up the baton as a leader in the field. So what we're hoping is that we can compress those 18- to 24-month projects by using our new platform to under 6 to 9 months. We've seen some of the savings that Marco has talked about. We've seen some of the savings that Ignasi has talked about. We are looking to go beyond that. Now I've spoken a little bit about how we're going to evolve Wynxx as a platform. What I haven't said is what the future for that platform involves. Now it's going to become an agentic platform. It's going to be able to communicate externally with other agents. So not only are we going to turn all of our features into separate agents. Not only are we going to add them to our agentic marketplace, both our internal client-facing agents and our internal ones that are actually going to be added. All of these agents will essentially be able to be accessed by our central orchestration or our brain. The brain will also be able to use anybody else's agents. There was a great question around the SAP. Are we going to be aligned with SAP's AI strategy? Or do we conflict with it? This is a great example of how we're going to work with it. Our brain will be able to reach out to SAP's agents. We will be able to embed those within our workflows. Our brain will understand the complexities of what the SAP agents are doing. We can teach it how to embed those in its sequencing to solve even more complex problems. So that's a great example. thank you for that question, by the way, teed that part up nicely. Now the next stage of the vision is not just to use this as a platform or a suite of products. It's actually to start to become a de facto member of a human AI partnership. We are going to embed Wynxx not as a tool for engineers to use, but actually as a worker as part of a holistic team unit. We're going to be able to offload a number of specific tasks. It will go and autonomously complete those end-to-end. That is the vision for Wynxx. Did I do on time? Almost.

Well, thank you very much, Dean. Thank you very much, Ignasi, for that insight and fascinating look into enterprise software and where AI systems learn that and collaborate intelligently and specifically. You have perfectly been in time. I think there are tons of questions in the audience. I suggest to have a coffee and address these questions directly to our speakers. And we will again meet in, well, roughly 20 minutes at 4:00 p.m., we continue with the next session. Thank you very much. [Break]

Okay. Welcome back, everybody. I hope you had enough time to grab a coffee and to interact with our tech experts and also have a chat with the management. We now shift to the real-world client perspective. What does AI-driven productivity look like inside the major financial institution. I'm very pleased to welcome [indiscernible] from Bradesco Seguros together with Ignacio to share how AI is making a measurable impact on delivery and efficiency. Please, Ignacio. The stage is yours.

Again. Again. One second. Thank you, Andreas. I was a little bit jealous when Marco used the video. And that's why as a nice breaking, I'm going to also add this video to start with and presenting officially what is wings, and then we will go into this conversation with [indiscernible]. [Presentation]

All right. Thank you. So as I said before, of course, we are 100% certain of what we are saying and what we are sharing with you in terms of KPIs, et cetera, et cetera. But I think that if we have the capability to bring a customer talking on our behalf around experience that they have had with GT using wings, that's the best story that we can share. And for that very reason, I'm honored to welcome our guests, [indiscernible], welcome.

Thanks, [indiscernible]. Thanks for having me here.

Now I see you, [indiscernible]. So [indiscernible], before we jump into a set of questions, please, for the benefit of the audience, can you introduce yourself, your role inside Bradesco Seguros and tell us a little bit about yourself, please.

Yes, sure. First of all, thank you, [indiscernible] and GFT, for the opportunity to be here with you. I'm running the IT shop here in Bradesco Seguros, mainly the development team, so responsible to make sure that we have the most of our IT investment, the investment in technology that the company has. I can also introduce Bradesco Seguros now if you think it's a good definitely to do that, [indiscernible]

Go ahead.

Yes. We have -- we are an insurance company, the largest in Latin America. We are based in Brazil, very focused on the local market. Last year, we had a revenue of EUR 20 billion, which was quite good. And it left us a net profit of EUR 1.5 billion. And fortunately, this year is -- the numbers are looking even stronger than that. So we are presenting good results. And of course, the challenge that we have is to keep going with this level of outcomes.

Thank you, [indiscernible]. And again, thanks for coming here today. We've been talking a lot around AI today, but also over the last 2 years, a lot of companies, everybody talking about AI. So when you decided back then early in 2024 to start adopting AI for SDLC, what was the main motivation that spark this change in your strategy and decided then to adopt artificial intelligence for your development processes?

Well, I'm pretty sure that all of you know that the investment in technology is one of the biggest constraints every company has. So the good side of our company is that we have an amazing sales team that is able to bring home a revenue of EUR 20 billion every year. But on the other hand, we are obsessed by efficiency. So we need to provide the most with the budget that we have for our annual investment. And that was the main motivation for us to look at AI. So this obsession to make more with the investment that we have. It's not little, but it's always a constraint, right? In every company, the investment in technology is something that you need to take the most of it. And the other side of this equation is that we have very nice and very good strategic partners like GFT and a few others that are connected with technology. So using these connections, we saw that AI was gaining momentum, mainly in code development, but we asked ourselves together with GFT, what else can we do instead of just using for accelerating coding, but can we use in the overall system development life cycle in other arenas as well. So that was the main trigger that sparked the whole thing and got us where we are today.

During today's session and usually when we are reading the market around the impact of artificial intelligence, we are very keen to explore what are the, let's say, the quantitative KPIs that are supporting this strategy. But I'm interested also to know a little bit more about the qualitative KPIs. Did you experience any change between before, in our case, using Wings and after using Wings in your operations, in your teams?

Well, yes, absolutely. And I think I can highlight 3 main things here. One of them is the shift left. So we are able to anticipate our gates of quality, security, so all of that. So it put us in a very good position to understand earlier where the showstoppers would be and get ahead of time in terms of being faster in anticipating what would be happening with the code when the code gets to a quality or security gate. So we have less code being holding these gates, which means that we are more productive. The other thing is the shift right. We are also being able to give the product owner more visibility about the outcomes when certain topic or story will be ready, how much it will cost. So it's more predictable now for the project owner to understand the whole thing. But probably the most interesting is probably what we are seeing in the screen now. We got a very simple word cloud from our tool is the tool that we're using to store all the data we have about our agile development, and that's the nice thing about having data, right? So before using AI, we got this word cloud and what was screaming in the cloud, the words that are screaming on the cloud were mainly about the process. So we could see things like daily alignment, so the ceremony of the agile methodology. After using AI, we start seeing these words changing to the actual objective of the squad. So in a very insightful way, we noticed that the mindset of the team was changing from the process to the final outcome from the objective of the squad, and it was very interesting.

And [indiscernible], that question we cannot miss, right? So qualitative KPIs is for sure an improvement in the attitude of the team towards the adoption of AI, but we should touch as well the quantitative one. So what can you share around those metrics that you are taking into consideration to say, yes, Wings was providing value to our operation.

Well, and that's all that matters, right? So the bottom line, so the results itself. And what we noticed is once we are in regime, so after implementing AI, we could see gains from 30% to 40% in productivity. What it means in the real world? It means that this year, 2025, we are delivering with the same technology investment that we had last year, 30% to 40% more to the business. So -- and we opted to do that, right? The other option that we could have is just save this money. And actually, we are doing that in certain areas, for example, in maintenance. We are able to maintain the systems this year, costing 30% less than we paid exactly for the same maintenance last year. So it's a cost reduction for Rio. But for investments to develop new things, we opted to protect the investment. But this year, we are delivering 30% to 40% more to the business in terms of -- the main metric that we use here is function points. So we are able to deliver more function points to the business, 40% more function points to the business this year than we delivered last year. And one good thing is that we are using these metrics for 5 to 6 years. So we are quite mature in understanding how our productivity behaves in the overall development cycle. How many function points we are -- each squad each agile squad is able to deliver month after month. So we are quite mature in these metrics, which helped a lot us to understand and to measure these gains in an effective way.

I think that it is instrumental to have a good methodology to measure performance of different providers, of course, including GFT. But clearly, that was the baseline for you to then determine that GFT was, let's say, one of the winners, right, in this AI race. I know that you also have some KPIs from Bradesco that you want to share with the wider audience today. Please, [indiscernible], tell us about what we see on the screen at the moment.

Well, yes, the gains that we had in productivity, there are many different flavors. So we saw these gains in application modernization, in code conversion and everything. The one that I like the most is probably what you can see in the left bottom of this screen. We are calling vulnerability for the lack of a better word, but it's prevention against cyber attacks. We have many tools in the market that are able to identify a potential threat that happens somewhere in the world. then we check our code against that threat and we start fixing to eliminate to prevent to be interested by similar attack. When we look at that 2023, we acted in roughly 60,000 preventions, and we spent BRL 13 million our local currency. Then the following year, which was last year, we did 3x more with half of the budget. So it's an impressive gain. And this year, we kept the same budget for this kind of activity. And very likely, we'll be delivering 50% more than we did last year. That's the kind of repetitive work where AI is very strong, where we can see the results in a very strong and very positive way. So this is for highlight one of the pieces that we have in the overall development cycle.

So more value with the same budget, I guess your Chief Financial Officer is super happy, I guess, with the outcomes.

Thinking about know very well that they are always asking for more. So and our mission is to challenge you guys to deliver more. So we are together on this.

And we are fully committed to deliver that as well. So thinking about and sharing more, let's say, details around the main features that the developers use the most. What can you share around this topic with the audience today?

Well, we are using for the entire development life cycle. So one thing that we actually, the analysts, they like quite a lot is to use agents to help in the story creation. And we measure the quality of the store by points. So we can see that the story that was Level 5, for example, last year, is functuating something like 7 this year. So we have better quality story is being produced and a better quality story affects the whole cycle, right? It goes smoother through the whole development cycle. But for the developer itself, what they did -- what they do like us in this prevention to cyberattacks is a very good example because the work that developer used to do was to get dozens, hundreds of codes, very similar pieces of code and do very similar changes in those pieces of code. So it was quite repetitive for them. They did almost the same change in hundreds of different codes. Now what they do is working a prompt that will be doing the work for them. So they love it. It changes completely the mindset of the developer. And then after the prompt is done, they see the end result and then they refine the prompt to get even better, and that's where the productivity is coming.

Part of it is cultural change. But if we talk about challenges, right, more than 180 professionals across more than 20 squads. What can you share around the challenges that we have had during this period?

Yes, if we look at this number of 20 squads is what we have with. If we look at the broader team, we have almost 200 squads running. And I think there are 2 things that we noticed. One of them is the time. So the productivity does not come fast. We noticed that we spend 3 months in a setup and ramping up the team. So if we have a squad that's not using AI for the same -- very same squad to get ready, we need to train the professionals. We need to get the tooling ready. They need to get used to that. And what we noticed is that it typically takes 3 months for the whole thing is ready. And in these 3 months, we didn't see any gains. Actually, sometimes we see even some worsening on the productivity. But after that, the magic starts happening. In the following 3 months, we can see this number stabilizing. So then the productivity starts to show up and the teams get their gears really going, and we see the numbers showing up. That was one thing. So you need to persevere in doing the work and get confident that results will come. The other thing that we noticed is that the results mainly at the beginning over these first 6 months, they are very heterogeneous across the different squads. -- mainly because what is very typical is that we start with the quads with the worst productivity. So we see the major gains in them because they had a bad productivity at the beginning. So if we look throughout the different line of businesses, we saw that the gains were as high as 70% in one side and as low as 20% in the other. But when all of this come together and stabilize after these 6 months, what we can see after all this turbulence is gone is that we end up with 30% to 40% gains overall. So you need to pass through that to see a stable result at the end. So -- and sometimes people are impatient wait for that and anxious to see the results and the results are showing something and they think they can take that for granted. So you need to give some time to get a stable number out of it.

I think that some part of the KPIs that -- or the aspects have already been shared in here in the following plot. But just for the sake of time, Gay, I'm going to jump to the next one, which is part of the story that you were sharing in the latest question is related to having this perseverance, having this strategic vision that needs to be implemented and wait for the results to come, right? From that angle, are you redefining the concept of your workforce in this new AI era?

Yes, we are doing that. So of course, we need to change the way that the team is working, and it's very important. They need to change. We need to change -- you mentioned before, right, it's a culture of change. So we need to change the mindset of the team to take advantages of the new tools that they have available. And being a developer, I know it's quite hard for the developer to believe that AI will be doing better than the developers. It's a very artisanal kind of work, and they love to do this stuff. But once they are challenged to do prompts that help them to do the work, it starts getting interesting. So that's a very important thing. But I think we are now moving to a different stage, right? We -- until now, we have been using AI to accelerate the work of the team. I think we are now moving to a point where agents will be doing -- AI agents will be doing a complete piece of work by themselves. And I think that's the next challenge we have together. How can we have mixed squats where we have some humans, but also some autonomous agents that we just ask them to produce a piece of code and test it and we get it done and join with the rest of the work to have the whole thing ready.

Very aligned to our view as well. Now for closing [indiscernible], what tips would you like to share with the wider audience that we have here today face-to-face and also virtually connected to this session of all the journey that you have started more than 1 year ago.

Well, there are 4 things that I usually like to share with the people. First of them is that security is a must. So if you have any doubt that your code will be protected, don't go there. So you need to make sure that everything is secure from the beginning. Your code will not be shared with anyone else, your data is protected. That's the very first thing. The second one is that you need to have the people train it. So investment in enhancing people ability is very important. You need to get the tools to train, to understand how the technology is evolving. So keep this space is very, very important. And then the other 2, I got this from Charles Darwin on his theory of evolution, right? He noticed that who survives in the nature is not the strongest or the fastest, is the most adaptable to change. And I think it's so true in our corporate world as well. So being able to change constantly is what differentiates us as a company. And we need to understand that and keep this space as well. And the last thing also comes from Darwin. He noticed that for IPC to evolve, competition is very important, but collaboration is even more. So that's why we love to share with other companies to learn from them as well to evolve together. And I read leaving here the invite for everyone for us to collaborate and go beyond.

And let me add, point #5 is working alongside GFT into that journey, but it's going to be me saying that, [indiscernible], not you. Thank you very much, [indiscernible], for being our trusted partner in that journey for being here today to share everything that you shared with us, your experiences, the challenges that, of course, it is not an easy task to do. But definitely, Bradesco Seguros is determined to go into that direction, and we are honored and proud that Shift is part of that journey together with you. Thank you, [indiscernible].

My pleasure. Again, thanks for having us.

Well, thanks also from my side, [indiscernible]. Thank you very much. Thank you very much, Ignacio, for hosting that session and for the insights and sharing these impressive results, real tangible results from real project that connects very well to the theory we had before. So we now turn into a topic that is absolutely crucial for the global financial stability, the role of AI in combating financial crime. And please welcome my colleague, Floris from [indiscernible] and Michael Laws from Deutsche Bank for our next slot. The stage is yours.

Thank you very much. Warm welcome from me as well. We'll be switching gears now. We heard a lot about Wings as one of our proud products in my team, Software Solutions. We actually also are very proud to have the products of Smart where we have been serving many banks, especially in Germany and in Switzerland, Austria, to stay compliant, a key capability of keeping the license and keeping the trust of your brand is actually being able to detect what kind of money flows through the bank and also adhere to all the regulations that keep on changing and keep on being added across the world to fight terrorist funding, to fight anti-financial crime, to fight money laundering. And we're very proud to have been serving Deutsche Bank for many, many years in this space. And I have a great pleasure of welcoming Mike Laws into the call. If we can switch to the screen. Perfect. Hi, Mike. Good afternoon.

Hi everyone. How are you doing.

So Mike is actually in London, and Mike Laws is responsible for all of the technology stack that Deutsche Bank uses globally to fight financial crime, which, as you can imagine, is a growing piece of technology because technology works both ways, right? You can use it for the good side, you can also use it for the bad side. And Mike is the one that tries to keep up or even be in front of the bad guys by having the right technology stack in the bank. So Mike, welcome to the audience. If you would like to briefly introduce yourself and your role in Deutsche Bank.

Thanks, Floris. Yes, Michael Laws, I've been at Deutsche Bank for over 15 years now. I've worked in many parts of the organization, started off in their markets business, but within technology. And then in about 2017, I was asked to take a professional risk and move into infrastructure, potentially to infuse part of the organization with a bit more of a commercial mindset. And I've been kind of in compliance and AFE ever since 2017, predominantly based in London. I've done some time in -- I did 3 years in New York as well. And more and more so, I'm doing regular trips to Germany. But I would say that's a nice brief introduction for me, Floris.

Very good. So 50 years in the bank, the last 7 years only in fighting financial crime. Let's first do a quick look back, right? So what would you say is the biggest trend in technology in specifically your area of the bank? What have you seen changing much in the past years?

Yes. Well, it's what's changed or almost -- I think what hasn't changed is often that's almost more relevant pointing out. So I think one thing that's different to many other sort of commercial parts of an organization is that I find that in regulation compliance and AFC, we're sort of never finished. So I'd like to keep with me a few examples of what was acceptable in terms of transaction monitoring or screening at the beginning of this journey. And I'll pick on one, which is client screening. So obviously, every bank has to regularly screen not just its clients, but associated parties for sanctions risks, internal watch list, a variety of things. And if I go back to that first day in AFC and compliance back in sort of 2017, we could probably screen about 500,000 parties a day, and that was probably okay. And if I look to what our plans are next year, we're going to probably be doing 30 million parties multiple times a day, and that's probably okay. So right? So the expectation kind of moves rapidly in this space, and it's always just okay. I've had a few instances where we've been called out for some particularly good work. But generally, it's trying to keep pace with the regulation. And of course, the next one is, which I'm sure we'll talk about that in a bit more detail later is the utilization of AI and machine learning. And I don't think anybody involved in this space has not noticed that more recently, there's been far more willingness and increasing engagement from regulators around the potential for AI, which wasn't always the case. You could argue that you could interpret the new EU AML regulation that, that kind of level of engagement has moved to support, which very quickly becomes expectations. So clearly, I would be surprised if there's many top-tier banks that don't have some AI detection next year. Just to round out this thing that we're never finished, there is this -- we must not only keep pace with regulation, but we always -- we also have to keep pace with business and market growth. So like if you pick transaction monitoring, the U.K. digital payment market, I think it's going to be about USD 450 billion this year. That's going to double by 2029. And real-time payment transactions for the last 2, 3 years have gone up approaching 50% and your monitoring environment is just expected to handle that. If I take some very real examples where we've introduced clients into Deutsche Bank that are merchant solutions, money service bureaus, payment service bureaus, we're talking about going from a system that probably has to handle 100,000 transactions a day or a month to tens of millions. And we need to be -- we need notice to do that. So I think the first big kind of challenge in the past few years is that sort of expectation like we're never finished. More recently is absolutely this expectation versus actual progress that can be achieved with AI. I think those commercial products out there like ChatGPT, Gemini, they are so gratifying, right? They are designed for the consumer market. They make you feel good. And then you start thinking about how I could apply that to a regulatory environment. And the margin for error is so low in what we do, all the requirement to explain an error if it happens. I would just use an example from last weekend where I was trying to understand the internal temperature that I had to remove a joint of beef from the oven. I didn't want to know what temperature it had to be. I wanted it to account for the resting period, the temperature increase. And it told me that I should remove it from the oven when it was 5,658 degrees Celsius. Now if I had somehow applied that to a threshold or something in my transduction monitoring system, I wouldn't be able to say, well, you win some, you lose some with AI. So I still think there's that barrier that we kind of have this expectation that's been put on us in tech and we haven't quite achieved. And I would say the last one I will talk about in terms of the challenges where I really think the value comes from is really understanding the bank's role in fighting financial crime, yes. So we don't arrest people. We don't go and find criminals. We are a part of a very, very regimented structure that informs kind of governing bodies and law enforcement regionally and in specific countries. We play a part. And sometimes, if you don't truly understand the laws, regs and guidance, you can actually sort of get a little bit ahead of yourself in trying to think, oh, surely, we need to do this. And actually, that's not our part in the overall picture. So one thing I think in more recent news is really understanding what excellence looks like in terms of our piece of the puzzle. So Floris, I think those 3 challenges of the last year, of the last few years for me have been that expectation. We're never finished, that AI is supposed to make everything we do super easy, but we still seem to be doing loads of work. And then this evolution around really understanding what excellence looks like for a bank and the role they play in financial crime.

Okay. So okay, okay is never good enough or is just good enough, right? That's your emotional state in most days. Is that a fair statement?