Alpha Bank S.A. (ALPHA) Earnings Call Transcript
November 25, 2025
Earnings Call Speaker Segments
Good evening, ladies and gentlemen. Welcome to the Alpha Bank Innovation Day 2025, an event dedicated to the transformative world of artificial intelligence. We are here today because AI is fundamentally reshaping how we live, how we work and how we do business. Throughout this evening, we'll exchange perspectives, explore real applications and highlight how Alpha Bank is at the forefront of this shift. We are driving meaningful change and are strengthening our role as a leading force in a rapidly evolving economy. A special highlight, of course, of today's event is the final stage of FinQuest, Alpha Bank's international innovation competition. Our finalists will pitch their AI-powered solutions across diverse sectors from agritech to fintech and beyond, showcasing bold thinking and actionable innovation. Later, we'll celebrate and reward the winning ideas, those with the potential not just to improve the future, but to actually redefine it. Our purpose today is to strengthen the innovation ecosystem through idea chains, inspiration across industries and international perspectives. We are very excited to welcome here today representatives from all major Greek banks, leading Greek organizations, members of the Greek government, innovative start-ups, venture capital firms and partners such as Microsoft, IBM, EPAM, EY, Accenture and Endeavor Greece. Innovation Day has been designed to spark dialogue, foster connections and highlight where the true value of AI lies. We are pleased to begin today's program with a welcome from Alpha Bank Group CEO, Vassilios Psaltis. Please join us on stage.
Ladies and gentlemen, distinguished guests, partners and friends. Welcome to the Alpha Bank Innovation Day. This is our inaugural forum that builds on the success of the FinQuest competition to shape a broader dialogue on how the financial and fintech ecosystem can bring forward new ideas for making sure that we're all going to be enjoying a better future. This is where bright minds on the one hand and bold ideas on the other hand, do come together to shape tomorrow. This year's theme, which is using AI power for the future of finance and all that, we're going to be doing this together. This, in my mind, it perfectly captures the notion, the spirit of where we're currently living in. It's a time where artificial intelligence and fintech innovation are converging, not only to transform how banks operate, but also to redefine the relationship between people, technology and trust. Just a few years ago, Greece was taking its first step as far as the innovation arena is concerned. And today, look where we are. We're marching ahead in full speed. According actually to Endeavour Greece, the Greek tech ecosystem is now valued at about EUR 12 billion, which is quite something if you compare to where we were 2 or 3 years back. Greek founders have raised $1.3 billion, both in terms and equity just in 2024, out of which EUR 400 million were actually related to start-ups that were established and were run here in our country. This milestone marks a dynamic and resilient ecosystem. The innovation engine is running at a fast pace, and Alpha Bank is quite proud to provide the fuel to it. We are collaborating with Moveo, one of Greece's most promising innovators. We have acquired FlexFin, which is a digital factoring that provides factoring for SMEs. Our venture capital arm has already invested something like EUR 16 million directly and indirectly into the fintech space. But at the same time, we're looking forward towards tripling it to EUR 45 million by 2027. At the same time, we do actively test fintech solutions. For example, we have run over 15 pilots with emerging fintechs just over the last 18 months. Each of this step is part of a clear vision. We want to blend the agility of the start-up world on the one hand with the trust, scale and experience of a systemic bank. This is to create a more efficient model of banking that moves faster and most importantly, that moves faster -- that moves smart. At Alpha Bank, we do pursue innovation in 2 complementary ways. At first, we're building an innovation ecosystem that brings together talent, technology, but also procure for partnerships. We cultivate future skills of people, develop in-house innovation, and we collaborate with leading technology companies that place the customer at the very center of what we're doing. Inorganically, through our dedicated innovation and partnership team that looks forward to forming partnerships in order to unlock new capabilities and value. And on the other hand, organically by building from within through a defined methodology that we're using that brings customer into the innovation process from day 1. We are actually co-creating solutions with them. And this is something that resolves real-life problems. The myAlpha Vibe app is a prime example to that. This was the first digital pocket money product that we have launched, and that was an outcome of exactly that process. The feedback that we got into practice, and we got out with a unique product. And now it is already evolving into the next even more innovative version. This is how we view innovation, not as a one-off disruption, but as a living process of learning and experimentation. Today's forum is as much about fostering the ideas that can upscale innovation in our sector as it is also about leveraging technology to create new experiences that empower our customer base. At the center of this transformation is artificial intelligence, which is fundamentally redefining how banks think, operate, but also connect to their customers. AI gives us the opportunity for radical repositioning. This practically means transforming banks from product providers into intelligent organizations that understand that predict and that respond to customer needs de facto in real time. That means that we are gradually evolving from being transactional to anticipating customer needs and proposing solutions, standing as a true partner vis-a-vis our customers. AI allows us to focus on what humans do best, and this is empathy, creativity and user judgment. As we become an AI-first bank, the way to ensure this is by remaining a human-first organization. At Alpha Bank, our purpose is to empower progress in the life of business for a better tomorrow and such progress needs partnerships. We don't believe that progress can be made simply in isolation. We believe in fostering partnerships that combine perspective, experiences and ambitions to create new solutions. Today's initiative is pioneering because it aims to bring forward the power of collective intelligence. We have come together today to exchange ideas, to share experiences and be inspired by what also other organizations are achieving, both within, but also beyond the financial services industry. This forum brings together start-ups, industry leaders and technology partners to imagine what comes next, but also to make it real. We are honored to host distinguished thought leaders and partners from global technology giants as Microsoft and IBM, technology integrators like EPAM to senior banking executives from UniCredit, Eurobank, Piraeus Bank, National Bank of Greece, and this demonstrates our commitment to collective advancement of the Greek financial ecosystem. And from Greece's broader ecosystem innovation leaders from Titan, Kotsovolos, Moveo and others who are bringing transformation across sectors. We're also welcoming academic partners from the Athens University of Economics and Business who ensure that we maintain the crucial link between research and practical application as well as venture capital firms like Big Buy who bring ideas to life. Special recognition goes also to Endeavour Greece, who has been an amazing partner in shaping this forum today, helping us to connect with broader parts of the innovation ecosystem. And of course, a warm welcome to the FinQuest 2025 finalist joining us from Switzerland, from Israel, from Spain and Greece, outstanding team that represent the creativity, agility and global reach of the fintech world. Their AI-driven solutions show us how technology can create value across borders and across industries. Each of these individuals, they do bring unique perspective into how AI can be used not only to create efficiency, but also to create meaning and long-term value. This spirit of partnership of openness and shape the vision is what makes this initiative unique because the future of innovation would not be built behind closed doors. It will be built together across industries, across borders, across disciplines. And we begin today's journey with inspiring discussions, visionary ideas and the FinQuest 2025 Finalists. Let's keep one principle in mind. The true promise of AI technology is not -- does not lie in technology itself, but rather in our capacity to use it wisely. At Alpha Bank, we are committed to leading this transformation, shaping a future where finance is not only smarter and faster, but also more human, more inclusive and more sustainable. We are committed to support Greece's innovation ecosystem as an investor, as a partner and as a creative test bed where innovative scale-ups can be put to practice. Thank you for being part of this journey. Welcome to the Alpha Bank Innovation Day and to the future we're building together. Thank you.
Thank you, Mr. Psaltis, for sharing a vision anchored in purpose and innovation. And of course, a big thank you to all of you for joining us here today at [indiscernible] and of course, via live streaming at home. And now let's please welcome our keynote speaker, Tey Bannerman, AI Strategy Leader and former McKinsey Partner in a fireside chat with Michalis Tsarbopoulos, Chief Digital and Technology Officer at Alpha Bank. Please join us on stage.
So today, I have the pleasure of setting the stage with Tey. Tey is one of the leading personalities regarding voices regarding AI. It's one of those people that has shown how to translate AI from hype into real value. And talking about that, you've been one of the key voices -- one of the leading voices in social media talking about is it hype, is it value? Tey comes with over 2 decades of experience in product management, in engineering, in strategic consulting, and he has helped leading global organizations navigate the complexity of AI transformation with a truly human-centered approach. So Ti, welcome. Welcome to Greece. Welcome to Alpha Bank Innovation Day.
Thank you so much for having me. It's an absolute honor to be here, and thank you all for being here as well.
And I'd like to start our conversation exactly by touching upon this topic around value. So everybody is talking about AI. Is it hype? Is it real value? I know you've spoken many times about the gap between what AI can do and what it really does or what it takes to generate real value. So in your view, are we in a genuine value creation phase? Or are we in the hype and creating another bubble?
It's a great question. And I think really, we're in both at the same time. I think there is a lot of genuine value creation. I think if you look at specific domains like personalization, like customer service, like pattern matching at scale, you're seeing real kind of business and enterprise use cases for AI that are driving real value. But at the same time, I'm sure you've all seen the numbers and reports 75%, 90% of AI implementations are failing. And I think that's what's driving a lot of the hype. A combination of that plus a little bit of overpromising from vendors about the real-world capabilities and what's required.
We have some of those vendors around. They are not the ones overpromising, right?
But I think the overpromising is interesting because I think if you take any topic in AI, like AGI, for example, you have a massive range of kind of estimations on time lines and disagreements between experts. So it's a world where nothing is concrete. And so I think that kind of adds to the hype cycle as well.
Okay. So you're saying, yes, there is an exaggeration. Probably there's a bit of a hype, but there's certainly real value to it.
Absolutely, absolutely.
Okay. Turning to another topic, which is about society and the workforce. When we talk about AI is accelerating capabilities, the conversation naturally expands beyond technology. And we are entering a phase where AI is reshaping business models, it's reinventing ways of working, habits, what have you. Question is, as AI is becoming more mature and more capable, what do you expect to happen to the workforce? Should we be expecting a replacement and augmentation, a redefinition of roles? What do you see coming?
I mean I really think it's going to be a combination of all 3. But I think ultimately, it depends on organizations like the ones here and the leadership of those organizations deciding what's the right ratio for that future workforce that we want to create because there are definitely clear areas where it does make sense to have AI taking a lot of tasks and automating tasks that might be manual and tedious. And there are many, many areas where it does not make sense to have AI doing tasks. So I do think that there's a lot of kind of very kind of decision-making with conviction that's going to happen now that really determines whether it's augmentation or replacement or a combination of the 2.
Okay. So you're saying it probably is going to be a combination of all, right?
Yes. Yes.
Okay. Good. And then turning to society. I think the main question is, at the end of the day, do you think AI is going to be beneficial for the society or to make it even more specific since most of us here are working for banks. How can we build AI systems and solutions that serve humans and communities and that not just maximize shareholder value. Is there a way to strike that right balance?
I think there is and kind of -- and I've seen both sides. So my background, I was a software engineer for many years. I was within IT and technology teams. So I was on the side where it was technology first. And -- but kind of over time, it kind of really moved to, okay, how do we put the human at the center as kind of Mr. Psaltis said at the beginning. And I think it starts with design principles. If you design from the principle of asking the questions of who could be harmed, what could the human outcomes be as a result of what we're building, you tend to really approach the way you design AI solutions very differently. You have very different teams. It isn't purely technical talent and engineers. You have different kind of capabilities and decisions that you're making along the way, but you also kind of start to think about the second and third order effects of some of these systems and decisions that you're making. And so I think it's all about the design principles that you start with that really determine the outcomes that you get to.
Okay. All right. So that's important to underline. It's not about focusing on the technology, but starting with the solution and the design and the kind of problem and the kind of objective that one needs to serve through AI. Good. Turning now into another topic, which is about organizational transformation, organizational readiness or AI readiness as we call it. Typically, big organizations like banks, for example, I'd like to think of ourselves as being quite ready for it. But that's not necessarily the case for the larger part of the industry, which is SMEs, for example, or medium-sized companies. What is your view on this? Do you see the industry, the broader, let's say, corporate world being ready for AI? And if you were to advise medium-sized or small and medium-sized enterprises, what would be your advice as to the 2, 3 things they should start looking at in order to get themselves prepared for AI?
I think it's a good question. And I think the fact that we're in a room with a lot of people from banks and banks have been at the forefront of this change for a long time. I remember the first bank I worked with on AI was 8 years ago, and it was a bank building machine learning and customer lifetime value models to predict customer behavior, but they also started building large data science teams and so on. So I do think a lot of enterprise can learn from what banks are doing. I think kind of programs like this are a fantastic way to really publicize, okay, not only here's what we're doing, but here's what we're learning and how. And I do think that a lot of examples from how organizations that are in the lead do things can really kind of serve as a template. So I mean, if we take a practical example, in wealth management, an area that I've done a lot of work, you're seeing a transformation where when you put that vision in place for what kind of future wealth manager look like, augmented by AI. It's AI doing a lot of the manual data entry and summarization work that a wealth manager might do now, but leaving the wealth manager more room to form deeper relationships with their customers to really look at helping customers through life transitions and so on. So a lot of those kind of deeply embedded human elements that require complex problem solving and relationship management. And I think you can carry that across to other industries that also require those types of relationships to go, how can we do this, but also what are the foundations? What do we need from a data perspective? What do we need from an organization and culture perspective. And so looking at you guys as a template for what they can start doing, too.
Okay. I think that's good advice. Can we get a little bit more specific? I mean, one thing is to tell them to advise small companies to look at what bigger ones are doing. But in more practical terms, would you say that they need to invest in people? Is it technology that they need to invest in? Is it frameworks? And then looking at AI, would your advice be to go to find problems for which to seek solutions? Is it to go maybe in certain domains and think about -- rethink the whole domains? How would you go about setting this up?
I think what we really see is that it's very rarely the technology that's the biggest investment or that's the first thing that you need to do. I think if you look at the kind of percentage of companies that are really doing well with AI, what they tend to do is start with kind of what we call diagnostic honesty, really understanding where we are on data, where we are on our vision and strategy and using that as a foundation to then decide what do we need to do first? And very rarely is that we need to bring in technology first because a lot of the foundations tend not to be there for a lot of I think that's kind of the first thing. I think the second thing is really not taking like a use case approach and going, okay, where are the AI use cases that we can actually begin, but looking at a whole domain. So for example, looking at the marketing function of an organization and going...
Well, we have a lot of marketing people in the room.
Fantastic. Good to meet you all. So looking at the marketing function and going, okay, what is our future vision for that marketing function? And we look at what are the actual opportunities, what can AI do well, what can our people do well and then you put together a vision where you go, okay, it's not just about bringing in AI image generation and AI copy generation here, but going to a world where, okay, our function looks fundamentally different because there are a set of tasks that AI is handling. There are a set of tasks where we're collaborating, but there are also kind of a set of tasks where data is being kind of managed automatically and customers are being communicated to automatically on the one hand. And on the other hand, we are having more one-to-one relationships, for example. So it's really taking a slightly more visionary approach than individual use cases.
Okay. And final question on this. What about culture and leadership? Do you think there are opportunities, catalysts or obstacles when looking at the typical SME company or the typical medium-sized company wanting to enter into the AI space?
Yes. I remember seeing a stat recently, and it shows that companies that have an innovation culture are 9x more likely to see benefit from AI initiatives than companies that don't. And that involves things like giving your employees kind of flexibility and room to innovate to actually experiment. It means kind of having feedback loops where employees can actually say this is what's working and not working. It means kind of co-designing with employees and with people who actually know the problem space really well. And also kind of being -- not being afraid to kill an initiative that isn't working because then you can divert resources to something that is working. And I think that all requires leadership conviction, but it all requires change in terms of culture because not every organization is set up this way. Some organizations are far more rigid, and it does require that change if you really do want to reap the benefits.
Yes. And I'd say this is not only typical for small organizations, but in some cases, also in large. So it takes a little bit of a leap of faith...
Yes. Definitely.
To move ahead before you can have full proof. Well, great. T, it's been a pleasure talking to you. Thank you very much, and it was an honor having you.
Thank you so much for having me. It's been a pleasure.
And we'll both be around for the rest of the evening.
Thank you, Tey, for those fascinating insights and we highly welcome on stage. I will now call our first panel of experts to have our first conversation regarding AI in banking. So please welcome on stage Eleftherios Kororos from National Bank of Greece. Harry Margaritis from Piraeus Bank. And connecting remotely from Italy, Alessandro Barardi from UniCredit. The discussion will be moderated by [indiscernible].
So hello. I'm really glad to be here, and I want to thank Alpha Bank for this invitation. This is the first panel I moderate and it has to do with banks. I have to tell you. I am a partner in consulting at EY Greece since a couple of months ago, starting a new career. At the same time, I am the Chair of the National Council for Research Technology and Innovation. And for the past 8 years, I was a Director and the Chairman of the Board of Democrators, the large research center of Greece, which actually got to be totally transformed from a dead end nuclear center into now the AI epicenter of this country. So in my experience in the research laboratories, it is breathtaking. The disruption that AI is bringing, it is breathtaking. The acceleration we see in various scientific fields is has never been seen in the last decades. Anything from protein structure and drug discovery to telecommunications to advanced manufacturing, it's -- you see Nobel Prize is one based on AI methodologies and machine learning methodologies. So this is all great, and I understand and I'm very excited about this. And of course, in my everyday life, I use AI now. And I think I'm doing better and better actually. I'm getting a little hooked, I'm sure you are. But then when it comes to banks, when it comes to banks, I think things are more of a particular substance where we all feel very, very conservative about banks adopting new stuff and adopting new things. And please just don't treat me as a moderator. Please treat me also as a bank customer. So be careful what you say. So in that sense, it's really very interesting to describe how things are developing in the banking sector, whether there is progress or not. And because we are all tied up to the exponential changes in technology, we have to keep on changing ourselves. We have to keep transforming. And I know you guys have finished big digital transformation in the last 10 years. And we were very happy and very confident about this. And now here comes a new round of AI. And of course, the very first question that goes to all 3 of you. I'll start with you, Mr. Tsarbopoulos. What do you think the challenge -- how big you think the challenge is? How ready are the banks today and your bank in adopting AI and actually in adopting AI in the various layers from administration to customer experience to security. And what steps should be taken further? Where are we? And how fast do we move?
Right. Thank you. Look, I'd say the short answer is we're fairly ready because banks have been in AI for years, not generative AI or agentic AI, primarily traditional AI, but we've been there. There's a lot of discussion about is it real? What we are listening about AI, is there real value? We heard Tey tell us about how 70% to 95% of AI projects fail or they fail to bring real value. But that being said, that does not mean that there's not real value. When you look at analysts, they have split views. Half of them will tell you that the markets are exaggerating. The other half will tell you, no, there's real value because we see the big companies leading AI that are producing huge profits. If you ask me, I have no view on the financial markets. But what I can tell you from experience is that there is real value in AI, tangible value. We've seen that over the years. You talked about digital transformation. You're right. We have been running a lot of initiatives in digital transformation over the last few years. The majority of them have been concluded and now we are entering a new wave, as you said. And they have been concluded with success. For example, in our case, we've managed to digitize the vast majority of simple products and daily interactions with clients. We are close to 100%. We have 98% of transactions happening outside branches these days. Clients have very few reasons to visit the branch these days. Maybe the main reason to visit the branch is to get advice, which you need a human to get. And we've managed to reach 30% sales of new products through digital channels. Now you'll ask me, okay, but that's all about digital transformation. Where does AI fit into all this? An underlying enabler for all this success has been AI because over the years, what we did is, a, we developed machine learning, NLP and NLU models, which we then leveraged to convince our clients to come and use the digital channels. This does not happen by chance. It's not that when you digitize something, you have people running to use it. So you use techniques, you use models, you use targeted and personalized communication, you used advanced segmentation analysis to entice people to use your digital offerings. Then on the internal front, on the operational model front, we have been transforming the way we operate. So we started by centralizing activities, redesigning workflows and processes, simplifying those, automating. And now AI comes on top. And AI is giving us the opportunity to drive additional value creation. I'll give you some examples. Currently, we are getting ready to launch a number of AI agents. A good number of them are essentially chatbots. These are agents that have been trained to answer specific questions to our staff regarding policy documents, product manuals, digital banking services, HR matters. There's another cohort of agents that are being produced that will execute certain tasks. For example, we are creating an AI agent that will support relationship managers in preparing for the next client meeting. We are developing an agent that will be helping audit when they are concluding an audit review to validate and conclude their report. And another one that will be bridging together all the information and data for a credit underwriting in a standardized format. What all that means is that once you have a number of workflows standardized, you can then apply an agent that what it does is it takes the trivial daily administrative tasks from the human and it reduces the time spent on those tasks from minutes to seconds or from hours to minutes. Then you talked about customer experience and service, another area of focus for us for years and especially these days. First thing we did is we used -- we leveraged NLP and NLU to process the feedback we collect from clients. So every year, we collect through campaign mechanisms, hundreds of thousands of open text feedback points from clients. If we were to process that with humans, we would need dozens an army of people to process that. Instead, we have deployed an algorithm. The algorithm contextualizes the feedback, standardizes it, processes the feedback and then turns that into actions, which we then take into our organization, and we execute on those actions to improve customer experience. Customer service. We were among the first banks globally that launched a Gen AI-based chatbot beginning of this year. What we -- in our corporate site as a pilot. What we did since then is we expanded the knowledge base of the bot. We embedded it into web and mobile applications. And what we're doing now is we're looking to connect that to the call center. Essentially, what we are trying to do, where we want to get at is to have a system of customer support and service, whereby customers will have seamless service, 24/7 support by using AI, remote servicing and human agents. So if I were to summarize, I'd say digital transformation has been supported by AI and now is the time to scale it up and drive additional value by transforming the organization using AI at the core of the transformation initiatives.
I'm impressed. Seriously. I don't want customer [indiscernible] experience this kind of transformation you're talking about. But indeed, this is big progress. It is certainly the digitization, data, data, data. Mr. Kororos, you want to say as well about National Bank of Greece.
Thank you very much for the opportunity to be here and see friends and ex colleagues and current colleagues. So thank you very much. Yes. So continuing from where Michalis left it, I'll say that data is very important, okay? This is the underlying -- the foundation of all of it. So what we did is we centralized data. We built on the enterprise data warehouse. We tried to collect all sorts of useful and important information for the bank in one place. And this is not only about the strictly bank-related data. We're talking about also about processes, circulars, all sorts of things that support also the bank to operate. So this is where we put the data. The next thing that we focus on, and I don't know if it has been a coincidence or it's the natural evolution of things, but it was the digital transformation, as we call it, where we digitalized a lot of the processes. We upgraded the systems. We are now at the final stage of this transformation by completing the core banking transformation we're doing. And this has been very important. Why? Because by having the data centralized in one place, we can leverage Gen AI to build chatbots where we can get information from the chatbot, ask questions, get the answers, et cetera. But by having a modern architecture, it will enable us to have agents that can act upon this information. So not only -- I'll give a very simple example just for the audience to understand. We launched like the rest of the bank, our own chatbot on the public portal. You can ask questions, you can get answers, how can I get a credit card a loan, what are the documents I need, et cetera. But then we saw that the majority of the interactions have been around certain areas. There are some things that, unfortunately, they still -- our customers, they need to do or they need to consult by going to a branch. So they started asking about how do I book an appointment or I want to book an appointment. So what we did is we integrated the chatbot with the appointment booking application. So now they can book through the chatbot the appointment. So they don't get information when is the next appointment. They can actually book it. They can say for what purpose they want to go there and get serviced. So in order to do this, you need to have the technology that will enable the integrations with all the [indiscernible].
How long has this process has been taking you?
Okay. When we talk about Gen AI because it's part of the Gen AI, this is something that we've been working on seriously for about a year, 1.5 years, not just for the chatbot. It was -- what we did is we took a significant part of an existing team. We built an AI and innovation team. And from there, we created an AI governance framework. So there are principles that we have to -- we have to prioritize the use cases to build the technology. And then we started producing -- materializing.
In your own opinion, your own feelings about value created?
I think there is a significant value being created for our customers, but also for the bank itself. I will put an asterisk here because I think it's in the human nature not to trust each other very easily. So imagine what happens when we have to trust a digital colleague.
So real value is to be proved, but we're all optimistic.
I think that we will see the value. We will see the value, but it is a matter of trust on the tools that we have or the tools that we are building, the digital agents. And we will see. I mean, in technology, we see it immediately, okay? We build software using Gen AI. And is it the most sophisticated or the best software engineer that you will find? No. But how many of those can you have? You cannot have. So if you compare this to a medium software engineer and then you have the expert to supervise it, then that's where you get the real value. it is not only in technology. That's what we have to do in the rest of the process of the bank.
Okay. And you, Mr. Margaritis, at Piraeus Bank, what are the steps? What's the momentum there?
First of all, thank you very much, Michalis. Thank you for the opportunity to be part of this amazing discussion and congratulations on this great initiative. Under EU AI Act, I need to disclose the use of AI. So my responses have been prepared with the o3 OpenAI model. But I think for us, big systemic organizations in the banking sector, this -- this new wave of technology has been -- is an amazing opportunity to improve the pace of innovation and the pace of change. Clearly, we are not -- and we shouldn't be satisfied with the level of customer experience, with the level of efficiency because there is a lot of legacy that we need to fight against. And the legacy isn't just technology, it's the mindset. It's the culture. And I think this is also very, very important for us, we try to address this by being very clear about what we are pursuing in AI, what are our objectives. Our objectives don't have to do with headcount reduction. They have to do with growth. They have to do with efficiencies. They have to do with improving our risk management practices. So again, we have also followed a very methodical approach because if you don't have a solid foundation, you can't do much. So in all aspects, the infrastructure, the data, the governance. So we also invested a lot in that. And we were fortunate that this new wave of technology founders at a place where we had done a lot of the groundwork to be able and move faster. By the way, our chatbot, if you ask a chatbot to tell you about how to get a credit card and reply as a pilot, you'd be surprised by the answer or [indiscernible] English as well. So these are -- there are new things you can -- no one thinks about, but who knows they might be relevant to a new way of talking to customers.
And talent -- skills and talent, are you -- can you find talent easily to growth, advance?
No, I mean I think answer is no.
How competitive that it gets?
Everyone looks for the talented people. Actually, I'm very optimistic since yesterday, I was at the Gen AI Summit and there was a hackathon and there were some very bright young students that I think they were just finishing their studies, and they were given a set of data, reviews of the app. And they were given 2 hours just we're using a tool called [indiscernible] to build something that would understand the insights, visualize it and what they were able to produce in 2 hours was amazing. And the way they also presented it and the level of their confidence there. So I'm very optimistic that we can -- if -- I think if we focus a little more on connecting the industry with the academia, I think we will.
Are you happy so far with the production of talent by the universities, the expertise that young people get?
So the supply still cannot match the demand. So we -- it is very important to be able and train and upskill your talent pool as well. And it's important for people to feel that they're part of this. I think there was a study by Oliver Wyman, the banking sector, financial services and communication are people are the most afraid of this change. People are afraid they will lose their jobs at a percentage that reaches 67% globally. So you need to -- if you want people to be on board and be copilots with you in utilizing these amazing opportunities, you need to address their fears.
Okay. Dr. Barardi, I'm coming to you now. You hear us well, I hope. So you share these steps. You share the progress, and it didn't take that long, like 1.5 years, 2 years. What do you feel these incremental steps and these small successes, the small wins, if you call it small win, I was very impressed actually. Do you think this is healthy for the banks or they should go faster?
Good evening, everyone. It's a pleasure to be in this Innovation Day. And I'm sorry I'm not there with you physically. So it's a good point. I got this question from my management during last years when I start setting up this global team leading artificial intelligence across the group. The question was we are going to be follower or leader. My experience during the last 5 years in UniCredit Group is that, UniCredit, as my colleague was saying before, as artificial intelligence, UniCredit has contributed with different local and significant impact in different domain of the bank with predictive AI, deep learning, automation, computer vision and so on. The moment AI with generative AI with large language model revolution start and give me the chance here to go even further in the future. The moment AI can start to create agent that can assist humans, augment humans with some superpowers it is no fair planning how agentic AI may change our workflow as a population, how we will change the back end, the front-end office, but also the relationship with the final customer. Going to your point -- coming to your point, so in UniCredit, we started in the last 5 years building a lot of scattered use case across the group. We learned, I would say, 2 years ago, that this approach does not scale because if AI transformation is actually a digital transformation and cannot happen in a fragmented way. So it requires end-to-end process redesign in the back office to make processes AI-ready because today, some processes are not even digitalized. So you need to review the process, redesigning the process to make them AI ready if you want to get the best outcome in terms of impact from artificial intelligence and agentic AI, you need also a new part for customer interaction. We may expect in the future that our customer will want to have with the bank endless live 24 hours context-aware conversation with the bank. As a customer, I would expect one day this to happen. You need to have the right foundation in place. Without foundation, okay, you cannot go fast. You cannot reduce the time to delivery with the technology that is evolving so fast. You cannot scale and you cannot have governance by design that in an organization like UniCredit with more than 10 countries and with the regulation that is coming with the Act, it's impossible to manage the scale up in a way that is sustainable. So in UniCredit, we have built an internal group-wide artificial intelligence platform share standard to scale across country, avoiding the duplication and ensuring compliance. So yes, small wins can drive transformation, but only if they are connected to a bigger vision and they are backed by strong foundation. Otherwise, the risk is that you risk a pilot trap and there is a long transition with, yes, scattered impact everywhere. But I have some doubts that this business case will turn now to be positive, okay, over the time.
Thank you. Thank you so much. Back to you, Mr. Tsarbopoulos.
Michalis, please...
Yes, it's okay, but it's a formal setting. All right. So Michalis, I know you must be one of the people in Alpha Bank feeling most of the pressure to move fast from -- and I know -- we all feel one way or another the same pressure, Believe me. So it is not so easy, though. I mean you do -- you are banks, and we all expect that you comply with all sorts of regulations and risk aversions. So how do you balance this? What is exactly the kind of dialogue you built?
Yes. Let me start with the positives. All right. What we have going for us is readiness and technology and maturity. Why? Because of the things I mentioned. AI is not new to us. We are experienced. Agentic AI is a natural evolution to traditional AI also from an organizational perspective. Also, we have great partners, right? You have Microsoft, you have IBM, EY, EPAM. You have huge technology providers that have been investing a lot of time in millions, if not billions of dollars or euros in not only developing the technology, but tailoring tools specifically for the financial services industry. Why is that so? Because apparently, the potential value creation is enormous. So we have a lot of tailwinds, if you like. On the other hand, as you...
They are nice. The tailwinds are nice.
Correct. On the other hand, can we do that full speed without control? We cannot. We heard Alessandro talk about foundation. Eleftherios talked about data. Let me talk about what I call foundation. In order to be able to scale AI and to do it in a productive and prudent manner, you need a foundation that consists of 4 parts. First of all, people, right? Right now, technology is progressing a lot faster than our people can catch up. So we need to make sure that we continuously educate and train and upgrade our people so that they do keep up with the progress in technology, number one. Number two, cost controlling and real value realization. We talk about value, right? Value generation in any organization, especially in banks, is not about the technology itself. It's about corporate discipline and about tight governance. So unless you do the right things in the right way, again, we had -- they talked to us about a domain-based approach. Alessandro also said the same thing, not go piecemeal case by case, but go in a certain domain and rethink, redesign the whole way of operating that domain. This is about design, about governance. This is about choosing what to do and how to do it. So you need to measure what you build, you need to measure what it costs. You need to measure adoption and usage and most of all, you need to measure real value generation. So this is the second condition. The third is data. You need to ensure that you have all sorts of data, structured and unstructured, that you have it validated, curated and readily available in a secure manner by all AI tools. And that brings me to the fourth point of the foundation, which is security and AI responsibility. So it goes without saying that unless your systems are fully secure, and unless your data are fully encrypted and protected, there's only so much you can do. So we are spending a lot of time and a lot of investment in elevating our defense systems against emerging cyber threats on one hand. And on the other hand, we're putting in place the quality checks and controls that we need to have to ensure that AI is used responsibly, that models are unbiased and in general, that we use the output of the models and the agents for creating value not only for the organization, but also for our clients.
Now you're tempting me to ask a question that is not planned. So how about your quantum readiness, post-quantum cryptography readiness.
How many hours do you have?
Just give me 2 words.
No one is ready for quantum. No one really understands what quantum is.
Not really, not yet.
It is one of the major topics of the Singapore FinTech Festival.
Exactly. Yes, exactly.
Agentic were the top.
So maybe next year, we will be talking about this.
Maybe or the year after, we'll see. But to conclude, you were 100% correct. These are opposite forces. On one hand, there is technology that -- and corporate pressure that pushes you to move forward fast. On the other hand, you need to have control in place to make sure that you do it in a productive manner, in a protected manner, in a prudent manner, okay?
And a question from Mr. Kororos. Maybe I call left I don't make discrimination. So even though I suspect your answer, are you for stricter regulation?
I think that regulation can provide clarity, so clarity on how we do things. We need to think of regulation as an enabler, not really an obstacle. So it can give us clarity. It can also help us have a level playing of things like risk management, for example. And of course, I'm not talking about only the Greek setting, but also pan-European setting. But somehow, we need to find the right balance not to block experimentation, okay? And the truth is that...
There's competition. This is why I'm asking. There's competition, it's the U.S., it's Asia.
The truth is that the world is so connected nowadays. But even if it doesn't happen in Europe, it happens elsewhere, and it affects us.
Yes. So I mean this is -- these days in the domain of defense innovation, everybody talks about a rapid adoption. And there are regulations and ethics there as well that are maybe even much more important to consider.
So I think that regulation needs to be in place, but we need to regulate more what is more important and less what is less important. And we need to follow a risk management approach. So where is -- where are the big risks of leaving it totally free? So, regulate that. But where are things that we don't mind if we fail somewhere. So we need to enable experimentation. That's -- and the truth is that we see the U.S., we see Asia, we see China in Asia, pushing experimentation with Gen AI, et cetera. In Asia and China specifically, they have even open sourced their Gen AI models because they are afraid of the criticism of what may be in the model, et cetera.
The black box.
Yes, the black box. So things are moving fast. There is a danger of being stuck in Europe just trying to regulate everything and...
And a question for Harry now. customer experience, I was visiting an older relative and she was trying to make a bank transaction over the phone. And she keeps saying [indiscernible]. She wanted to be understood that she wanted to speak to a real person. How do you think -- how is society perceiving? How are your customers? How are they responding and not only in their everyday transactions, but for financial advice as well, which is the heavy stuff.
So I think customer experience is definitely the new frontier for all traditional banks because it's not just the people that are not so tech savvy, the people that are older and they are now -- they need to do everything digitally and you need to explain to them and make it easier for them. It's also -- I think the younger generation they expect a totally different level of experience from banks. And who knows, maybe in a few years, I think Gen AI, they only use their AI assistant now for everything. They will use it for shopping. They will use it for who knows what in the future.
So the rest of us also.
And probably everyone shortly. So how do banks address this? It's not a trivial matter. And I think all the investments that we've done in the past, the infrastructure, the security, the trust -- and of course, in banking, trust is the currency. You cannot -- you spoke in the beginning about perhaps why are the banks innovating faster, which I think banks are doing many things that are not visible perhaps to the wider audience. But one reason is that you cannot jeopardize trust, and you need to find the balance. This is why this whole -- I think this is why a lot of banks are doing this program to try to innovate with young people, with partners, but when the time comes to operationalize something, then you need to follow a much more rigorous approach.
So Dr. Bernardi, how ready are the customers to trust AI? Today, I mean, you think a bank can use as a marketing tool, the fact that they are utilizing AI models and they're advancing their AI technology. Is this going to be an attractive marketing approach?
Well, if there is -- yes, if there is clarity with our customer. So trust is not assumed. Commercial Bank has UniCredit with other banks in a complex landscape with all the emerging fintechs trust is all for us. So we need to be transparent with our customer how they're using their data and how we are using the data to provide them insight for offering them a customized or personalized service. We need to be transparent with the client. We need to be -- to start being transparent also with our employee and our front line of our customer. And artificial intelligence, we think will not replace human, but I think that the humans will be replaced by those humans that learn how to interact with Agent. So at UniCredit to start being also transparent with the population, the front office, we are using an AI copilot for bankers and start providing personalized digital experience for customers. As my colleague was saying before, so predictive AI in the past 5 years ago or more, 6 years ago, already power next best offer churn model, helping our colleagues in the branch, helping bankers anticipate needs. So agentic AI will take this further because it will enable hyper-personalized product, maybe one day real-time pricing, but transparency and explainability are not negotiable, not only for our customer, but also for our front office. It happens to my team in the past to more difficulties in having machine learning model outcome adopt by the front office than building the machine learning model itself. Because whenever you are providing a tool to our customer, to our front office in the branches, you need to explain then the outcome of that model with agent AI that will be using this model as a tool will is a matter of transparency and explainability will be even more important. The governance framework that we put in place also in line with AI act principle ensure that every, let's say, AI recommendation not directly to the customer, but to the front office before going to the customer is traceable, fair and compliant. So will customer trust AI? Yes. when it delivers value and clarity.
Sorry, I think we've run out of time. I'm giving notice that we've overrun actually our time for quite a few minutes. Thank you so much for participating. It is true that the competition is going to be fierce, resources limited. The talent, very difficult to get. So I believe next year, I'm hoping to see you guys more involved with AI factories.
writing code ourselves next year.
AI factories, we will go back to writing code because [indiscernible] there are many incredible start-ups being created that can be a valuable alliance for you. And cloud sovereignty, what you do with your data, all these are very -- and of course, post-quantum cryptography and cyber. So see you soon for more discussion. Thank you very, very much.
Thank you all for this panel. Now a small change in the program. We will now invite the Minister of State, Akis Skertsos on stage to discuss on AI and how it reflects on society.
Dear friends, good evening. I'm by no means an expert on fintech, so don't expect from me advice on how we can develop AI applications in the banking sector. But I will try to guide you through the governance, technology, AI and digital strategy and how we have developed this since 2019. It is a pleasure to be here today, and I would like to congratulate Alpha Bank for convening us to discuss an issue that is not only timely, but also urgent. A few days ago, I had the privilege to visit Singapore with the Prime Minister, a country not larger than the island of Cephalonia with double the GDP of Greece and the per capita income of $100,000 per year that has embarked on its digital journey 10 years ago. As we were told by the ministers, digitalization and artificial intelligence are not an option. They are a necessity for survival. I was impressed by their broad support program for 250,000 small- and medium-sized businesses aiming at their digital transformation and utilization of artificial intelligence. I was also impressed by the government's strategic decision to create a strong mechanism for digitalization of the state in the field of GovTech, which employs nearly 4,000 IT engineers and product developers as civil turbines, very well paid. Their mission is to transition the state and economy into the era of AI in a way that leaves no one behind. AI for the public good is their slogan. And it fits perfectly to the concept of responsible innovation that we also share and apply here in Greece as our own national digital vision. In the context of government, responsible innovation is measured by how efficiently it distributes the tech dividend, the societal and economic benefits generated by technology across all citizens regardless of social class, access to power or background. This aspiration is deeply rooted in our Prime Minister, Kyriakos Mitsotakis, one that has guided Greece from the brink of exiting the Eurozone only a few years ago to becoming one of the continent's best performing economies today. The results speak for themselves. Greece is now consistently outpacing the Eurozone in growth, restoring fiscal credibility and building an environment where businesses can thrive and citizens can prosper. In an uncertain global landscape and while the Eurozone grew by just 0.9% in 2024, our economy expanded by 2.3%, more than double the average. And unlike much of Europe, our growth is anchored in fiscal discipline, creating a primary surplus of 4.8% of GDP in 2024, while the Eurozone remained in deficit and also the fastest declining public debt in Europe. Fiscal stability and economic progress is just one part of the equation. The state has undergone a profound digital transformation since 2019. With the launch of Gov.gr, we have digitized public services and made the interaction between the state, citizens and businesses seamless. Building on this digital revolution, we are now implementing our AI strategy, a blueprint designed to help us leapfrog in sectors such as education, health care, innovation. It is no coincidence that Greece has been selected as one of the first countries to host an EU AI factory. Over the last 6.5 years, Greece has moved through 3 interconnected phases of digital transformation. First, digitization, building digital rails, universal connectivity and a unified government portal. Second, productization, turning services into user-centric digital products and mobile applications, redesigning state services around citizens and not bureaucracy. And now today, we enter the AI phase, embedding artificial intelligence on top of this digital foundation to deliver proactive, personalized and secure public services, modernize the state, strengthen democratic capacity and above all, strengthen institutional trust. These layers reinforce each other. Strong digital infrastructure enables applications and applications become exponentially more powerful when they are AI enhanced. We have executed one of Europe's fastest digital transformations by consolidating the state into one digital front door, Gov.gr, scaling from near 0 in 2019 to more than 2,500 services and 8 million users today. Citizens can now access their certificates, prescriptions, tax services, notarial functions and licenses digitally. This front-end shift was supported by nationwide infrastructure upgrades, 99% 5G coverage as we speak, 46% of fiber to the premises by 2024, enabled by EUR 3.3 billion under our RRF Greece 2.0 Digital Investments Plan, including major support for public sector services and SME digitization. Greece's ICT market has grown to EUR 8.7 billion and is on track to reach EUR 15 billion by 2030. The country becoming a regional data infrastructure hub by attracting major cloud and data center investments by Microsoft, Amazon, Google, DATA4 and others, which are establishing, as we speak, 18 and more facilities across Attica, Thessaloniki, Crete, leveraging Greece's strategic geolocation and upgrading digital infrastructure. Greece is building also our own sovereign AI capabilities through DAEDALUS, an 89 Petaflop national supercomputer and Pharos, one of Europe's first national AI factories, EUR 30 million cost, co-founded with EuroHPC. These assets support research, support start-ups, public sector AI pilots and critical sector uses such as health care, education and civil protection, ensuring Greece is not only a consumer of global AI technology, but also a producer of sovereign solutions. Greece is also transitioning from building portals, delivering scalable digital products and AI-powered citizen service. In our digital government ecosystem, there's a prominent place for our publicly produced digital identity and citizen services. The [ Gov GR ] wallet delivers secure mobile identity and verifiable credentials, ID, drivers' licenses, vehicle registries, health insurance, academic and disability IDs used by more than 4 million citizens. And this is the basis for a mobile-first state integrated with daily life and economic transactions and secure by design. Other AI applications in public services like Wildfire AI, a real-time AI fire detection system via cameras and drones, identifying more than 100 wildfire events last year or AI legal checks for our land registry mechanism, which automates and cuts review time from 2 hours to just 10 minutes, reducing costs from EUR 15 to EUR 0.14. Finally, our newest baby, a government-owned AI incubator for in-house AI delivery like the Singaporeans are doing, which is recruiting private sector technical talent to develop sovereign tools for the state. Its mission is to internalize AI production by building high-impact AI applications for national needs, citizen interaction and policy delivery at a fraction of the cost at a fraction of the time. But is it all rosy? No, it isn't. The state is only one piece of this puzzle. The rise of artificial intelligence has revealed something unprecedented up to now, the concentration of extraordinary power in the handful of technology companies, power and capitalization that, in many cases, exceeds the strategic capabilities of nascent states themselves. These firms operate as digital nations in their own rights. They command vast compute resources. The shape global information flows. They influence daily lives, even the mental health of billions of citizens. Their reach is immeasurable and their decisions, technical, commercial, ethical can have geopolitical consequences and can affect the fabric of our democracy. That is why we need strong, modern independent state institutions, institutions that are capable of understanding the technology that regulates fairly and keep the balance of power in check. There's a fine balance indeed to be kept between overregulation that suffocates innovation and complete deregulation that over dominates the public sphere by the interest of the very few and the very mighty, and we stand in the middle. We need the private sector. We need academia, we need civil society, and we also need a smart state. Still, we are not blind to the challenges ahead. Greece and Europe have a long way to go. We do face fundamental questions, and some of these are the following: Will we remain as Europeans consumers of technology or producers of technology? Will we limit ourselves to being regulators or will we also be innovators? Will Europe lag behind or will we lead? While the answers to many of these questions may seem obvious to those in this room, the reality is that institutional and regulatory roadblocks persist. The EU single market is far from complete. Inter-European buyers act as a de facto tariff of 44% on goods based on the IMF report. And this is why deepening the capital markets must become a European priority. For our companies to compete globally, they need access to deeper pools of interest of capital of more efficient cross-border investment and the regulatory environment that supports rather than constraints risk taking. The idea of a 28th regime, a harmonized framework that coexists with national rules should be embraced as a practical step forward. It would also allow innovative firms to operate seamlessly across the EU without being burdened by 27 different sets of rules, accelerating growth, integration and innovation. And here, I believe we do have a compass. Mario Draghi's report offers Europe nothing less than a road map for competitiveness in the age of technological disruption. His call for a more unified single market, greater investment in frontier technologies and the creation of a European scale set of champions must not remain words on paper. It must become our third agenda. Greece is ready to contribute to this vision, leveraging our own reforms, our own AI strategy, our own digital transformation, our own growing tech ecosystem to ensure that Europe does not merely adapt to the future but actively shapes it. Ladies and gentlemen, artificial intelligence is no longer a distant promise. It is here. It is already reshaping all sectors of life, state and the economy, the financial sector, your sector at remarkable speed as well. Across Europe and here in Greece, banks are integrating AI into risk management, lending, investment decisions, customer service, fraud detection and back-office operations. These technologies enable financial institutions to process vast volumes of data, take faster and more accurate decisions and automate high-cost procedures that once required extensive manpower. The benefits are, of course, substantial. AI reduces operational costs and increases productivity by streamlining onboarding, KYC services and compliance checks. But these opportunities also come with new risks. Complex and opaque models challenge transparency and challenge explainability. Poorly designed algorithms may embed bias or lead to inadequate predictions and overreliance on automated decision-making can erode human judgment and can erode the trust in institutions. This is precisely why the European Union adopted the AI Act, the world's first comprehensive regulatory framework. It promotes innovation while safeguarding citizens' rights. For banks, it establishes clear obligations, transparency, risk assessment, accuracy, robustness and continuous monitoring of AI models. Dear friends, I truly believe in the emancipating power of technology and artificial intelligence, but digital technology will remain aligned with the public interest only if its development is anchored in integrity, in transparency, in accountability and a sense of higher purpose for the collective good. These are not abstract ideas. These are the guardrails, which ensure that innovation expands opportunity rather than in equality, that innovation strengthens democracy rather than undermines democracy and that innovation serves humanity rather than shaping it in ways we did not choose. Thank you for your attention.
Thank you, Minister, for honoring us today with your presence. And now it's time for the pass of FinQuest, the pitch event where innovation meets opportunity. FinQuest by Alpha Bank is an international innovation competition that supports start-ups and scale-ups worldwide, fostering innovation in banking and beyond. This year, FinQuest 2025 focuses on artificial intelligence across 5 key themes: banking innovation, family ecosystem, travel, smart living financial tools and employee training. After carefully reviewing applications from over 20 countries, 7 finalists were selected to take part in a 2-month accelerator program where they received mentoring and guidance from Alpha Bank executives, venture capital partners and experienced founders. Our 7 finalists represent the very best of this year's cohort, visionary teams harnessing artificial intelligence to redefine how technology empowers people and transforms industries. The top 3 teams will receive awards from our partners, while the ground winner will have the opportunity to pilot their solution in Alpha Bank. Before we meet the teams, let's take a moment to introduce the FinQuest Judging Committee, the panel of experts who will be evaluating today's speeches. These distinguished professionals have dedicated their time and experience to assess some of the most promising AI-driven innovations. Let's roll the video and meet this year's judging committee. [Presentation]
So please let's give a warm round of applause to our judging committee. We are honored to have you here today with us, and thank you for guiding the next generation of innovators. Now come on. Let's clap everybody. This is the best part. Now let's get started. Please welcome on stage our first finalist -- sorry, disconnecting, deCODE represented by Michal Zkakis, the CEO. Michal.
Good evening, everyone. Hope you can hear me. Thank you for being here. My name is Michal Zkakis. I'm CEO of deCODE, the company that has built Dikaio AI. I have 25 years of experience in product, having worked in companies like Google, and I'm an ex-CPO of Workable. But tonight, for the next few minutes, I'm going to discuss with you the transformation that AI is bringing for the legal and the compliance teams for smaller and larger organizations and financial institutions. Before we dive in, just a quick word about our team. We combine deep legal experience, advanced AI engineering and product leadership for over 100 years of collective experience across world-class companies such as McKinsey, Google, Ernst & Young, Amazon, Mastercard, Intrum to name a few. So this blend is what is allowing us to build both reliable and innovative products at scale. Across Europe, regulatory and policy complexity is accelerating. Legal, risk and operation teams face a consistent inflow of new EU rules. At the same time, they need to accommodate internal mandates, Board Director decisions, policies, et cetera. Most of this work is still unfortunately being done manually, which means that research, drafting and reviewing needs to happen by the teams by themselves manually. There is no unified source of truth. There is very little automation and no reuse of past work. The result is predictable as most of you have certainly experienced, overloaded teams, long cycles, bottlenecks in procurement and credit, delays in strategic initiatives like M&As or portfolio transfers. With these problems in mind, we designed Dikaio AI. Dikaio AI is the first AI legal assistant train in Greek and EU legislation. Dikaio offers instant answers, drafting and compliance checks. It keeps legislation updated daily. It allows organizations to bring their own data, offers smart prom libraries and supports necessary workflows for the work of lawyers and professionals, things like translation of legal documents, anonymization of legal documents or validation of the compliance. Most importantly, it ensures enterprise-grade security with EU hosting, encryption and 0 use of customer data for training. For legal and compliance teams, Dikaio accelerates legal research, automates drafting and standardizes reviews. This leads to up to 50% faster research and document creation, reduced external counsel hours and stronger compliance with fewer errors or penalties. When we are looking at other teams, operational teams, like procurement, credit, risk, treasury, those teams handle heavy regulation workflows. Dikaio enables them to bring internal data into AI workflows to collaborate in private workspaces and use automated red lining and approval checks. This results in shorter contract cycles, fewer policy breaches and faster, more consistent decisions across the organization. In large organizations, such as financial institutions, for example, significant friction often appears in strategic projects, things like M&As, carve-outs, portfolio sales and spin-offs. For those type of projects, Dikaio can support cross-document research, can summarize obligations and risks, it can enable structured collaboration. The outcome, faster deal readiness, fewer legal emissions and optimize terms. There are 3 key differentiators that make Dikaio AI unique. We have built the only self-updating legislative infrastructure for Greece. This means that the infrastructure we have built is also automatically codefying legislation, which is one of the big challenges when it comes to Greek regulation, but it's also doing the same for you. So it's updating daily all Greek and EU legislation and core decisions. This allows -- ensures that the answers that the legal system provides are hallucination free. Secondly, it provides rich workflow features, things that the modern lawyers and professionals need, things like translation of legal documents, anonymization, red lines, policy checks, et cetera. Last but not least, it's built with the high security standards in mind by supporting encryption everywhere, only you hosting capability to anonymize documents, sensitive documents before publishing them and capability to deploy customer on-prem. This is why organizations trust Dikaio for mission-critical legal work. We have only been around since March 2025, so less than 10 months now. And things have scaled rapidly, thanks to our lean but exceptional team. We have more than 12,000 right now lawyers and professionals using Dikaio AI, hundreds of companies of all sizes. It has answered 0.25 million legal questions, and it's answering at a rate of 3,000 legal questions per day, which is close to what the Gov AI legal assistant also is answering. We have also drafted tens of thousands of legal documents. But this is not the end of our story, probably just the beginning. In 2026, we plan to provide more tailored compliance features for SMEs, corporates and the public sector. We will launch AI-powered Compliance as a Service modules, starting with the AI Act and GDPR, which is being redefined now with the new reality of AI data. And in order to allow companies to offload them from the compliance hassle and let them focus on growing and innovating. In 2026, we plan to bring this service to Southern Europe countries. In 2027, we want to add more compliance modules such as DORA or the European Accessibility Act and expand across Central and Eastern Europe. And by 2028, we aim to have the full EU coverage for most of the compliance needs for all sizes of companies. Our mission is to give every organization a reliable, secure and deeply knowledgeable AI assistant to remove friction -- legal friction to accelerate operations and improve compliance at scale. Thank you for your time. You can try Dikaio AI and happy to take any questions.
Congratulations. We will start with a question from Mr. [indiscernible]
Congratulations for the idea, the application, of course, your presentation. I have 2 fundamental, I would say, questions. The first one is which do you believe is your competitive advantage in order to avoid a heavy competition by copiers, by followers of good ideas in the future? And the second one is, why do you consider a bank, Alpha Bank in our case, a good candidate for this participation?
I will answer the first part of the question. Initially, for Greece, for example, we have built the infrastructure, which can update legislation automatically, something which is not trivial. It has not been done before, and we have spent actually months and months of manpower to actually implement this. But going further and looking at further applications also in EU, not only in Greece, we have a great experience when it comes to product, how to build, how to make things that are hard, are not sexy, how to make them easy for organizations and remove the friction there. So for example, AI is going to take everyone by storm. It's going to impact all the companies. They are going to need -- and there are not any decent solutions right now. They're going to need solutions which can help them offload this work easily without needing to spend hours and hours into deep diving into their processes and their documentation. So we believe that our experience in product, which has already been tested, let's say, in the Greek market as well as the different applications, tailored applications that we are able to build in order to provide a superior user experience is what would differentiate us from the rest of the competition. For your second question, we are actually in different stages with quite a few financial institutions when it comes to trying different use cases for our service. Different -- the compliance -- the legal and the compliance departments are probably the easiest, let's say, scenario where there's not a lot of implementation, let's say, needed for those departments. They're able to check policy compliance of regulatory compliance for big volumes of data. They are able to ingest all of the different contracts, for example, that they may have for customer data and be able from out of those to pull reports and generate, I don't know, credit risks or things like that. So there are different applications that can be used with this type of infrastructure, which is always up to date with regulation is able to process large amounts of data and large amounts of company files and be able to process those into legal documents and artifacts.
One more question. Just let me -- I will go to Mr. [indiscernible] because he is. Let me clarify the judges, we can have up to 2 questions, okay, from the team.
Congratulations. Based on the use of your system until now, can you report any productivity improvements or resource savings?
This is tough. It's a tough question because you need people to actually be able to measure those and report back and they need to already have this data available. So I could throw around numbers, but reality is that this is still early days, especially for larger organizations, they have larger sales cycles. We started 6 months ago, so it's only the beginning still. So I don't -- we have anecdotal feedback that things have been much easier for them. But yes, you have to take my word or their word for that. We don't have any data as of today for that.
Okay. A big round of applause for the goat. Thank you, Michal.
So now let's welcome our second team, Dry Runz represented via video link by Yariv Varel, Founder and CEO. Yariv Varel.
Just 5 minutes, and I will try to make the best out of it. So I will share my screen and we can do it. Okay. So what do we do at Dry Runz? It's actually very simple. We make people better. We believe AI is not going to replace us all and for sure, not in the upcoming years, but we believe that the opportunity or the real hidden gen opportunity in AI lies in making people better. Let's start with the team. I founded Dry Runz just a bit of like 18 months ago when Gen AI started to pick up. Before that, I created a different start-up by the name of Justed. I was Co-Founder and CEO, and we sold it. Some of the people from Justed continue with me, some are new. We are a small nimble team from all over the world, U.S. people handling go-to-market and sales, people from Europe, Sweden. We just hired a temp in Greece, and we also have someone from Africa. We are really global in the teams that we are looking. We're going to continue to be nimble. I believe any AI start-up will be like that because we are all augmented. We are all using AI, not just to develop it, but as part of our job, and I think it's part of the promise. Now what are we trying to solve in the world? You have salespeople, right? Alpha Bank will have salespeople. You will have support people doing their job. Normally, big organization will have quite a lot of people there. How do you make them better? Well, to cut to the chase, I have 5 minutes, if you can bring a live coach to speak with them one-on-one, that would be great. They will understand what they're doing wrong, where do they struggle and they can give them hands-on treatment for a long period of time. That's amazing. However, it's extremely costly. It's very expensive, right? To put a coach for every person, just the schedule of that, it's enormous cost. And so many organizations worldwide, and we're seeing it on a daily basis, have amazing plans on how to make their sales teams better, how to save cost on support. But it gets stuck rolling into the market. Another good example is think about communication skills. Can you say to your manager, be empathetic and they will be empathetic. Can you say to your salespeople, just ask questions, make sure you ask more than you listen. You can say it, but it will not happen. Why? Not because it's -- they are bad employees, just because it takes time and time is money. And the overall improvement of human beings at the workplace, of course, in sales, we also see it in support. We also work with health care institution, which is less relevant for fintech and other industries is huge. So what do we do? We have a simulator, an AI role simulator. So instead of talking with your prospect or with an angry customer that cannot do something, you speak with an AI. You speak freely in Greek or in any other language, and it responded not like an AI would respond in a mature coherent way, but as a customer. So you can -- when you create a simulation, you can create a disoriented customer and the customer will speak as the disorient person or as an angry person or I really love the passive aggressive ones that always stick it to you, all of those guys, so you speak and people love it. But the real thing, I don't have a lot of time, so I'll focus on the main thing. It's the instant feedback. We give feedback to employees on something like 20 different metrics. how much time did you speak versus the customer? How many questions did you ask in that time zone? How many filler words did you use? So many other things. A lot about communication skill, your empathy level, testing understanding, confidence level. All of those things are being reported back with suggestion how to improve it. When you finish a dry run simulation, you get missed opportunities, which is one of the feature people love about us. They go in, they click missed opportunities, they see the transcript and what they can do better. And we have abilities to help them create better sentences, et cetera, and they all do it on their own pace. So if one salesperson is struggling with something like that with a specific topic, they can get direct help on that. If some support person is struggling on a different topic, they will get their own support for that. And everyone learns on their own time based on their own tools and their own progress, making the promise of AI, making people better a reality. Let's talk a bit about the challenge behind the scenes. So when we created Dry RunZ, the first decision, the tech team, of course, really pushed for creating our own LLM. At the time, it was very fancy. We can -- we thought about it. Eventually, we chose to work with off-the-shelf large models, and I think it was a very good and economic decision. The challenge, however, it is to use those heavily trained model to act against their nature. So you want the AI to be angry. You want the AI to deliberately misunderstand what you're trying to say to make sure that you do it. And we do it all in live speech to speech, speech to speech is a very new technology. It's less than 12 years in the market. We cannot use live front because it's live. So the training that we can do for the model is very unique and we can do more unique things in problem solving, et cetera. And that's really what's making our company successful from the tech point of view, where we are able to harness all the great features of the large models and really apply them for the simulation. Some results. So what do you get if you use Dry RunZ? You close more deals. I think it's very simple. If you have a large sales team and you work with us, you can do better objection handling, better report building across the teams. You make more calls for sales calls each time, more calls for customer rep each time it's huge dollar saving. We decreased the tendency to rely only on live coaches. Live coaches are still part of the puzzle, but they are not the Achilles. They are not the main. You don't need to schedule with this person and that person, you're on vacation, they're on vacation, it never ends. So live cost training costs, which are huge are out. All of the organization and travel time. You don't need to travel to the main office to get training. You can do it in your own place at your own time where you have the time to train about what you need to train. So this is what we do. We offer the technology. We empower the team within the bank or any other organization to create their own simulation on their own. We focus on creating the best tech, the easy to use, the scalable and of course, the data part. And that's it. We have multiple customers today. We have launched something like 8 months ago. We are already working across multiple industries from multiple hospitals, credit card companies, food service companies and more. And if you would tell me 6 months ago that we will be in this place even presenting to you guys, I would not believe it. And it's an amazing rise, and I'm very happy that to be able to share it with you. So this is everything I can have say in 5 minutes. And now I can answer any questions that you have.
Okay. Congratulations, Yariv.
Thank you. Thank you. I'm so sorry, I wasn't there. Again, I have so many friends there in the audience, sorry. I'm really sorry.
Okay. So let's please take one question because time is up.
Congratulations for the idea and the product that you have developed. I have a question. You said that you have -- you are 8 months in production. So 2 questions. How much time do you get to train this model for clients? So how fast they can put it into production? And also with this you define actually the sales schools, all the education process that they are doing on their company. Give us one concrete example from one industry that you have tried. What is the real return on investment that they did using this solution?
Perfect. So I'll try to be as transparent as possible. If it's a model that we know like a cold calling or selling something or an angry customer, you can create your own simulation in 5 minutes. If it's something new, we are now working on a new more high-end knowledge process where, for example, for doctors about making sure people when they sell to doctors, they know a new scientific article, et cetera. It's a new -- there are differences between this and the other models. Each model is a bit different. So we train the model in the backstage for quite a long time. But then when we reach the production for a user to launch a simulation, it's 5 minutes. Maybe 10 minutes if you're first time, but it's easy, and you can change it and amend it each time. I will share with you a use case that is relevant for fintech for the second question. We are working with a large credit card company. You all know it and use it. Fortunately, I cannot use their name. They use it for 2 use cases. One of them, they sold the loan. I cannot say what type of a loan, but the loan was a bit troubling for the local authorities, and they wanted to make sure that when they sell this loan, they don't be -- they're not too pushy and they're not trying to really manipulate the audience, the users because the loan has multiple tricky parts, let's call it like this. And they used our AI to certify their salespeople to make sure that the person that sells this new innovative, let's call it, loan is not being pushed. They adhere to all the rules. We're talking about many people on multiple occasions. To do it manually or with the manager for each one would take months. It would be a very long and expensive process. With us, they did it in 3 weeks, maybe 4 weeks, if I'm counting also some early trials. So within 4 weeks, they trained hundreds of people in this way. And if you didn't pass your certificate, you can do it over and over until you pass it and you get the feedback. And of course, you get better as part of it. So that's an example of how you can scale training. In this case, it was certification. It wasn't training just to make sure that they follow the rules of that specific case.
Thank you very much, Yariv.
Thank you. Thank you.
So now before we pass to the next please know that we have 5 minutes for every pitching and 2 minutes in total for the questions, okay? So now we are ready to have with us here the third team, Nettle, represented by Richard Guga, Co-Founder and Executive Director. Richard, please.
Hello. My name is Richard Dugan, I'm with Nettle. For the first time in history, humanity doesn't scale. In fact, every day, every single second, we're losing billions of dollars. And it's not because of supply chain issues or interest rates, it's because of one simple unsolvable human problem, latency. The gap between a need and the perfect human response. The banking industry also suffers the same kind of challenge, lack of availability of specialist resources, the reduced operational hours of the banks. This causes engagement friction and lack of personalization. When you couple this with the competition of online banking, only online banking and the high cost of operations, it becomes a real challenge for the banking industry. So what if we could scale humanity? Our creativity, our intelligence, our empathy, not replacing people but extending them. At Nettle, we build digital human solutions that does exactly that. It shares the best of who we are, our knowledge, our values, our presence. Let me show you. [Presentation]
So that's not just a demonstration, that is reality. Let me introduce you to [ Vecna. Vecna ] was the first digital human banking assistant ever deployed in Europe. She was deployed in 2021, and she is still active today. She's part of a larger solution that was deployed to the Slovak Savings Bank. It was an integrated solution that had a core conversational AI platform as the engine, and it was deployed to 3 different user channels. One was an online avatar that was used for support of campaigns and product launches. The other was a mailbot that was deployed and synchronized with the bank's Genesys Contact Center. It actually delivers 85% of all incoming e-mails successfully, so hugely successful for the bank. And the third channel was the digital avatar. It was a 3D holographic image displayed at the bank in the branch, and it provided frequently asked questions and general guidance. It was also trained for transactional banking, meaning that it could block a stolen credit card for customers and also provide banking activities for that customer. Our recommendation is for Alpha Bank to use a similar kind of solution and gain the same kinds of benefits. We would recommend a more measured approach, one that builds trust over time. We would recommend that the avatar be deployed at the branch level as a meet and greet avatar, answering frequently asked questions, possibly scheduling appointments and even providing general guidance. A Phase 2 could be extending the capabilities of that avatar so they would have specialized support for operations and guidance at the digital corner. And ultimately, we would be able to extend that avatar even further for a specialized training or onboarding avatar for customers and employees. We believe the benefits to Alpha Bank would be huge. The 2 most important ones would be by deploying this really modern and innovative customer engagement experience, the bank would skyrocket in terms of brand identification and availability in the marketplace. But most importantly, by deploying the solution, you build once and deploy everywhere, meaning that not only are you maximizing the efficiency and the impact, but you're also maximizing the investment of the solution because you could deploy it to multiple channels and multiple branches at the same time, reducing costs as well. And this should be the music to the ears of every banker in this room. A little bit about Nettle. We are -- we have 5 co-founders, 3 IT people and 2 business people. It's the perfect blend of business acumen and IT expertise. Really pleased to be invited as strategic partners with some of the largest AI vendors in the global space, Google, Microsoft, NVIDIA. Also really proud of our customers who aren't only testing, but they're actually transforming their engagement models using our technology. And finally, just an example of a couple of the platforms and form factors. Here, we have digital HoloBox as a brand ambassador. We have a digital insurance agent or a broker on a mobile phone. We have a multilingual, multicultural avatar that maximizes inclusivity. And my favorite here, a live digital avatar deployed to your TV. Adam was deployed during the hockey World Hockey Championship in May. He was interviewed on a regular basis daily for over 2 weeks. It was a live avatar interacting with the live co- interviewers on the television screen, very, very impressive as well. So I'd like to thank everyone, and I encourage and welcome Alpha Bank to join us on this innovation journey. Thank you very much.
Thank you, Richard. Congratulations. Now let's start with a question.
Thank you very much. A couple of questions from my side. How your product -- how much your product is scalable cross country because I'm from Italy and [ need credit. ] So I would be interested to understand also if you have a plan in this perspective. And also consider that competition, you mentioned in the [ previous ] chatbots and similar solution is very, very high. What is the real advantage that you see vis-a-vis your peers?
Sure, sure. So we can provide support for over 42 languages where our technology was actually built in 2018, 2019 before generative AI. So we had to use very good knowledge and data scientists on our team to develop our natural language proprietary engine, so we could actually integrate easily to the other languages in our region. We're LLM agnostic, so we can integrate with any kind of LLM. We have serious business domain expertise because we've been delivering to the finance space and the online consumer space. So from a competition perspective, not very many people have that capability. Also, our digital avatar is also proprietary, very nicely synchronized and complementing our NLP engine. But most importantly, when it comes to interoperability and the innovation that we provide, nobody can do this. So we have speech to text, which is pretty common, but we also have gesture recognition, eye tracking recognition. We have a browser-based rendering model that increases scalability tremendously. So we've already got pilots that we're working on that are tens of thousands of avatars across different markets. And we're also providing other capabilities in that space, emotional intelligence. And when you bring that entire solution together, nobody can compete with us. So we're really pleased about that. Thank you.
Okay. I see we have very, very little time for a very quick question.
What is the actual customer satisfaction beyond the element of surprise that the customer has facing an avatar. And what are the actual gains in productivity and in financial terms for an institution that uses this AI avatar?
So much like some of the other guests, I can't share the actual numbers of some of our customers. But like I mentioned, the back office avatar actually successfully deployed or successfully answered 85% of all questions. But interestingly enough, Slovak Savings Bank, the largest bank in Slovakia, largest bank branch in Slovakia, largest bank in Slovakia, have the largest customer base, but it was considered a dinosaur when it came to technology. So from an anecdotal perspective, when we deployed this, all the other banks in our market and in our region started looking at this and they wanted something. And within a year, they all had their own capable avatars that were either online and some of them are actually experimenting with in-branch avatars as well. And the feedback from the bank has been very positive. It's, like I said, increased the brand identity in the marketplace. And now Slovak Savings Bank is considered probably the top 2 innovators in the market space. So from that perspective, it's great. And as I mentioned, you can offset the capabilities of -- and you can extend the capabilities of your team because of the lack of availability of specialists, you can actually have the avatar do the redundant recurring work and have the specialists actually doing the work. So you're saving time and money. And there's a lot of churn in the banking industry as well. So you're able to save money from that perspective. If you deploy this also as a training and onboarding tool, you would be able to actually, like I said, create one logic engine with your business domain knowledge and then actually deploy it to multiple channels. You can have multiple branches, you could have it in the back office training employees, you could have in the front office onboarding clients. So the potential is huge. Thank you very much.
Thank you, Richard. Let's welcome now, Agrinow, represented by Korina Chatzigeorgiou.
Hello, everyone. I'm very happy to be here and introduce you to my world. So from fruits to vegetables to olives in order for them to get to our table, they need care from human hands. And you know what, farmers can't do everything alone. They need help. And that's why they hire agricultural workers. Can you imagine what happens when farmers don't have the stuff that they need, it's a disaster. In Greece, in 1 year, only from the olive oil farmers, we lost EUR 27 million. And this is a problem reported in other countries as well, such as Italy, Germany, France, U.K., across various crops. So you might be wondering why this happens? Well, there are 4 different aspects. First of all, farmers still use outdated methods of finding workers such as phone calls. Furthermore, because those workers are coming from different nationalities, communicating in a different language is challenging for them. Without proper communication and with different hiring paths that they have to follow based on the nationality, it's very hard to navigate what to do for the paperwork. And as you can imagine, without proper paperwork, we have informal payments. And how do we know all this? Because both me and the co-founder, [ Kyakos ] are coming from farming families. Those were not just numbers and facts for us. It's the reality that we are facing every day. So we decided to do something about it. We did research, thousands of interviews. And we combined our educational and professional background across programming, economics and agricultural engineering. And together with a team of 9, we're set to solve this problem in the simplest way as it was possible. I'm happy to introduce you to Agrinow, a conversational and genting AI system, a place where farmers can find, hire and pay their stuff, a place where workers can maximize their employment and feel safe. But let's see how it works. Everything starts from a chat. Farmers simply have to explain what they need and the job listing is created. They are automatically matched with available vantage workers and language is no longer a barrier because everything is translated and transcribed in real time. Paperwork is streamlined and easy to follow steps that everyone can do it even if they have never done it before, but also the AI collects, checks and submits the document to the relevant public authorities when it is time to pay, again, it is one step away. So just like that, we made hiring as simple as chatting with a friend. Now AI can take care of the complexity and users just have to follow a chat. So far, we have gathered feedback for over 75,000 unique MVP users. We received the first place at [indiscernible] together with the support of Visa and Microsoft that supports our AI with the credits that they offered. We were featured in the Economist magazine, collaborated with the Greek Ministry of Agriculture and funded in cash by Google. The market is huge. In the world, 1 out of 4 people work in agriculture. We focus on the European Union and especially in middle-aged people that use the Internet and particularly in the countries, Germany, Italy, France. But that's not it because according to this year's world future of job reports by World Economic Forum, it is projected that the #1 largest growing job for the next 5 years are going to be from workers and laborers. And this makes sense because as the population grows, so does our need for food, medicine, clothes, all of which come from farms. And there is also another aspect. AI is silently replacing millions of jobs. And while some people will be able to get reskilled, some of them will turn to occupation that is harder for the AI to reach, which is laboring jobs. So what about Greece? Farmers pay each year more than EUR 1.2 billion in wages, but the system supporting them are stuck in the past. So -- but the majority of the payments are in cash. So that's like Uber that totally transformed the tax industry that was preliminary [ car ]. Agrinow is set to do the same thing and unify and streamline the entire hiring process, something that farmers and workers are eager to adopt. We are getting a fee for our services that includes the [ mass make ] and the paperwork and also the payroll transition, and we are here seeking a partnership with a financial institution or a bank in order to help people create bank accounts to exchange remittances and get access to micro loans. With the right partner, we foresee having EUR 60 million of revenue in the next 5 years. But what I want you to remember is that we're not solving only a laboring problem. We are building the infrastructure to bring banking services close to people that want to but get access it. We are reducing inequalities among different nations. We are promoting decent employment on the fields, and we are ending the food waste that starts from the field. So because we are using AI to empower, not to eliminate our communities, we are backed already by 75,000 people and by major companies. We are here asking you for the FinQuest prize because we help banks get to the beating heart of agriculture. And we honestly believe that with your support, we can help thousands of people that at the end of the day, bring the food to our tables. Thank you.
Congrats, Korina. First question.
Congratulations and the holistic approach. My question has to do with a specific segment. Usually, the workers and the farmers, they are not so digitally savvy. So how do you plan to ensure adoption?
Thank you so much for your question. Indeed, and that's why AI fits in because it can simplify complicated workflow. So it feels almost like talking with a human. That's our competitive edge. And so far, we have much more than 1,500 job placements, which looks like we are on the right path to do something extraordinary.
We have time for a second question.
Congratulations, you presented yourself and your partner in building this excellent application. However, which is the team that supports that?
Thank you for your question. So we are a team of 9 actually. We have the background in economics and agricultural engineering, but we also have a team that helps in product, UI/UX design, legal payments, et cetera. So we offer a holistic approach to this problem.
Now let's move on to Travelr, Yannis Dinapoyias. Please come on stage now.
Hi, everyone. My name is Yannis Dinapoyias. I'm the Founder and CEO of Traveler. The hyperlocal infrastructure for business travel, expenses and benefits powered by AI. I will tell you a little bit about our back story and how we are here today. Starting from my childhood, I grew up in a family -- in a travel family, basically, family-owned traditional travel agency based in Athens that was founded in 1974. My experience was there since I was a little kid. And my road took me to the U.S. to study economics and finance in Chicago. On my way back here, actually, home was calling me back. So I came here thinking business travel is a $3 trillion business, $3 trillion global business, and it's working on fragmented systems. So I was crazy enough to say it was accepted. And I was lucky enough to find this amazing team actually, out of all of them, only Eleni is here, but we pulled together decades of experience in travel and tech. We've all been basically traveling for life because of our different backgrounds and different works before joining altogether the team. And we've been rocking it according to us and according to some competitions in domestic competitions and international ones as well. We've been certified by all the needed bodies to act as an infrastructure layer for travel. And we've been doing it since the beginning of last year. Actually, in less than a year, we've achieved EUR 1 million in GMV in sales, and we have projected confirmed sales pipeline for next year of north of EUR 12 million, and we are still having 11 months ahead for sales. But let's dig into deeper in the business model and our vision as well as our mission. So business travel is begging for optimization. It's a global problem that touches us all. We've all been needing to travel for work, also travel for leisure. Problem is when traveling for work, lots of people in our organizations need to mingle in the process. So we need to research the travel options, then get approvals for it as well as passed by budget controls. Usually, prices change using the old way, the traditional way, the ways of e-mails and calls. At the same time, it's super frustrating. It's a complex thing that involves different providers, different vendors as well as suppliers, and it's very costly. So you need to outsource it to an agency, which means that we need to pay for fees or do it on our own, which is hours and hours and hours pre and post trip. At the same time, it's a huge opportunity because out of the hundreds of thousands of SMEs and mid-market companies in our area of Mediterranean and Southeastern Europe, 3 out of 4 people are still using calls and e-mails to manage travel internally in the organization until we were born. So we're presenting Traveler, the super app with 0 friction in travel, in managing travel expenses and benefits all in one travel ecosystem. We unite business travel, corporate expenses and employee benefits into one single unified environment, and we do it in style. So it's the first AI-powered travel expense and benefits super app in SME in Southeastern Europe and Mediterranean, reducing cost, saving time and automating ESG reporting for the whole organization. We are source agnostic. So basically, we function as the infrastructure on which all the whole ecosystem and the whole supply chain of travel works on. So hotel chains, airlines, ferries, transfer providers, you name it. So we are providing the platform on which the value chain of travel builds on. We have integrated roles, reporting, and we are super user-friendly if an organization chooses to use us via our UI. But our expertise and our specialty doesn't end there. We have some screens of the UI that you can see how simple it is to search something or manage bookings if you are a travel manager and of course, make sure that everything happens on budget. We integrate with different third-party platforms as well. So we need to be working together with ERPs and HR platforms that organizations are using already, and we are having a 2-way sync capability with those. Business model is very simple. So we are acting -- we're positioning ourselves as the infrastructure layer for suppliers, international suppliers and partners such as Alpha Bank as well as different partners on the supply side. And we meet them, we merge them together with end users, meaning people from different size of organizations from SMEs up to mid-market as well as enterprise. How do we reach them? Of course, that's why we are here. We reach them by -- via the AI agents that they can work on. So if people are working on Slack or Teams or Micro Copilot, they are using our APIs to serve travel content. How do we do that? By acting as the brain. So business benefits and expense, all is embedded by design into the traveler brain. So we are basically converging the brain for approvals, the brain for HR and expense into one embeddable AI brain that is being served to the companies via the AI agent of their choice. We combine global B2B rails, global product sinking different providers from all around the world as well as our hyperlocal service network in Greece as well as the Mediterranean, in ferries, et cetera. [ So superstar with Travel Compil AI ], which is the breaking -- the backbone of our solution is that we are source agnostic, meaning that we can serve our travel content via all different AIs out there. Competition is traditional travel agencies, but we are inviting also travel agencies via white label solutions as well. And we are also in enterprise -- we're on the other side of enterprise solutions like SAP Concur, Egencia, BCD and TravelBrick because we are the only solution in the market with AI-first design serving banks, fintechs as well as corporates of all sizes. We have done 80 customers are already channeling their travel buying through us in Greece as well as abroad. And we are a deal amount of north of 12 million into next year. Our revenues are coming from 3 sides. So supplier side revenues, commissions based on the people -- on the companies that are participating in our ecosystem from the supply side, meaning airlines, hotels, et cetera. We have a subscription model for clients and companies and as well as a payment rail coming from interchange. Market is huge. We have shown that already last -- in this current year, in a few months, we've signed north of EUR 1 million in GMV, having confirmed north of EUR 12 million for next year. And our mission is high and our calling is huge to be answering to a $3 billion market by 2030, achieving EUR 1 billion in sales by 2030. I would like to thank you for your attention. And sorry for my stress. We're building for travel. We're building for Greece. Thank you so much.
So now 2 minutes for questions.
Two quick questions. The first is my impression is this is a very crowded market globally. You said something about Mediterranean. You need to explain to me why this needs to be local. There are companies like obviously, SAP Concur or RAMP or Expensify. There are tons of companies in this space. That's the first question. Second question is, when you say 12 million GMV, this is total merchandise that goes through your system. I mean, what is your margin on that? I would expect very, very little.
That's the second question. You have to wait for until I answer the first one. So yes, actually, it's a very competitive market. It's actually why our team is so passionate about working on this problem. It's a global problem and lots of companies from different angles are trying to answer it. Expense platforms is one. RAMP, like you mentioned, is one of the expense platforms that are Expensify and PayHawk. There's all these different, even [indiscernible] business is doing that, is actually trying to answer how to manage expenses for business, categorizing them, auditing them. But what they're missing is content. What they're missing is most of them, the brain behind the transactions. So we are positioning ourselves in front of them, actually -- sorry, behind them, serving them with travel content globally. So RAMP is actually pulling from travel vendors. PayHawk as well. So we are in the position of serving that content and the brain that each company has of who is working in which department, who is approving whom, who is checking whom, who is the auditor of the travel budget of each department. That's information that lives -- has to live in a tidied up manner in one platform. And that's what we're doing, okay? Yes. And second question is, of course, profitability. Because it's been a bootstrap company since the beginning, since our birth, we are very focused on lean, being lean and having a really rock-solid unit economics strategy. Actually, we have managed to be self-funded and client funded in some ways of changing the model from a credit facility, which is the norm in business travel to prepayment. And we've been able to do that because of our value proposition and of course, because of our competitiveness. We're keeping our margins low. It's a volume business for us. So we want to be the alternative for hotels and travel providers to big OTAs. So that's why when we are onboarding a hotel chain or we're onboarding an airline, we don't take a big chunk out of their selling rates as Booking.com does, for example, which is something that they are very angry about. So we go in a more fair model.
Thank you. I believe we're out of time. So thank you very much. Congrats. Okay. Let's have myTeam represented by Dimitris Sereleas.
I'm really happy to be here. My name is Dimitris Sereleas, I'm the founder of myTeam. myTeam is an ecosystem that helps sports clubs and academies to optimize their daily operations, either this has to do with registration of athletes, either this has to do with communication with parents, coaches and the whole -- all the other stakeholders or this has to do with money collection, which is one of their main problems. Some things about our traction. Until now, we have a bit more than 400 clubs that work and use our main application, myTeam. Among them, we have [indiscernible]. We have 160,000 profiles within the app and 100 different families. We have 20 million of GMV happening within the platform, and we managed last season to convert 1 million of this amount of subscription paid towards the clubs in online payments through our marketplace system. And one other metric that is important for our business is the number of page views that we have on our mobile app, which is about 2.5 million per month. It is similar to a midsized news portal. Our team -- the founding team is George Malamatlis and myself. We worked with George more than 10 years together. We have worked in several tech business projects, either with start-ups or with large corporations, even with banks. We have been funded from VCs in Greece and abroad. We have raised until now EUR 1.6 million. And in our cap table, we also have Rasmussen, who is the Co-Founder of Google -- Google Maps, sorry. And which is the problem that we identified. There is a huge market. Only in Greece and Cyprus, we have more than 12,000 potential clients, including academies, sports clubs, dance studios, pilatus and stuff like that. And all these organizations use outdated systems to manage their daily operations. And they use not only outdated systems, but they do not interconnect these things together. They need to have one Excel seat for the registration and another one for the payments. They need to use Viber to communicate with the parents or the athletes. And I think what you get the point. So what we decided to do, we decided to build an ecosystem where my team application is the operating system. It's where every detail goes, and it's spread among other applications and things where we share knowledge and data. Where is the innovation and AI edge in our company? We use AI internally, of course. Our sales rep is an AI built system that sends more than 500 call e-mails every day to potential clients, while we also have synthetic AI editors who create customized content for our users, either it's an athlete, a parent or a coach or a sports manager. But we also use AI to create product value like we do with the AI copilot that helps within the myTeam app, the users of the sports labs to optimize their daily operations and create useful reports to talk to the AI agent and create trainings, create a subscription, get reminders about payments and having an assistant that they would otherwise pay even EUR 500 per month. We also have SportsLab. SportsLab is a second application we created in our ecosystem that it's mostly focused on the athletic development of the kids that use MyTeam, and it's sharing details based on their performance, analyzing them and then provide tips and tricks on how to develop their skills where it's needed. Last -- sorry, another one application that is under development right now it's CoachLab, where we help coaches to optimize trainings. We do that by creating smart coaching plans where they can add the number of athletes and the number of people that they're going to have in the training, the equipment and what they want to work on. And then they create the custom plan as we see. And last but not least is the [ Mighting Hub ], which is a hub that we create customized content based on the user, sport age type, different for athletes, different for coaches or sport manager. And it creates daily articles that we serve through our platform. Thank you very much.
Thank you, Dimitris. So let's have our first question.
Since I am a football fan, I believe I'm a fan of yours also. However, at the end, in which way do you believe that the cooperation with the bank will amplify your idea and your initiative?
Yes. There are many ways that we can work with the bank. The most obvious is the payment gateway. We already manage more than 1 million in online payments, and we expect this number to grow even more. we have seen opportunities like buy now, pay later. When a parent goes to pay for a camp, he might have to pay upfront even EUR 1,000. We see even opportunities with something like a factoring system. When you have to organize a camp as a sports academy, you need money upfront. So there are a number -- a huge number of opportunities for myTeam to make a pilot with a bank.
Thank you. We still have time for one more question.
So I'm using your app. So this was really interesting in terms of organizing when it comes to young children, especially getting on board on the team. Apart from the organization and the logistics, I don't understand what is your value proposition in order to continue and be this bigger with regards to all the content that you said. What is your vision on the next day?
Okay. So our vision is not to sell just the initial product, which was a Software as a Service and was serving the sports clubs. Our vision is to become an ecosystem of interconnected applications that will provide information to my team and will help the whole ecosystem control of the sports community, parents, athletes, everyone involved to be easily connected to find opportunities and easily manage their days. And also, we want to help more kids to be active and to spend time in the sports field instead of their mobile phones and tablets. In order to achieve this, we need to help these organizations to become sustainable.
Thank you. Now our last team for today eFrontiers by David Giron.
Hello, everybody. My name is David Giron. I am the CEO and Co-Founder at eFrontiers. And I believe Mr. [ Saltiz ] said it best. With all this AI, banks need to innovate. We need to predict customer demands, but how can we do that if we don't understand the customer. Well, that's exactly what eFrontiers is doing. But before I dive in, I want to highlight one of the biggest issues in the industry today. Today, over 67% of the world's population is financially literate. On the other side, financial institutions are spending over $90 billion every year, trying to promote products that people don't understand. So we have a massive disconnect. People don't understand their financial options and financial institutions don't understand how to better reach them and educate them. Well, the key is in education. So we created a gamified learning platform that makes it easy and fun to learn about money. You can think about it as a Duolingo of finance that is white labeled by financial institutions to promote financial education and boost customer acquisition. It follows a progressive learning journey where users will go on locking levels, starting with the basics, going to more complex topics. Everything is built in a modular micro learning methodology so that users can learn at their own pace and at their own time. We have gamified the entire experience so that as users are learning, they're collecting trophies, achievements, and they collect points that allow them to compete in real competitions that offer real rewards sponsored by the financial institutions like critical points, loyalty programs like the bonus, for example, or even books. Let's look at a real case study to understand how this works. We partnered with a large financial institution that was struggling with high customer acquisition costs and, of course, struggling with engagement. So the first thing they did is they select the features they want to include into their platform. We then build a curriculum that is tailored to their market and built around their financial products. And there -- our AI engine content allow us to create all the content optimized for our platform. But even better, we can leverage any of your existing materials, blogs, articles, videos to train the model and put that content into the platform. And in less than 2 months, we can be live. Now we offer this as a stand-alone solution. It can also be implemented embedded. But what we've seen is that there's 2 main reasons to have it as a stand-alone. First, there's no tech integration needed to allow it to be fast and simple. But most importantly, for new users to avoid friction, if it is a stand-alone, people don't have to download another banking app from another competitor to start learning about finance. What we've learned is that our partners by promoting our platform have seen up to 20x better performance of their marketing efforts. The best part is that once a user is learning about finance, they also start learning organically about the products of our partners. And now the real magic is what happens in the background. Our platform is learning about users, studying over 100 different data points to understand better their goals, their needs and the future requests. And it's not just about new customers. It's also about your existing ones. What we have seen time and over again is that most clients have a check-in and a savings account, and they don't really understand any of the other products that you offer. Well, through education, we can help them find the right product for their financial journey. And what we get is this result, an unfair advantage. While your competitors need to have market research, generic personas, we have a real-time data layer that allows you to understand your audience. For example, what's the demographic, what assets are they interested in? What are the goals? What's the risk appetite? And we take it a step beyond, and we actually narrow it down more granular to each user, understanding each user so that we can serve up with the content that they need so that they can improve their financial journey. We have a straightforward business model, which have a licensing fee on a monthly basis and a performance fee based on number of users and conversions. Today, we have over 100,000 users across platforms. We have tripled our ARR over the last 8 months, and we're now working with 3 financial institutions. The best part is it's not just about business, but we're helping our partners achieve their social impact goals. My co-founder, [ Harry Chan ], is a 2-time tech entrepreneur. He was building business simulations for institutions like Yale and MIT. My background is in wealth management and venture capital. Today, we are also backed up by 2 venture capital firms that are helping us with our growth. Today, we're here because we want to connect with more financial institutions who might want to learn and use financial education as the next big marketing tool. Thank you.
Okay. First question.
Okay. It's very clear what you do. And of course, you have a good success. My question is a bit different. Can you develop a system that actually can be used for training kids? In case we have a problem with financial literacy at that age.
Great question. So that's exactly what we do with our partners. What we do is try to understand what audience they want to target. And in some cases, they can say, well, we want an academy for the younger generation and one for an older generation. So what we will do is that our AI will tweak it, so they will make easier content. So something like Johnny has 2 Apples and so on, right? And then for the older generations, we can use more typical content. So yes, we can tailor it.
Okay. I see we have time for one more question.
Congratulation for the application. The business case is clear. What is your ambition for the future? What is your goal for the next 3 years? What do you want to achieve?
Awesome. So today, we hit 100,000 users. We're very lucky. It's been an amazing year. I think next year, we're talking to some of the leading institutions in Europe, but I want to take this worldwide. I'm actually originally from Guatemala, and I built this with the ambition that we can actually let anybody in the world learn about finance. If we really want to get people out of poverty and from difficult situations, donations is only like a tape. It doesn't really solve the problem. We need to give them the knowledge so they can make better decisions and have a better future. So for me, it is to get to millions of people, especially in those countries that need it the most.
Okay. Thank you very much, David. A big round of applause for all our finalists. Our [indiscernible] will now deliberate to decide this year's top 3 winners, and we will take a short 10-minute break. Get ready because the best is yet to come. [Break]
Welcome back everyone. I hope you rested well. And now we will turn our attention beyond the banking sector. From retail to industrial goods sector, we'll explore how AI is redefining entire ecosystems. Please welcome our guests for the next panel. Christos Karagiannakis from Kotsovolos; George Karagiannis from Moveo and Antonis Kyrkos from TITAN. This panel will be moderated by George Papadimitrakis from EPAM.
I feel the energy might be dropping a bit, which is normal after a break. But before we get started, and we go to our lovely guests that we have here, and we explore what's actually happening beyond banking, just to rewarm a bit the atmosphere. Let's give a huge warm round of applause for what we just saw from the 7 companies. Fantastic propositions and from Alpha Bank setting up organizing the event. And for all of you, that you're still around waiting to see who's actually going to be in the finalist run. So the main goal of this panel is actually to a bit trying to understand the different types of views from different industries. So today here with us, we have Christos from Kotsovolos. I'm sure the Greeks understand the difference in the industry land between Christos, let's say, George, who is representing Moveo, which is a technology company, and we can speak a bit about it. And obviously, Mr. Antonis Kyrkos, who is representing Titan Cement. So very, very, very diverse group, very different industries, very different aspects of experiences.
So the first topic that we're going to try to explore, and we'd love to hear your reviews starting with Christos, it would be around how has AI shifted, changed, transformed your industry, let's say, in the past 5 years.
I will start, let's say, with the bold statement that AI turns retail from a product-centric business into knowledge-driven business. So this is a totally different approach and fundamentally shift happening. So retail is being reshaped across every part of the value chain. And AI is not anymore an add-on or a password or even a tool that we need to use. AI redefines every aspect and transforms the whole operation end-to-end. And I believe that it is a very powerful driving force that acts simultaneously on three layers. At the front end, we are witnessing the spectacular rise of agentic commerce. So people will no more browse product by product. They're going to dedicate their journey into AI agents, that search, compare, evaluate and compare on their behalf. This is a fundamentally different approach and change the rules of engagement. In the middle, where we have the core operations, we've seen a shift. Actually, AI pushes retail operations into a real-time mode to be more specific. For example, forecasting is going to become -- became already actually hyper granular. Pricing is becoming adaptive to individual level workflows regarding stock are self-regulatory, naming just some processes. And at the back, the third layer, let's say, AI opens new frontiers for retailers because AI open ups -- give the ability to pivot into different adjacent sectors. They give the ability to retailers to monetize data and knowledge. And in many ways, it's not just about optimization. Actually, it creates new landscape and new...
Almost a new operating model.
Yes, exactly.
And Chris, just a follow-up question on this. Have you seen any, let's say, tangible results? It doesn't have to be KPIs. I know we use the term quite often. But have you noticed from your organization's perspective, when deploying those solutions, any tangible results as far?
This is tricky. Because I mean, the fundamental problem that all the businesses have is how you measure value, how you measure success, how you compare the value that you create against the cost, how efficiencies, let's say, is calculated in a way. To be honest, we are all the business, irrespectively of the states that they might exist. They are on the early stages of acknowledging and identifying the value that they have created. So for us, it's a journey that started. Of course, we have started -- we believe that we are early adopters, but we are not in that position in order to measure and give specific tangible benefits.
Fantastic. George, from a technology POV, producing a product that actually goes into different lens of clients from any type of industry, working with LLMs, Greek type of LLMs, which have very interesting, let's say, aspects you were saying. George was saying to me earlier, have you ever thought that people go in and type in Greeklish? Like it's a good one. It's like Omega is Omicron and i.e, everybody has a style. So how basically AI as a technology has been, let's say, impacting -- let's say, all of those industries that you're actually trying to serve from a product lens. And have you seen any sorts of difficulties in this initial kind of run over the past few years.
Yes, of course, 100%. First of all, thank you, everybody, for being here, and thank you to Alpha Bank for hosting this great event. So from my side of the equation, let's say, we are into, let's say, conversational AI and AI. So we have the privilege to work with those frontier LLM models every day. It's kind of a curse and a blessing, to be honest. It's a blessing because every day, we get to witness something extraordinary, but it's kind of a crush because our industry, the AI industry and the providers, it gets disrupted like every month. There is something new that we will have to adjust. So what we say internally in the company is that we're building the product, we're building the car at the same time, we're driving it, okay? And this is a challenge of its own, of course. But the promises, most of them at least are true. AI, I think that many people compare it to the industrial revolution, but it's going to be -- it's going to have twice the magnitude and it's going to be 2x faster, okay? So if you put this into perspective, maybe many people see that us moving a little bit slow right now, but it's going to move real fast, real quick. And this is why do we see that the technology is improving at such a rapid pace that I think every company, especially in the banking and financial services sector, moved from a nice to have tool into a must-have like I need to have this because without it, I'm going to be left behind, and I'm no longer going to be relevant as a business. And AI is not really new. It didn't come out with ChatGPT. I was doing my PhD in the U.S. working on machine learning and on the predecessors of ChatGPT on Bard and GPT-2. And my brother and I were in the U.S. during that revolution before ChatGPT, and we saw it coming. We saw a big revolution coming, and we're like we have to build something. We have to build a product that truly transforms businesses that makes them use the data that they have to automate, to personalize customer experiences. And that's how we started. And now like fast forward like 4.5, 5 years now, we see this becoming truly a reality that we have deployed GenAI agents to more than 100 customers worldwide. We're serving more than 15 million users right now every month. And we're doing this with foundational technology that we build and we build in-house. And we've seen the results in -- from customer service, let's say, automating from 50% to 70% to even 80% of the requests to proactive use cases where agents go out and reach out to customers to get them to do something that is revenue generating for the company. We've seen use cases in account receivables and debt collection that our AI agents are 20% more efficient than financial analysts. And this is like kind of the tip of the iceberg because the way we see it is AI is not really one single scope. It allows a bank because now we're talking about the financial sector to kind of like redefine customer experiences and use AI as the bedrock of the new infrastructure that they have. 100%. So disruption is huge.
Huge, and thank you for giving a few examples in terms of measuring and trying to give us, let's say, particular use cases around banking as well. Now shifting to cement, right, a very different, I would call it, like heavy industry. I wanted to ask Antonis, which is -- I don't know how many people would have this knowledge in here. So I think it's a good segue. How has this industry been transformed over the past, let's say, 5, 6 years? It could be from traditional ML AI to agentic to -- but how this has been applied in your case, has transformed the industry affecting, let's say, Titan as well.
Thank you, George. And it's true that I feel like a flying in the soup here in the middle of service industries, I represent heavy manufacturing industry. But you would be surprised to hear that this sector is also being radically transformed by digitization. And in fact, it's not only the 5 or last 6 years, Titan started applying AI 9 years ago. We were probably the second company in the world that applied artificial intelligence in its manufacturing facilities in its plant. And the transformation we see covers most of the value chain of a heavy industry like ours, primarily in the manufacturing sector. So how we produce the material in our factories, then on the supply chain, the distribution of the heavy material, very different from distribution of financial products, I can assure you, or even of white goods or electronics from the retail. And then also in terms of the customer interface, we are a B2B company with a very particular, I would say, traditional client base that is currently now gradually adopting digital technologies. And Titan's experience is that we have seen the biggest impact so far on our operational -- core operational domains. We have applied tools like optimization of the performance of our equipment in our cement plants. And to the earlier question, you said, can you give us numbers and metrics? I can give you very specific metrics... I can tell you exactly how much we have improved the productivity of our cement assets. The cement production assets has improved by 5% to 10% in the last 5 years. It may not sound revolutionary, but I can assure you in a process that has been micro optimized and fine-tuned for the last 150 years, a 10% improvement is nothing short but revolutionary. We have also increased our reliability by predictive maintenance and anomaly detection. On average, we are preventing something like 10,000 hours of downtime in our plants globally by predicting faults before they happen. We have reduced our energy consumption on average by 7%. That includes electricity as well as thermal energy. We have reduced 40,000 tons of CO2 emissions in the last 12 months. And going to distribution, we have increased the productivity of our concrete trucks by 9% and we have increased the on-time delivery of our concrete to the construction site by 10%. So if you put all of this combined, the business case for Titan and for the industry at large is a very compelling one. We have seen fantastic return on investment, and we are now moving towards transforming also the customer experience by moving away from the traditional construction site where everything is being ordered by phone or e-mail to modern digital portals and the ability to track your order and things that look very straightforward for any other industry, but they are totally transforming the way our customers operate.
I have to say these are very impressive results. Just a quick follow-up and just also looking at time, has this -- as you mentioned, as you started 9 years ago. So doing the math, that was the era hyperscalers, Yotta services, boom boom, AI, ML, but this was primarily in the U.S., right? Greece was a bit behind on that. How that -- sorry, I'm going a bit off script, but I was just curious.
No, let's say, the whole program was orchestrated from Greece. You're right in saying that our pilot plant was in the U.S. for the first time we tested the solutions, but we brought things in Europe quite fast. And it has been -- we are a global company, so we don't see ourselves as a Greek operation.
Fantastic. Okay. Christos, from -- let's say, from experiments to POCs as all of our tech friends here would call them, which some hate them, some love them, some find the value, some not. But from experiments to what we call scale of AI technology, what have you noticed in your own organization? How does that work? Have there been any lessons learned that you could share with the group? Good ones, bad ones?
Yes. Of course. First of all, definitely as early adopters of GenAI because AI has been always around the business for retail, even for cement industries. But as early adopters, we started with POCs, even with Moveo, we started some experiments in order to understand how things evolve and how can we use LLMs, GenAI under a new operating -- other new, let's say, systems and facilacilitate the experiences. But what we have seen is that we should treat this new technology, not as an IT product, not as an AI tool, but as an AI collaborator. So our own knowledge, our revamped or updated manifesto says that we should treat AI -- GenAI through the lenses of HR. So we think that this is a fundamental, let's say, choice in how we're going to implement and how we're going to move forward in the future. Why I'm saying that? Because for us, AI, it's not the next chapter of technological or digital transformation, but the beginning of organizational transformation. So we believe that we should put people in the loop by design and trying to create this shared accountability between AI tools and human. So we believe AI is a future collaborator. We implement the HR logic, and we have, let's say, created the term coined the term AI twin. What that means? It means that in the existing human organizational chart alongside this organizational chart, we create an AI organizational layer so people get people, colleagues or teams get their own augmented collaborator. That means that next to a commercial buyer, we have a digital buyer. Next to a pricing analyst, we set up a digital pricing analyst. And that means that people should onboard the AI agents, should train them, should mentor them and even promote them if they get new skills through this evolution of the LLM of the knowledge that they get. This is a totally different approach.
Indeed it is. It sounds like you're pretty...
It's a tough decision because...
It's a very interesting one.
But we've seen exactly as the colleague from UniCredit said, skated POCs and tools that act like black boxes with not controlled results. And this is not the way we will go forward, especially if we're talking about a customer-centric business like the one that we run. So this is a fundamental shift from the POC rationale or the approach to build tools in order to improve efficiency. We want to create a totally new layer alongside to the existing human organization chart.
So what we hear from Christos, and thank you for saying this is an interesting approach because in terms of -- if you had to pick the three layers, so technology, organization, people, talent, et cetera, et cetera, you're putting on your strategy right on the core of it and you're saying, as I'm defining my entire org chart, I'm putting AI in the forefront. I'm trying to define basically what sits right next to them and how to basically scale it up within my organization from part, not just do spot POCs and see what it works and then scale it up. Very interesting. George?
Yes. I agree with...
Now it's a section where we have to talk about the hardest parts of the...
Yes. I am...
We can also...
It was one of our debates with Moveo, I think, at the beginning when we were trying to collaborate on that.
Yes, yes. So no, our job is to make POCs production, okay? And 100% of our POCs have gone to production. That's our job. But no, kidding aside, I think that, to my friends, what I tell them is that you live in the epidemic of POCs, on GenAI. Everything is a POC, everything is a promise, until let's say, an agent hallucinates and the bank executive sees this and like take it out. It doesn't work. I cannot trust this. It's not compliant. It's not safe. How am I going to deploy this on my web banking? Well, you can't. If you treat the LLM as your product and you treat LLM as a black box thing that magically works, then you can't. And part of our job is to take this very powerful but still an auto complete -- a glorified auto complete, but still very powerful that we treat as an engine and we try to put it into a car, and we want to give the car to the bank, to the institution. Because at the end of the day, people care about the car. They don't care about the engine that we use. And we're very good at creating engines. We're really, really good at it, and we have multiple different engines, but LLMs I think of them, they're starting to become sort of like a commodity, okay? What's the difference between cloud and GPT and Gemini? Okay. Maybe there are a few differences, but still, on average, they behave the same. So at the end of the day, at the space that we're in, enterprise practical AI, we're trying to build those vertical agents that have actual ROI for the bank and institution that chooses to work with us. And when I said that we're living in the epidemic of POCs, maybe you've shared about the MIT study that 95% of the pilots never actually make it to production. This is insane. And still, even though there's been huge investments in GenAI, many, many companies don't yet see positive ROI when they treat the LLM as the core black box product, okay? So that's exactly what we're trying to do, just basically give the product and take the POC and make it into production. And Fun fact, we're working very, very closely with Alpha Bank, it is one of our closest partners. And we started working with them through a POC because obviously, first, you want to see until you believe. We started with a POC with Alpha Bank and Alpha Bank became one of the first banks worldwide, to deploy a production GenAI agent on their landing page and on their e-banking. And this is huge, okay? And we did it by closely working with them, closely trying to see that the data exists, the correct flows exist and the agent is compliant, it's safe and it's applicable for customer-facing experiences.
So to your point, keeping up with the speed, not just reading it as a POC that if you leave it alone, it might die, but basically taking to create value, take it to production, try to scale it.
Yes. Exactly. It's not magic. It's not going to work by itself. Garbage in, garbage out in general sense.
So shifting to you, Antonis, what has been the -- let's say, the hardest part because 9 years ago, we heard the story, very specific numbers, by the way, impressive in terms of ROI. I think very few organizations can support a similar one. But what has been the hardest part? Has it been talent, organization, technologies, lack of specific skills, adoption, training the people, making them utilize all of those capabilities, making them believe, what has been the hardest part?
So George, industrial AI is a bit different in that it is not plug and play. The models and the solutions that optimize manufacturing facility typically are invented or developed for other industries. And then you need to take that and adjust it and calibrate it and connect it with the operating system of a plant and adapt it to the way your engineers are working. So a proof of concept for us typically lasts more than a year where we have to take a model.
You say, wow, that's a difference, right? Great partner that exists here, friends in technology of banks, et cetera, that's a huge difference from an industry lens.
Back in 2017, our first optimizer took about 15 months to make it work. Our first predictive maintenance solution took more than a year of testing. Our logistics solution took the best part of 2 years. So that's, I think, the difference. And for us, making it work is the hardest part. Once we are convinced that it works, that there is a business case for it. Then immediately, we switch from experimenting and doing multiple other pilots to industrializing the solution and rolling out fast. And we have built a dedicated organization that is dedicated to actually take the solution plant by plant, asset by asset around the world and deploy -- even a deployment for a mature solution takes 3 or 4 months. It doesn't take 3 days. So you need a whole organization around it to go around and make it happen. And the final part here is that what we realized, and I think the secret sauce for us was that you need a matrix of capabilities to make industrial AI work. You need a data scientist that calibrate the algorithm and make it suitable for your facilities. You need a data engineer that capture the terabytes of data from your sensors and manipulate the data. You need the integrators to take this and put it into the machines of the facility. You need the project managers and then you need, of course, the change managers. And I think the right mix of skills was in the end, what made us being able to productize and roll out quickly.
That's fantastic. And again, wow for the POC aspect of it, wow for the patients and trying to figure out the right matrix and wow for the level of tuning customization that needs to happen to be able to work for the particular industry, right? I'm sure it will take a lot of it. But to your point, once it's set, having the right, let's say, maybe not manual, but the right skill set and capability in terms of the people enables this. So I haven't seen anybody bugging me about time. I see time is up. But if I can have one more minute and go to the reverse approach with Antonis, George and Christos. One punctual question, 1 minute for an answer from each one, if we can. How have -- we're talking so much about AI, GenAI, everything AI, so many events AI. Sometimes I think even the technologies are being a bit contradicting themselves because of the numbers, it's a hard part, right? Because you mentioned, it's very difficult to measure and so on and so forth. And it's objective because we're humans. So to your point, Antonis earlier, you were talking a little bit about the journey, right, of Titan, which is very interesting. One question I have is, how have the, let's say, expectations from a customer perspective, B2B on your case have changed if and to what level, if you will, over the past, let's say, 5, 6 years? Have you noticed anything different? Is this shifting as well from a business to business? Or is it still, okay, you don't have to digitize everything. We can live with that.
So as I said in the beginning, George, our customers are very particular types of companies. Most of them are small family-owned businesses, traditional in the way of thinking and valuing human interaction very much. So the way we have seen customer expectations and behavior change is, first of all, as they become more familiar in their personal lives by using digital platforms to do their daily shopping and commerce, they expect this also for their business. So I think one demand that came clear is that they want to interact with us on top of the telephone and the human touch with the electronic platform. And the second thing, I think the industry and the product is slowly but very surely decommoditizing. People expect service on top of the material. They expect predictability of time. They expect consistency of the quality, they expect ability to flexibly change the times and so on. And all of this cannot be done without electronic platforms. So I think this service orientation is what is changing for us.
Fantastic. Thank you, Antonis. George, is the moto, pressure, pressure, pressure on your side? Technology? Let's put it out. Let's make it work properly. No mistakes, by this date.
Essentially, yes. There's a lot of pressure. A lot of pressure to move fast, a lot of pressure to innovate. Expectations are the highest they've ever been. Obviously, you play around with ChatGPT, you think it's God, right? And you expect this kind of experience to give to your customers. So expectations from the business, so our partners have risen a lot, I would say, but there is still a lot of unknowns to be figured out around compliance, safety, how do I make this or hallucinate, especially in areas like financial services. But also from the end customers' perspective, I think expectations have risen again. There's kind of like we are in the transition period sort of and transition periods are always kind of weird in the sense that many people have gotten used to talking to the IVR or talking to the stupid chatbot. And when they see that it's an AI, they're like, "Oh, it doesn't work. I want to speak with an agent, agent, agent.
Human human human.
Human human, yes. Which is it's kind of the precoception of the end customer, but I think this is slowly changing and it's part of the exercise to educate the market as well by giving them actually good experiences such that users will come back -- customers will come back and still want to interact with the AI.
100%. Christos, same question, are customers now expecting not the three clicks, not the two clicks, one click, I buy everything, give me my recommendation. Why should I search? Why should I do this? Or has it been aggressive on the retail side?
Yes. I think retailing is on a Darwinian moment. Customers actually will no longer browse. So the main question for retailers is, are we ready to be found or trusted by an AI agent in parallel or more than a human as a human does. So this is very crucial, change everything. The old paradigm does not exist anymore. So we are in front of a totally different journey, a totally different operation. This is my view on that.
100%. Everyone, ladies and gentlemen, this is Mr. Christos Karagiannakis, CEO of Kotsovolos; George Karagiannis, Co-Founder, Chief AI from Moveo.ai; and Mr. Antonis Kyrkos, Group Strategy and Digital Officer for Titan Cement Group. Thank you very much for your time.
Thank you for this interesting conversation. Thank you George.
And now the moment we have been all waiting for, the FinQuest award. So the judge commitee have already returned. Let's get started. Today with me I have a very special co-host, which I will present right now, Temi. This is great to have you here Temi. So let's get things started. Could you please help me with the third prize. And while Temi is going to bring the third price, I will just inform you that the third prize is a start-up award pack offered by Microsoft. I would like to ask now Yanna Andronopoulou, Microsoft General Manager for Greece, Cyprus and Malta to join me on stage for the award presentation. Hello Yanna. Nice to be here. Good to have you here. Thank you. And now the third prize goes to Agrinow. Korina. Congratulations. Okay, now for the second prize -- we have it already here, and I will call on stage Mr. Primpas. Please, Mr. Primpas. Now Dimitris Primpas, General Manager for Greece and Cyprus at IBM. And the second prize is a prize of EUR 10,000 and it goes to Dry runZ. Congratulations. Thank you very much. And now -- before we continue, I want to remind you that all three winners will also receive three weeks of [ tech consulting ] from our partner. It's time for the first award, could you please give us the first award here. The first prize is consisting of EUR 15,000 along with an opportunity to run a pilot project with Alpha Bank. And this will presented by Dimitris Efstathopoulos, Alpha Bank's director of Innovation and Group Business Solution. And the first prize goes to dikaio. Congratulations. Congratulations again to all our finalists and of course, the winners and to every innovator who dare to dream bigger. As we close, remember the purpose of this event with its Innovation Day, Alpha Bank and this year's FinQuest Alpha Bank sparked meaningful innovation, brought together start-ups, industry and key ecosystem players into the same conversations on how to accelerate AI-driven transformation. Thank you to all our mentors and partners, supporters and the FinQuest community and of course, to you, our audience here today at [ STEGI ] and our audience at home via live streaming for being part of this celebration. Thank you. Of course, a very special thanks to my co-host, Temi and our sponsors. Now please join us for the -- for a celebratory drink. Good night, everyone. Thank you.
Read the full transcript via the API
You're viewing the first half of this call. Get the complete Alpha Bank S.A. transcript - plus 251,000+ transcripts from 12,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.
Get an API key View API docs →For developers and AI pipelines
Programmatic access to Alpha Bank S.A. earnings transcripts and 251,000+ others is available through the
EarningsAPI REST API and the hosted MCP server.
Quarterly plans from $105 - full transcripts, speaker segments, full-text search,
and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.