From Ferrari to AI-powered banking

Transcript

[00:00:00] Alla Gancz: Welcome to Pragmatism in Practice, the Thoughtworks podcast where we explore practical approaches to becoming a modern digital and agentic business. I'm your host, Alla Gancz, payments vertical leader at Thoughtworks. I partner with leaders across the global BFSI ecosystem to translate strategy into real-world outcomes. I'm delighted to introduce today's guest who has one of the most distinctive career journeys in financial services.
 

Piergiorgio Grossi spent more than a decade working in Formula 1 at Ferrari, moved into data, AI, and entrepreneurship, led digital transformation at Ducati, and is now Chief Innovation Data Officer at Credem. Piergiorgio, your career has taken you from environments where performance is measured in fractions of a second to a bank where innovation must also be trusted, resilient, and sustainable. Today, I would love to explore what banking can learn from high-performance engineering, how Credem is turning data and AI into real business value, and why people remain at the center of the transformation. Let's get started. Piergiorgio, good morning.

[00:01:11] Piergiorgio Grossi: Good morning.

[00:01:12] Alla: Thank you for joining us today. Before we talk about technology, how would you describe the common thread connecting Ferrari, Ducati, entrepreneurship, and Credem?

[00:01:22] Piergiorgio: I try not to think in term of differences or this time this kind of stuff. My job basically is reaching a company and helping the company through technology and not only in methodologies and what is needed to innovate, basically. Sometimes it's more about tech, sometimes it's not. I'm much more interested in understanding where they start from, which are the resources, which is the culture of the company, which are the tools and the resources that I can use to help the company to change.

Of course, industries are different, topics are different, culture is different, but still from my point of view, they are the same. It's more understanding how to adapt to the new place, to be in a way the outsider so the person that can bring a different point of view, but without being too much an outsider, because otherwise the culture will kill you in a way. If you think from this angle, there are not so many differences. They start from a place, and then you want to understand the direction and help them.

Again, the most important thing is which are the tools that you have. Sometimes it's money, because you have a budget or something like that. Sometimes it is people, because maybe, for example, at Ducati, I was the head of IT and not only for the digital transformation. I had the power of the IT team that could help me in understanding and physically doing stuff. Some other companies, you have the power because it's the board or the CEO that wants you. Depending on the tool and where you start from, you define the journey.

[00:03:17] Alla: Wonderful. It's your knowledge of technology, really, that is a common denominator across all the roles that you've had, and that knowledge of technology that you bring to drive, accelerate transformation. You're absolutely right, culture can eat strategy for breakfast.

[00:03:36] Piergiorgio: I'm at that level in which if you're a tech guy, you think, I'm not a tech guy, but if you are not, I look like a tech guy. Technology, especially in this moment, in the last two or three years, with all the generative AI stuff, I think it's an important tool that we have in our hands. Understanding it, being able to explain it, being able to play with it, to me, makes a difference, especially because you know when not to use it. It's so easy. In any company I've been, there was always a problem, and the solution was, "Okay, we can buy this software, and we will solve the problem."

I've never seen a problem solved in that way. It's never the case in which a software solved one of your problems. Of course, the software, technology, can help a lot sometimes. Usually it's more about a process or a way of thinking or a way of doing stuff. Then, if you are willing to make the change, then of course, software, for example, an agent today, the magic word agents can help you in solving. First, how's the process? How can you change? Do you have the courage to make the change? Do you have the tools, internal, the culture, to make the change? Then technology comes and makes the difference, of course.

[00:05:03] Alla: We're going to explore it a bit further. Let me ask you this question, which is, you once described yourself as a software guy and argued that banking is fundamentally a software business. Has your experience at Credem reinforced that belief, or made your view more nuanced? Is a modern bank becoming more like a technology company?

[00:05:22] Piergiorgio: I like to see myself as software guy. I have to say that thanks to generative AI, I'm much more a software guy today than I was 10 years ago, [chuckles] and it's very funny. It's an incredible period for people who love software, this one. Like Kent Beck says, software development is a lot about trust. Of course, it's about software technology, but a lot is about trust. I think bank is a lot about trust. It's all about trust, if you want.

That's why I think a bank is a software house from that point of view. Of course, it's a lot about people. It's about relationship, but again, developing software is about people. It's about relationship. It's about understanding the needs. It's about understanding how to help someone with a need. In a way, there are a lot of similarities. It's not just because money are electronic in a way that they are inside databases and application and blockchains and whatever, but it's also because it's similar.

When you help a customer in achieving his goal, when we help an entrepreneur to achieve his dreams or a person to buy his house, or whatever, you are creating a relationship, helping the person in the process. It's exactly the same that we do when we develop software. We understand the need of a person, and we try to transform it in something that's helpful for them. Yes, probably it's much more about technology now in the last two, three years because now, even companies that used to outsource all the software development and the IT stuff, because it was like water. It's something that you can buy everywhere, so who cares. It doesn't made the difference.

Today, we understood that-- I hope it's not too late. I'm not talking about Credem. I'm talking about in general. That, well, that's not exactly water. It's more an asset that, in some way, you have to be able to manage. I see a tendency, and maybe you see even more than me, in going vice versa. Banks, and not only banks, are saying, "Hmm, you know what? Maybe now I want to do that stuff here by myself, because now it's a little bit maybe easier or I can manage it in a different way than three, four, five years ago." That's a trend that I see. I don't know if you see the same trend. 

[00:08:31] Piergiorgio: That's not easy because, again, it's going to be a peak. We know about the stories. It's a bubble that everyone now wants to do software internally. There is this myth that say, "Okay. You have tons of agents that can write perfect software, you don't need engineers, you don't need anything." In a couple of years, we'll say, "Hmm, maybe not," and so let's mix and match. Maybe some things can be done easier, and can be maybe done internally. Some things, it's better that Thoughtworks do it for me because they know how to do it at a scale or at the complexity or at the security, safety level that is needed. We know the story.

[00:09:18] Alla: That's what we call enterprise AI. Vibe coding is very different to enterprise AI when you need to build and deploy applications and run them in production. We need to think about it holistically.

[00:09:31] Piergiorgio: Probably this software development life cycle, even if you don't write your code yourself, is going to change forever, because, of course, now it's even easier. It's something that, if you are in Agile, you know about. Now, it's even easier for people to show you what they think in term of running code, a prototype, or something, than in writing a document or in a Post-it or drawing in a whiteboard. For example, that's a point about trust, about working together, that AI, even if it's a technology, can help us in being, if you're more human, more design-centric. That's an interesting pattern.

[00:10:17] Alla: Absolutely. I agree with that. There's a lot of focus on providing context, and we call it spec-driven development with the agents. The context is becoming a lot more important as we capture those requirements.

[00:10:32] Piergiorgio: That's an interesting topic. A trend that I see, or at least something that we are trying to do here, is that a bank is extremely rich of regulation, documentation, manuals, and so on. We have all the rules in the world, especially in Europe. We have tons of that stuff that has been written for humans. Understanding how to first write in a way that makes sense more also for a piece of software, is one thing.

The other thing is how to make all this context, as you said, all this knowledge available to agents or to AI in general so that they can then handle data, not only software development, but everywhere in a way that makes sense with the rules that you have defined. That's another interesting challenge that any company has, especially a heavily regulated one, as I think.

[00:11:28] Alla: We're seeing a lot of focus on knowledge management, especially focusing on unstructured data where documentation can be scattered across emails, shared drives, and different PDFs and different formats, and that knowledge management is almost your critical piece of the puzzle, especially as we move into agentic development. Let's go back to Formula 1 for a second. I just wanted to hear your reflections on when people talk about Formula 1, they usually focus on speed. From your experience, are the more important lessons actually about data quality, feedback loops, and trust between people and teams? What are your reflections from Ferrari experience?

[00:12:13] Piergiorgio: First, Formula 1 is a lot about trust because the magic here is that between one race and another, that is, probably one week or two weeks. Talking about technology, probably, you change the car, or at least you changed a lot of the car in the past with different rules. Now probably a little bit less, but every race has its own car because the track is different, because the car is changing. It's getting better race by race. Of course, you change a lot of the software.

Being able to change so quickly, continuously every other week, you need a lot of trust in the team. Not just for the software, the same for the wings, for the engine. Trust is not enough, so you need a process of testing, of validating, and so on. That's why, for example, at the beginning of year 2000, when I started working at Ferrari, we had the chance to work also a little bit with you. We came back and so on, and understanding how we could embrace this agile stuff because we were agile in a way. How could we leverage that patterns to be able to deploy software in a way that makes sense for the race? That's an interesting point.

Another one is feedback loop. More than speed, being able to tune with the feedback loop that makes sense for your business. For example, in Formula 1, you have a race every two weeks. If you want to make a change to a software to the car for next race, you need to understand about it immediately, the same day that the previous race end or the Monday, because then you have five days to do something. You have to build your processes in a way that you get this feedback from the guys that maybe are flying back from Australia to Maranello in that case.

It's not the same for the car of next year, because while you build the car of next race, you also start building the car for next year. You need a different pace, a different speed of loop. In one week, it doesn't happen anything because maybe you need one month to make all the simulation for the new bodywork or whatever. That stuff here, aligning with the right feedback loop, to me, it's essential. I call it heartbeat. How you define the heartbeat of your organization? Maybe it's more than one, because again, there are things that change continuously, and there are things that change slowly. It doesn't make sense to make a daily walk, a daily something different. You understand what I mean. Daily update on something that changes every month. Aligning everything to this heartbeat, to me, it's pretty clear in Formula 1 because you go on TV every two weeks, so the heartbeat is just watching the TV. In regular companies, it's a little bit more difficult because you have to see it, and it's not clear sometimes. To me, it's a lot about this feedback loop alignment.

[00:15:45] Alla: This is so fascinating, just to hear the focus on agility and the need to be agile and communicate with business and the drivers and the team. That alignment at the heart of it is fascinating, and it's so relevant for all of us. It's also nice to hear that Thoughtworks played a role and you had a chance to work with Thoughtworks during your journey.

[00:16:08] Piergiorgio: Many times in different companies.

[00:16:11] Alla: Excellent. We'd love to hear more stories during the podcast. Let's talk about creating the innovation ecosystem at Credem. When you arrived at Credem, you were not simply being asked to introduce new technology. In fact, it was not just about technology, but it was about business vision and outcomes. You were helping the organization to build a repeatable capability for innovation. What was the real innovation challenge you found when you joined Credem? Was it a shortage of ideas, a shortage of execution capacity, or difficulty connecting ideas to business ownership or technology?

[00:16:44] Piergiorgio: Definitely, probably not technology. Technology, again, it's been super important. Probably not that all the technology that was needed was there and still is not here, but that's more about banking than about Credem. The first thing was to understand where to position our team, if you want, because there was no innovation team. We created an innovation team, the highest level in the organization. The idea was, for sure, there will be many people in the company saying, "Well, we are innovation. Why now do we need an innovation team, or what does this guy want?" For example, IT. We are talking about technology.

It was very important for me to create a strong relationship with the CIO that I still have a very super good relationship with him, and say, "Okay. You know what? Technology is your field. I will never be an enemy for you. You are the owner. You have hundreds of people. A lot of knowledge. It's nonsense. I'm not going to teach you technology." The same was about, for example, marketing. Strategic marketing, whatever. "You know about banks. I'm not the banking guy. I'm not going in your field, if you want."

We try to understand which were the field of play. One was, for example, understanding that, as every company, Credem was extremely focused on what McKinsey call (H1) Horizon 1 type innovation, so incremental stuff. We are super great. We are 120 years old, growing, being one of the best banks, more solid in Europe every year, so we know how to do it. The problem is that if things are going a little bit faster, maybe the H2, H3, so Horizon 2 or 3, or more disruptive innovation that used to come in 20 years in banking, maybe now it comes in 5 years, or 3 years, or tomorrow, like GenAI.

How do we prepare for it? Who is in charge of understanding it, seeing it, and bringing the company and saying, "Okay, guys, we have this new stuff. Let's go." That was our job. We first of all position ourselves like the H2, H3 guys. You are the champion in incremental. If you need me, I'm here. Like H2, H3, we are the champions because you don't have time. We bring you there, or we bring the future innovation more next to the H1, so you can eat it and understand it.

We created extremely distributed way of doing innovation. First of all, it's of course open innovation, so with the outside but internal. It's not like a team out of the company doing magic, and then trying to sell this magic to the old company. It was internal. We decided to stay inside and to create a network of, we call it innovation leader, any company call it in different way, agents or whatever, human agents that, in a way, are the innovation of that department, of the IT, of the legal, of the HR, of the marketing, of the CRM, and whatever.

We started asking them, "What's your strategy? How can we help in seeing the future that maybe you are not watching because you are super busy in the every day? Which tool do you need?" For example, we have a budget. If you want to elaborate and explore that stuff, we can help you. We have this money. We created processes like saying, "Okay. One thing is doing a two-year super beautiful project. One thing is exploring for one month a wallet to understand if it works, if it makes sense, what is a wallet?"

We defined official processes because a bank is something serious, and so you need a committee, you need a process. We created what we call a fast-track. With the fast-track, we are allowed to do things in a kind of garden where they are allowed to do that stuff with some money that we put. Then, for example, we started working with startups. Startups started to saying, "You're a bank, we need money." We said, "Nice idea." We created a venture capital initiative internally.

Now we are able to invest in startups. We have more than 40 startups in our portfolio in this moment. We invest in startup that make sense for us, so it's not a financial investment. Well, it is because we put money, but it's industrial. We put money in startups that can help us in doing stuff in a better way, in a different way, explore things that we cannot explore because we don't have time, we don't have knowledge, technology, and so on.

We created a place. I'm talking here from a place that's called Officine Credem, so Credem Garage in English. That's a physical place that's different. You don't need a tie to come here. You can have a hackathon here and sleep here during the night. You can bring your pizza here. It's not the office of someone. It's a place where everyone in Credem can come because it is a place that's makes sense to make innovation. It is a journey in adding tools, as I call it, to help the company, help the innovation leaders, help the managers to make their own innovation.

[00:22:44] Alla: What a great story. Credem describes innovation as culture, as methods, and processes, and business, not simply as technology. You've just told us the full story. You also recognize that perhaps only 1 in 10 ideas will become reality. How do you create permission to experiment while remaining disciplined about stopping the other nine?

[00:23:05] Piergiorgio: That is extremely difficult. We don't like to fail in a company like this that is made of very, very, very, very good people. We were talking about how companies are different. One of the things that, to me, Credem stands out is that it's made of very, very knowledgeable and good in every way possible way people. They like to do things in a good way and to have success. We struggle to kill activities, but that was another of the tools that we gave to the company, saying, "If we fail, it's my fault. I'm the guy who fails. I have a salary because I fail, so it's my fault." Sometimes it's not that easy.

It's a little bit simplistic, but in some way it's not because when I go to the board, I say, "You know me, and you know that I'm here to make mistakes. If I don't make mistakes, I don't find the right direction." A very good old friend of mine that's now the head of Audi Formula 1 Team, that's Mattia Binotto, used to say, "We have to break engines not in the race, because, of course, no one want to break engine during the race, but when we test them, we want to break because if we break engines, we know where the limit. If we don't break, we never know if we can get some more performance from them."

Innovation is the same stuff. If we don't break things, you don't know if you are at the right level of innovation. Maybe you are, but probably you are not. You're just doing incremental stuff. We want to break engines. It's all we want. We push, trying to break as many engines as we can, but it's not easy.

[00:25:05] Alla: It's great that you have a safe platform to experiment and test it out.

[00:25:10] Piergiorgio: Again, it was not something that you do like a pilot or under the desk, or something that in secret, or how we do experimentation, it's a secret stuff. We said it has to be official. We are a bank. We are the supervision of ECB. We cannot cheat. We have to be transparent. We need a transparent way of doing that stuff.

[00:25:32] Alla: Yes, but it's all about the learnings. If you fail, and the learning during that experience is most valuable, isn't it? Let's dive into and look ahead now and talk a little bit about the most exciting topic of the day, which is GenAI, agentic AI, and maybe reflect on the journey you have been on, which is you have described Credem's GenAI journey in two stages, explore and exploit. Many organizations are comfortable exploring. Far fewer build the mechanism needed to exploit the technology at scale. You also said that Credem has moved from exploration phase of GenAI into exploitation phase. What changes when organizations crosses that line, and what does it look like today? What are you focusing on?

[00:26:22] Piergiorgio: First of all, it's difficult. Even if everyone is coming to give you a good receipt, no one knows if that receipt works because it's few months that we have those tools. No one knows if, after two years, you die because the receipt was wrong. That's why it's difficult. That's why we like an explore-exploit way of working. That is not like saying we are in explore and now we are moving to exploit. It's a continuum. You are always both in explore and in exploit. That's the key to me. The explore-exploit is something that we read in a book that's called The Invincible Company, but it's a common way of thinking.

There are two phases. The first one is explore, is where you explore, you try to find good questions, and you understand if there is an answer to that question. We don't talk about business models or return, or something in that phase. We want to be sure that there is a question for someone that be an internal customer, external customer, and there is a possible answer. Then we go to exploit, that is, "Okay. Now we put it in production."

First, we say, "Is it needed the application or not? Do we have to build it or not?" Because maybe it's okay, but it doesn't help really our customer or our revenues or whatever. When you have this mechanism, because the difficult stuff in every big company is that everyone always think in terms of exploit. It's always like a big project, give me a business plan and so on, even if there is a lot of uncertainty. That's why it's important to keep the explore phase active and saying we have different processes, different way of thinking, different results they expect from the two phases, and then you start the loop.

At the beginning, GenAI was a lot about explore, because we didn't know anything about that. What we did was deciding that it was a priority. My team two years ago was just about innovation. Now is also about AI and data. We create a team because it's part of the information system, if you want family. We are not technically inside the box of the IT, but we are twins. The message to the company was, "Look, this is important," full stop. Another message because we are working completely 100% in agreement and together with the IT.

Then it was, "If it's important, we have to invest in it." Investing in Gen AI is first of all investing in people. Giving them tools instead of saying, "Wait, we don't know. The return is not clear." No, you have to go. Also, to defend if you want the company, because otherwise, people is going to use ChatGPT. That's the standard in their own personal devices, and maybe putting data of the company into the free tier of one of those tools, and we don't want that. It was a defensive but also an offensive way of understanding and then approaching Gen AI, saying, "Okay. Let's do all stuff that is needed to be sure we use a tool internally." This was a challenge, not not an IT challenge. It was a challenge for the IT risk, IT security, compliance, and so on, because we want to, again, be compliant.

We are one of the strongest and more solid bank in Europe. It's ECB that said that, it's not us, and so want to keep the posture. Want to do things in a good way. Being sure that all the boundaries are okay. Then helping the company and all the people in the company to make their own journey. Sometimes it's a personal journey. We call it everyday AI, saying, "How can AI change your everyday life inside, maybe outside the company?" Thanks to AI. "Where is your starting point? How much formal training do you need? How much bootcamp or icebreaking do you need?"

It's people in my team going to the offices and at the beginning and showing them how to do stuff. Then maybe after one month, getting back and say, "Okay. Now, let's work on your problems and how you can solve your problems." It's not someone, the IT central in our organization doing it. You can do it.

[00:31:14] Alla: Excellent. It sounds like you've already building the foundation for embracing and scaling AI in a very responsible, safe, secure way. It looks like Credem's AI agenda is now moving into everyday work, like you mentioned. It creates a great opportunity to improve productivity and service and so many other processes. It also makes your adoption governance and employee confidence central to success. Tell us a bit more about AI Next, which is the central hub connected to domain teams across the group. Also, where are you seeing the value coming from AI today? In terms of employee productivity or customer service or technology or operational processes.

[00:31:58] Piergiorgio: AI Next is the name of this team that we've created focused on AI. Internally, we have the innovation unit that it's helping the company, as I said before, not just in AI, but today probably 80% with AI. Then we have a governance team because again, we want to be sure that we do things properly. We have an AI team, meaning data scientists, engineers that can do some of the things that the teams need. Plus we have a data team, the guys that looks after the infrastructure, the cloud, and so on. We have all the capabilities to do the explore and exploit inside the AI.

What is most important is that we have exploited this concept of hub-and-spoke that I said also before with innovation saying we want to be sure that any team inside the company or any company inside the group because Credem is a group, we have Credemleasing, Credemfactor, Credeminsurance, so we have companies that are part of our ecosystem has their own capability of developing at least a little bit of something.

Here the goal is, "There is the everyday AI. Anyone has his own tools, can do it, can play with it, and get help from them." Then we have the big projects that we do centrally, the big end-to-end transformative, very important project. We have one, for example, for the software development or for the operation. There are two of those that we are now working on. In between, while I wait for the big transformative project that is going to come maybe next year, what can I do in my company of the group or in my office? Maybe I have the legal department, or I have the HR department.

While I wait a complete transformation end-to-end of my way of working that maybe will come in five years because there are other priority, what can I do? The idea is having the hub-and-spoke mechanism and giving them tools, rules, knowledge, so they can do something. We have what we call-- well, no, we are going to production. The idea is more like having-- we have what we call risk complexity matrix, saying there are something that you cannot do because it's too risky or too complex. We do that. The IT do that. There are things that are sufficiently local, where inside rules and technology that we provide you with the IT, you can be autonomous.

It's a challenge. It's difficult because you need knowledge everywhere, but it's still a direction because everyone today that is getting out from a university has some knowledge-- well, not everyone, but many, that have some knowledge in software development. Also, if you come from economics or whatever, maybe you know about statistics and so on. How can we leverage using AI, your knowledge, in a professional way, but with no risk, because you are acting on things that are not so complex and not so risky.

[00:35:26] Alla: My next question is around those guardrails that you need to put in place. Now you're going very big on AI and unlocking business value. Before a bank can deploy AI safely, what must be true about the data foundation, the architecture, the ownership, the governance models, security controls?

[00:35:45] Piergiorgio: Also, because there are tons of rules for the DORA, the AI Act, all the IT risk, and all the cybersecurity. It's really an incredible job of the IT department, the compliance, audit, and also of my team to be sure we comply with everything. Here, to me, the real challenge is how can we ensure that we are compliant. We want to be compliant, not just because there are rules, but because the rules often are good. We don't want to harm. We don't want to make bad things, and so rules help us in doing that.

It's more like, how can we create those rules in a way that doesn't completely stop the company? That's, I think the same for every bank of every regulated entity. Creating playgrounds or creating sets of technology in a way that we can say, "If you are here, you are safe because we have all the boundaries, all the data are okay, all the privacy rules are okay, and so on." Creating those, if you want sandboxes, but not for playing, for doing production stuff, it's important because you can multiply the number of things that you do because otherwise, for every application to start from the beginning, is the IT okay, is the DORA compliance stuff okay, is the IT risk okay? It's huge. If you create a place that you know it's okay, maybe it's not the best, but it's okay, so let's use it.

[00:37:37] Alla: We call it path to production. You create a repeatable path to production that works many times all over.

[00:37:44] Piergiorgio: Exactly. Plus, of course, all the work that we are doing, the IT department, they're doing an incredible job in reinventing their way of working is, of course, agents that can help in controlling, in shaping, in proactively helping you in being sure you are okay. That's another important stuff. Having this team of agents around you that protects you, control you, and quickly, in real time, it's another incredible stuff that can help us in not being completely paralyzed in banks. I'm not talking about Credem in specific.

[00:38:23] Alla: We call it a bit of a harness around it. [chuckles] Harness engineering. Let's talk about AI agents because this is now a great opportunity. Perhaps you're looking at AI agents executing more complex tasks. Where do you see opportunities, and where are you in terms of deploying, exploring this?

[00:38:50] Piergiorgio: Again, we are doing it from two different angles. One angle is playing in the everyday eye with the agents that everyone can build to help themselves in doing stuff. I'll call it small automation, or small intelligent automations. We are leading with the tools that we have. We have the Google system, Google Workspace, so we are not the Microsoft Teams stuff, but it's the same with Microsoft. It's not a matter of the tool. It's more like, "Okay. Which are the components that are there? Are they safe? How can we help everyone in the company, 8,000 to 7,000 people in using it properly?"

That's one of the directions that we are doing. Almost everywhere. No matter if you are in a branch or in the HR or in the credit risk, anti-money laundering, everyone is doing his own stuff. From the other side, we are facing it from the more big project architectural point of view. As I said, there the first area in which we are moving is the software development lifecycle, the operation stuff. These are the two first area. I would say the official answer should have been. We did a very detailed analysis, and we extensively understood all the benefits for all the area, and we decided those two.

That's something that we are doing, but things are moving so quickly that when you start the analysis, after two months, everything has changed, so the priorities are no more the same. Second is how ready is that area to receive this end-to-end change? It's because there might be a very good area in which you can apply agents, but for many reasons, sometimes technological, sometimes regulatory stuff, they are not ready, or it's much more difficult.

[00:41:03] Alla: It depends on the maturity of each business group.

[00:41:06] Piergiorgio: Exactly, but also regulations, because sometimes it's a tricky area. For example, HR stuff, you know that. Of course, it's about people, it's about the information about people, sensitive information, privacy, and so on. Is it the right place we have to start? For something, yes, we did something. For example, for recruiting, we have in place stuff with a startup for helping in recruiting. Maybe for some other stuff, it's better to wait to see how the AI Act is moving, how all the guardrails around our agents are moving, and so on.

To me, it's a balance. It's learning and doing the best thing today that can give us learning revenues, but also learning. Also, I prefer not to start something today that maybe in six months will be done overnight. Now, there are projects that we say, "It's a two-year project, but if the AI is going to this direction, maybe we can do it in the weekend. Let's wait a little bit." There are a lot of things that we can do. Let's start from something else. Even if that makes a lot of sense, it's a very, very difficult choice.

[00:42:25] Alla: It's fascinating to hear. I think agentic AI is bringing so many opportunities to transform not just the employee productivity, but we're building individual task agents, but also we're accelerating workflows, and we're building AI platforms that allow enterprise-scale enablement, so to speak. It's extremely fascinating. Let's look ahead in terms of future banking. As AI becomes more embedded in products and services operations, like you said, will banking become more intelligent, more agentic? What do you think?

[00:43:02] Piergiorgio: There is a range of stuff, and we have seen it with the new banks, for example, payments. That, of course, are going in one direction, speed, automatic. Maybe embedded into other applications like, I don't know, ChatGPT probably will be able to make transactions for us and so on. The other side is again about trust. For those stuff there, I want to be sure I can talk with someone, I can sit with him or her. I hope that he's going to use AI because I want to be sure that he can augment his knowledge to help me in doing my mortgage or my investment or whatever. That's the way in which I see it.

Probably interfaces are going to change. There are a lot of talks into the newsletters or on the social about this stuff of being able with AI to create an interface that it's just for me, just in that moment, just for that need. Instead of having applications that are configurable, that are the same for everyone, the interface is just composed for me because it's me. Maybe I'm 59, so I cannot see small characters that is different from my son that is 21, or because I have some knowledge that my son doesn't have, or vice versa. The interface is going to be created at a glance just for me, just for that transaction.

Probably, the direction of having a conversational everything is a thing. I don't believe much in voice but that's me because I'm old, probably. There were, as you know, a lot of attempts in the past. It didn't work. Now, of course, the interfaces are so good, and the models are so good that it probably makes sense. I prefer to have a chat with my friend instead of typing. Still, I'm not sure when I talk about business, about my money, I want to chat. Of course, I want to chat to have suggestions, to maybe have info instead of going to menu, blah, blah, blah, to see my account balance or whatever. I want to just ask and give me the balance. Still, I think interfaces are going to change a little bit.

[00:45:47] Alla: We're going into this transition from digital to agentic channels. Especially younger generations are relying on agentic channels to validate that they're doing a lot more research a lot faster. There's a lot more intelligence.

[00:46:00] Piergiorgio: The problem is that young people, especially the guys that are probably around 18 or something, they almost don't remember how was the world before the agentic stuff. Of course, it's just a small amount of people, but it's the people that is coming to the Credem, in the future, is going to be a customer of Credem. For the digital, it was about 20 years or something. Here, it was about two years or even less. They expect a different relationship, how they expect a different relationship with teachers, for example, at school.

[00:46:44] Alla: Correct. Everything, shopping, travel, everything. The consumer behavior is changing so fast that we need to be able to respond and prepare for that. A quick question on you're doing so much incredible work within Credem. When you work and make your decisions, how do you align, or how do you decide which capabilities Credem should own internally or which technologies to adopt from the market, and where a strategic partner can create the most value? Also, maybe share some experience of working with Thoughtworks.

[00:47:19] Piergiorgio: That's not something that I usually do a lot because it's more the CIO and all the other guys. I think it's changing. I'm not sure I have the answer. Also, because it's a little bit different for every company, so I don't want to spoil. Of course, again, it's changing because now there are capabilities that allow us to do things that were extremely difficult two or three years ago. It's still something I think every company is working on. Of course, one side of the equation is how the stuff is strategic for us. If it is strategic, you want control over it.

In the past, probably control was having a very good contract with the supplier. For every company, I'm not talking about Credem in specific, with very good lawyers and very good contract. Now, maybe it's going to be more about having good engineers, if we are talking about software, and a good engineering practice. That's going to be an interesting challenge for every company that is insourcing capabilities because one thing is managing suppliers, even if technical and even if staying next to the software, but one thing is producing software. You need not only developers, but also managers and a way of thinking that it's pretty different. That's where usually Thoughtworks can help because your knowledge in good practices of producing good-quality software it what makes you so special. It was in the past, and I'm sure it's now.

[00:49:11] Alla: Thank you. We're so proud to have an opportunity to be part of your journey and work together with Credem on scaling Agile, DevOps, CI/CD, test automation practices in the past a few years ago. We covered quite a few application clusters across the bank. I think what's interesting is that our work was focused not only on delivering technology, but on strengthening the architecture, the engineering practices, and organizational capabilities to sustain innovation. That's what we are proud of.

[00:49:46] Piergiorgio: Practices to me it's what makes the difference. I always say that I'm much more interested in the how than in the what because if you have a good how, I'm sure that is something very more intelligent me that knows a lot about the what. If we are fast, good, and safe, and so on in doing stuff, I'm sure we are in good posture to be good on the market.

[00:50:13] Alla: That's right. If you were advising a banking CEO in Europe, what is the single most important action they should take in the next 90 days to prepare for the agentic future?

[00:50:25] Piergiorgio: A lot depends on where they are now. What I used to say a few months ago was start now. You have to start now. How much are you using AI for your everyday work? How are you experimenting different tools? How you are playing with their hands? The beauty of this technology is that in a way, everyone can play with it. Three years ago, if you want to talk about AI, you have to be an engineer, you have to know about Python and libraries and machine learning and so on.

Now anyone can play with it. You have to start because it's so different. Thinking with your heads, as they say, it's something that makes a difference. Start now. Show your manager that you are doing it and you are expecting that they are doing it because I still see-- I'm not talking again about Credem, but in general. There are projects that start, and maybe you are in a room, and you talk about a project for two hours, and no one says AI and said, "Okay. Guys, I know it's the buzzword at the moment.

It's not that AI has to be everywhere, but you know that there is this stuff here. How this stuff can change the way in which you think of this project? Is this project good even without AI, or it is good just because of AI, and now it's possible or not?" Also, understanding how your projects are of these two categories, something that you can do even without AI or something they can do because of the AI in understanding.
 

[00:52:07] Alla: Well said. I agree with all that. Well said. Thank you so much for sharing your journey, your amazing work at Credem, and your vision. What I take away from our conversation is that high performance is not simply about moving faster. It comes from creating the conditions in which people can make good decisions, trusted data, clear accountability, rapid feedback, strong engineering, and a shared sense of purpose. Your focus on people is very, very profound. It was very clear that the future of banking will not be built by technology alone. The future will be shaped by organizations that can combine intelligence with judgment, experimentation with discipline, and speed with trust. Piergiorgio, thank you so much for joining us. It was such a fascinating discussion.

[00:52:57] Piergiorgio: Thank you. I will bring all your compliments to the great team, our CIO, our control functions, the cyber guys. Everything is possible just because I was in a way able to enter into this family, and they trust me, and I trust them. It's something that no one can do alone. It's their capabilities.

[00:53:24] Alla: Innovation is a team sport. That's a key takeaway from all this. It's a team sport, and we can see on how you brought everyone together, and that's connecting the dots between all the stakeholders in the organization. Innovation is a team sport.

[00:53:38] Piergiorgio: Thank you.