The Future of Disruptive Technologies in Finance with ​​Patrick Mullot

In this episode, Louise Newson speaks with Patrick Mullot, Head of Innovation and Disruptive Technologies for Investment Banking and Finance at BBVA, about the transformative impact of disruptive technologies in finance. They discuss Patrick's journey into his role, the importance of collaboration between tech and business teams, and the significant role of AI in shaping the future of banking. The conversation also touches on changing consumer expectations, the rise of neobanks, the challenges of integrating new technologies while managing risks, and the need for innovation to be user-driven with clear business objectives.

Disclaimer:

The views and opinions expressed by the guest on this podcast are solely their own and do not necessarily reflect the official policy or position of BBVA.

TRIPTK is your partner for brand-led transformation. 

Website | LinkedIn | Substack 

Transcript

Introduction & Guest Background

SPEAKER_00 0:00

The tech is a tool. The tech is not the goal. And many companies and in many departments of engineering, they usually make the mistake that thinking technology is the goal by itself.

SPEAKER_03 0:12

Hello and welcome to Codex Futures, a triptych podcast decoding how we live, learn, and play in the future. Each episode, we dive into a different industry to uncover the trends, cultural shifts, and powerful forces reshaping it and what those changes mean for the people driving innovation forward. In today's episode, we're thrilled to be joined by Patrick Mulow, Head of Innovation and Disruptive Technologies for Investment Banking and Finance at BBVA, a global financial services group headquartered in Spain. Together, we'll explore how disruptive technologies are transforming finance and how those innovations are shaping the future of consumer banking. Welcome Patrick. Wonderful to have you here.

SPEAKER_00 0:52

Thank you very much. Nice to have me here.

SPEAKER_03 0:54

Brilliant. So you have a very brilliant title, Head of Innovation and Disruptive Technologies.

SPEAKER_00 1:00

Could we start by I'm sorry, it's a little pompous, but yeah. It is descriptive of what I do.

SPEAKER_03 1:07

Could you, I mean, yeah, could you share a little bit more about what that actually means and maybe your journey into that role?

SPEAKER_00 1:13

Sure. Um, well, I began, I began my career on the trading floor, so in global markets and in in other bank, and then I moved to BBVA. I've been leading some teams in derivatives and sales and other activities very related to the global market activities of BBVA and other banks. And at some point I wanted to make a shift into a more tech related job. And um I started building a team that would be at the time that was, let's say, the hinge between business and and the tech team, the engineering team. Usually the engineering team they have a very specific way of talking about things, and on the business side, they also have a very specific way of talking about their business, and they don't necessarily understand each other very well,

Bridging Business, Technology & Customer Needs

SPEAKER_00 2:00

usually, from my experience at least. So there was a gap there of communication. And since I had both uh what's a skills, I was able to do the translating of managing the expectation, translating exactly what the business wanted to do with it with the to the tech guy, etc. So that led me to have a very much deeper understanding of how things are done on the engineering side, and also by my trajectory, uh how the business works usually in uh in global markets or or any kind of investment banking activities, I would say. And then, you know, a year pass, years pass, and the career leads you to some points. And at some at some moments, there is uh an opportunity to lead the innovation team inside a specific area of the investment bank of at BBVA. Uh, and and I took the job, and it's really, really interesting. So, not anymore in charge of just executing new projects, but more in charge of finding new ways of doing the business, mostly on the tech related. So, maybe finding new solutions to improve, let's say, improve the way people work, make it easier for them to work, make more efficient, and also enable them to offer a better quality of product to the clients or maybe a faster approach to the client, and at the end of the day, to respond to the client's need in a much better way.

SPEAKER_03 3:18

Fantastic. No, really, really powerful sweet spot, I guess, kind of understanding the technologies and trying. I really like how you talked about translating it in a way that means something and is understandable across the rest of the bank. You mentioned kind of innovation solutions and new and better ways of doing things. Do you have a past project that you're particularly proud of of a solution that you developed that you could share with us?

SPEAKER_00 3:38

Well, many, many.

SPEAKER_03 3:39

Yeah, I'm sure many. Could pick one, pick one.

SPEAKER_00 3:42

Yeah, I would say the project I'm I'm most proud of would be the one that was that were not necessarily very tech intensive, but more about working together with the teams, uh the business to understand exactly what were the needs and how we could solve that in an efficient and quick way. Because this is not specific of a bank company, any big company, very big company, you are usually drawn into bureaucracy and paperwork to do anything specific. So we find we found a way to build some solution around very light development that didn't need to go through the whole process of the bank for approval of those big engineering projects. And uh we've been able to do very quick deliverable to the to the users. They will really, they had a real impact in their day-to-day. So they really liked it, and and the adoption was almost immediate. And the more you give them, the more the more IDs uh arise between all the different users that you know you establish a communication with them, and then you know, step by step you start building much more complex applications, but that it that fits exactly their needs, which I'm proud of that because usually when you when you work with that many users inside uh inside of a company, so we're talking about 100, 150 users, it's very hard to get their attention, it's very hard to get all the data, all the information out of them and build something that responds exactly to what they are expecting. Because while you're still building something, the world is changing. So their expectation they are also changing. You're not that aware of what is changing. So when you deliver something, it might be outdated yet. So so it was it was a really great experience, and we're still working on that. So now it's a big piece of big part of their day-to-day business, and it's a very, very, very for me, it's a very huge success because I'm I'm still in touch with many of them and you know, get in in touch with anyone and grab a coffee and and they tell me, wow, what about doing this and this? And what we shouldn't should maybe we should change this for that, and maybe you could you have this dialogue which is fundamental, I would say that to be able to innovate, you have to be able to answer exactly what the needs are.

The Rise of AI in Banking

SPEAKER_03 6:00

Oh fantastic. So it was a process, was it, that you changed and you innovated? Was it or was it a product?

SPEAKER_00 6:05

What was the yeah, the process was imposed by the management, but but not the way to implement it. So basically, the the the change was instead of working in silos and in, I would say silos teams, so each team were doing one one just one kind of business, we would have to change the way we work and try to work in a pool manner. Maybe some people working very specifically, but then a number of people, maybe usually juniors, that have less specificities in their skills, able to do one type of product or one another's much more liquid resource force. And we would have to track all that. So all those people working in different projects and different products and different clients and changing all the time from one project to another, it was a bit a nightmare. So the first idea, as always in any company, is to use Excel, to use a spreadsheet. But from day zero, we realized that it was not possible. You can do that for five or six people, but not 100 people. And there's no way to do that in Excel. So we sat with all of the uh the different product managers and and and understand what were the specificities of their business, and then took everything into account and then build a solution and be very fast in the first delivery and be able to correct and to fix things that were not exactly what they were expecting, and go step by step and little by little baby steps each time and deliver small, small increment.

SPEAKER_03 7:28

Yeah.

SPEAKER_00 7:29

And at the end it it worked.

SPEAKER_03 7:31

Oh, fantastic. Oh, interesting. Interesting. It was it's not a tech solution then. It was a very kind of human-centered.

SPEAKER_00 7:38

It both, I mean, the tech part is also very heavy in terms of hours of work, but it's worthless if it does not respond to something that the users are expecting or or want or can be useful for them. Yeah, absolutely.

SPEAKER_03 7:52

Yeah, that makes sense. It kind of you know it's marrying what the technology can do with actually what's the human problem, what's the need and the use case. Fantastic. When you when we talk about all, you know, your role as disruptive technologies, what what are those, you know, are there some key areas that you're focused on? What what are those technologies that you're focused on?

SPEAKER_00 8:10

I would say now it's it's AI. There is no way around that.

SPEAKER_02 8:14

Yeah.

SPEAKER_00 8:15

Maybe three or four years ago, I would have said blockchain.

SPEAKER_02 8:19

Yeah.

SPEAKER_00 8:20

Probably with less enthusiasm than with AI. Blockchain is was a promise. It's still a promise. It's not really realized on the business side. I'm not talking about crypto stuff, et cetera, which is very stable and very I won't say it's mature, but it's a market that exists, and but it's not, it's not our business. AI is really present everywhere. AI is gonna change everything. It is so important that the bank has decided that from one of the five pillars of the growth for the next years, one of them is AI. So it is one of the most important things that the bank is doing is AI. Is trying to put, not trying to put AI like hammer it everywhere. It's not the point. It's it's just that we know that we're at the end of the story, we're gonna have AI everywhere. So better to think about it right now and think what we can do with AI that we are not doing right now, where it can really help. What also are the limitations, obviously, what are the risks of using AI, and what the different the different approaches of using AI inside a big company, because there are different aspects of it. You can, I would say mainly there are two. You can give your employees some tools, like any AI or LLM-backed chat bot, chat JPT, Gym and I, whatever brand you want. But this is a user, it's individual help. And on the other side, you can do also development using their APIs and and try to integrate AI into your processes in the bank to make automate some some other processes. So they are the two the two big sites, and there's some gray in the middle. But yeah, it it is it is gonna be something very, very, very disruptive. Obviously, it is already basically. You can, I mean, you probably know we all use it. Even if you don't want to use it, you're already using it. Even now, if you if you want to Google something, do you have an AI answer without even asking for it?

SPEAKER_03 10:13

Yeah, yeah, that's true.

SPEAKER_00 10:14

Yeah, it's gonna change many, many things, I think. And uh it is already changing some stuff. It is there are some very big risks, and it has to be uh very we have to be very conscious of what what AI is bringing on the table. Um, a lot of good things, but also as any new technologies, some some potential big risks. Remember the uh the the images of the Pope with the Versace coat, stuff like that. We all believed in it. Only can be worse now because AI is progressing.

SPEAKER_02 10:45

Yeah, absolutely.

SPEAKER_00 10:45

It's more powerful. So yeah, we have to find a way. But I I think it has always happened with new technology, not just AI. Every time there's a new technology, there's some fear about job loss, uh losses. There are fears about uh some offices that are gonna disappear, some job that's gonna be disappear, and some risk of fake stuff. At the end, we have to deal with it and and try to regulate what we can regulate and try not to regulate too much, but also regulate a little bit, and try to protect the users of misuse of any kind of technology, which is a very thin line, it's a very difficult in gray area because you know, respect pream privacy and free will, but also protect the people against their will. It's always that's the job of our politicians.

SPEAKER_03 11:30

Yeah, yes, yeah, yeah. And when you say, when you think of AI in the context of innovation, which I guess is your role, do you have any kind of opportunity spaces? Like where do you think AI and innovation, you know, what are the sweet spots for that in the the finance industry at the moment?

SPEAKER_00 11:46

Well, everywhere. To be more precise. I'm I'm more I'm more working on the investment banking part of the of the bank, but I would touch on a little bit of commercial banking that we that we all know, which is you know, uh the relationship that as individuals we have with our with our banks. It is obvious that we are gonna have less and less interaction with real people because AI can take the place of uh of some people for very basic stuff. We don't have to talk to anyone to search for anything inside the bank because AI will be able to go through all other information and find very, very exact what exactly what we need. They will understand what we ask. So obviously it reduces the contact surface that we can have with with a human interaction. But it's not bad in itself because we, I mean, at least in the banking world, this kind of reduction of the human number of people working at the offices has been already reduced for many years. So it's it's not something that is going to be drastic right now. I mean, this this is an ongoing process for many, many years. And people were really overloaded with a lot of administrative tasks that bring almost no value to anyone. Uh, you know, reporting to regulators and inputting data in spreadsheets, blah, blah, blah. Things you have to do, but you know, and all of this is going to be automatized. So people would have a much more, I think a much closer relationship with the customers for when the customer needs a real human interaction. I mean, just you know, call your bank or check your app because you want to know the balance of your account, absolutely zero value.

SPEAKER_02 13:22

Yeah.

SPEAKER_00 13:23

You don't need someone to do that. You can do it by yourself, or if you have a chatbot, you can ask and it will tell you. But now, if you want to set up a company and you need some, you know, at the beginning, you need some financing and you have to explain to the bank what the business is about, what is your business plan, blah, blah, blah. That could be done by an AI, but there will always be someone in the loop. The AI can't take this kind of decision and and can't. I won't, I won't, I won't see the AI take the place of some people there to take this interaction with your customers. This is where you you really add value, uh a very specific problem, a little bit more complex than just do a transfer. So on my side, in terms of investment banking, it's very similar, except that the transactions are much bigger, they are much a lot more and uh more complex, but you can also try to standardize that. So instead of when you're talking to the CFO of a big company, usually this kind of interlocution is very rare. It's necessary, but it's rare. And those people, those kind of people of profile, they are very, very solicited, so you don't usually have a lot of time. So it's much better if everything which is not really which doesn't have a really good added value that could be done by a machine. That's great. And then you take advantage of the time you have left with those people, with those high profile to discuss really to the point and strategic stuff and things like that. So you get rid of all the uh industrializable industrial uh uh issue that you can have about very standard standard products or standard questions.

SPEAKER_03 14:57

Yeah, absolutely. Yeah, I think we're seeing that theme of efficiency in in all industries, isn't it? So kind of AI optimizing and making things more efficient, so elevating the role of the human to the service that is more needed, yeah. Do you think I guess from I think we've talked a lot about, you know, from the bank's perspective, from the consumer's perspective, or the you know, the everyday person, how they're thinking about live, learn and play and how finances and banking is, you know, helps enable that. Do you think, how do you think that will change? You th how people are thinking about their finances, their banking? Is AI affecting that?

SPEAKER_00 15:29

Well, yeah, well, actually it is already. I mean, not it's it has happened before, before AI. We have seen a revolution with uh what we call neo banks.

SPEAKER_02 15:38

Okay.

SPEAKER_00 15:38

People like Revolute or banks like that. It actually responds to the needs and and expectation of younger generations, more used to to use, you know, mobile phone apps and schema like Netflix subscription, not take not giving that much importance about who is your bank that maybe my generation is more inclined to. I mean, your it's your bank, you've already, you always had your money there, blah, blah, blah, this kind of stuff doesn't make sense anymore. No, it's it's just a question of how great the app looks, how fast is the service, and how quick I can do and do what I want to do. And that's it. And if it doesn't work, I just delete the app and that's it. I just onboard another one. So it is very different from the from the world we knew uh, I would say 10 years ago, where we had to go to the office to open an account, and then you you know, I still know a lot of people that have a sort of a personal relationship with the director of the branch, people like that on the personal aspect. And for people, you know, for for the teenagers or maybe 20-ish people, this doesn't make any sense for them. It's it's something that is not their world. So it has to be fast, it's it has to be an app. You can access it everywhere. They are much more global people, usually they travel a lot. So being able to use multi-currencies account and to without any friction. I mean, on a traditional bank, you can also pay in in sterling or or whatever currency you want, but you have to transfer, you have to call, you have to do the exchange, it takes the time, it takes money. On those apps, it's completely transparent for the user. You know how much you pay. So yeah, it's completely different. AI will uh, I would say, maximize this kind of work for for the for for the for the for the client, for the customers, because they don't have any interaction with anyone, but they could not they just are able to do very simple stuff in neobanks, no very big complex. It's just very standardized because they're everything is automated because there are no human in the loop. With AI, I think they could be, they they will be able to do more complicated transactions because the interface will be able to understand exactly what the client wants, because using natural language, just explaining what it wants, it will be able maybe to find a solution. Yeah, I don't know, a student that is moving abroad to finish his studies or to do his PhD or a master's somewhere else, and just explain that to the to the app. And maybe the app will say, Well, I would maybe, I don't know, uh, in this suit in Zurich, for example, if you go to Switzerland and Zurich, the uh the real estate markets in this and that. So if you want to rent a flat, it's gonna cost you about this amount. We can finance that on five years, but it could just you know bring new solutions according to one simple story that you're not able to do right now. So I think it will be much a much uh streamlined experience with for the customer. Instead of talking about finances, they just explain their problem, and then the bank will give them a solution based on this request. Let's call it request, but you you're not requesting a specific financial product. This is my issue. I want to move to another country for two years, and I don't have any, uh I don't have maybe I have my parents could help me a little bit, but I don't have any guarantee. And then the AI would understand exactly what the problem is and try to find inside the catalogue of product of the bank what product could would fit the best for this kind of solution. Right now, to do that, you have to talk to someone. And this someone has to know exactly all the products available into the bank, which can all can as good as AI, because AI doesn't make any mistake on that case, just to query to a database. So the knowledge is always accurate.

SPEAKER_03 19:18

Yeah. Oh, that's interesting. So it's almost the role of a bank will expand beyond being finance and money first, actually to lifestyle first, and then the products are just the enablers of that.

SPEAKER_00 19:32

But actually, I know I know there was an example that that is always on top of my mind because at BBVA we bought a long time, many years ago, we bought a company that used AI to evaluate the real estate price of properties in Madrid, for example. So we have this data. So those guys they are always checking the market, but how much cost any flat or house cost in Madrid? And so if you don't actually go to the BBVA app to request a mortgage, you say, I want to buy the house and at this address. And then the app is able to know what what what what is this this house like and know how it works, how it's worth, how much it's worth. So it's able to know, well, how much money do you have in your account? I already know that, and how much the what is the price of this house? So they can offer you uh the perfect mortgage for your condition, your risk profile, and the property you want to buy. Then you only only ask for how can I, you know, buy this house? So it is the first step. But yeah, it will go further than that for a trip, for travel, for maybe holidays and to buy a car or whatever.

SPEAKER_03 20:38

Okay. I mean, in that sense, do you see like other platforms like Google and Airbnb, for example, as competitive? Because I guess it's who's the first entry point. Would you go? You can ask that question that touches on all aspects of your lives. But which app do you open first to be the first portal into helping you find a solution?

SPEAKER_00 20:56

Well, I don't know. But what I can say that in the banking industry, I think we've been worrying about Google or Apple coming into the banking business for maybe 15 years, but they've never did it. I think the point is you can't go into the big banking without going through all the regulators because banks after health, it's the most regulated industry in the world. And it's very, very, very cumbersome. Is there's a lot of bureaucracy and there's a lot of things to do, and a lot of things that you can't do. And for those big tech companies, if they have not done it, sorry, I think probably it's because it's not worth it for them.

SPEAKER_01 21:39

Yeah.

SPEAKER_00 21:39

It's much better to be on the side and provide our the bank with a lot of services that we are hiring from them, but we are providing the financing and we are providing all the interfacing with other regulations. Maybe at some point they will decide that it's worth it. Some of them have have built small banks, but for their own purposes, not commercial banks.

SPEAKER_03 22:00

Yeah. Okay. Makes sense. 'Cause I guess I'm I'm kind of curious. You mentioned, you know, neo banks uh the the the the now. What what do you think the in ten years' time, what will what will be the new revolute?

SPEAKER_00 22:11

What will be the new Well, I don't even know what's gonna be in one year. It's yeah, it's it's moving so fast. I don't know. Actually I don't know. If you look at China, for example, they it's true they have a very different market, but they can do what the the I think it's I don't remember exactly I don't remember the name of the competitor of WhatsApp over there. It's WeChat, I think. Oh yeah, WeChat. WeChat is a bank. You can do financial transactions inside WeChat. You can do everything in WeChat in China, but it is it's easy for them because it's state controlled and there isn't there's no competitors. So if you control everything, it's always, you know, it's pretty easy to innovate. You just impose it and that's it. It's not really innovation. But I mean, if you see, it is very hard to see the the difference now between the chat that you can have in Trade Republic or Revolute and WhatsApp or any kind of messaging. They have an or they have their own messaging platform where you can send money to people inside the chat. So maybe the next step is people like WhatsApp or or Telegram to create a joint venture with any kind of bank and enable payment through their platform. But again, both in the US and in Europe, there are a lot of regulations to do that. I mean, you have to be a registered payment company, but it's it's much easier than to be a bank. So being means being able to finance people to loan money. Payment institution is just moving moving money around. It's a very different business and it's much lighter. So you're you're not you don't have any balance sheet by yourself. It's just I'm taking your money and I'm giving it to someone else.

SPEAKER_02 23:46

Yeah.

SPEAKER_00 23:46

Just transaction. So maybe they would do that. Maybe there's a way to get some income also and get some valuable information also.

SPEAKER_03 23:54

Yeah, interesting. And what are BBVA doing, I guess, to kind of future proof the role of a traditional bank? Are you shifting into other areas or you yeah, I don't know, how kind of if we talk about you know the new model of a bank is changing, how much are you evolving to keep up with that or to hold your position as a traditional bank? What's what's the start?

SPEAKER_00 24:18

I don't I don't think, I mean, for the time being, it's it's more about efficiency than any other thing. So I would guess that using those technology and AI first, we will gain a lot of time because many, many things that were manual before now will be automatic. So we will be able to free people from a lot of of tasks so they will have more time to do things. So maybe we will improve on client satisfaction, or maybe we'll also be able to create new IDs and new products because we because we won't be, you know, uh that that involves so much involved in the day-to-day business and you know, never you know, taking a step back, don't have time to do that. So probably people will have more time to do that, take a step back and say, well, maybe we could offer this client this and this and mix with this or come up with new ideas. I guess this is the creativity that you need to do innovation and creativity, you need time. You you can't do, you can't say from I don't know, from 9 to 10:30, I'm gonna have this meeting, 10:30 to 11:30, another meeting, 11:30 to 12, I'm gonna think about things and then doesn't work. Doesn't work like that. You can't, you know, stop everything and and clean your head and and come up with brilliant ideas. Doesn't work the way at least for me, it doesn't work like that. Uh so I guess that's this is the way. And and I would I would say that we've been seeing in the institutional market some new products uh which are very, very similar to existing products, but much more complex in terms of volumes. For example, would you do I would go very, very, very technical, not very small technical stuff. There is a product in investment banking called syndicated loan, which is very easy to understand. It's just a loan, but the loan is so big that not just one bank can loan the money. So the bank is you get in touch with other banks, and all together you create a group, which is called a syndicate, and then all together you lend the money to uh the company. This is the typical product that, for example, company airline would use to buy a new fleet of planes. If they need maybe two billion euros or two billion dollars to be to buy new planes, uh you can't ask one bank, one bank for those two billions. It's too big for the bank. But maybe between 50 banks, we can do that. It's not a question that the we don't want to lend the money to this company because of the credit profile. We are okay with that, but it's so much money that we can't do it by ourselves. It's too heavy for us, shoulder broad enough. So usually when you do a syndicated loan, it's with maybe 20, 25 other banks. What about doing a syndicated loan with 2,000 banks? Same product, same credit issues, but in terms of management of that, it's a nightmare. We don't have the resources to manage everything. So each time there is a change, you have to communicate with all the banks, you have to do everything. So unless you mechanize everything, there's no way you can do this kind of product. Even though theoretically it's it's exactly the same as doing it for 25. It's just a scaling issue. So we are seeing a lot of products emerging right now that are like that. It's both some use AI, some don't, but it's it's it's using innovation in terms of how you can structure your systems to be able to afford these kind of volumes for investment banking and reporting, all the reporting that goes with that, and all the communication management, because we we need to have humans to manage the loan during during his life. And obviously, you have to talk with all the banks. If something is happening with the company, you have to talk with all the banks and you know keep them aware of what what are the changes or the situation. You could use AI to do that. If you can't use humans to talk to 2,000 banks every day.

SPEAKER_03 28:01

Yeah. That's interesting. So it's almost it's be that's using AI to scale and expand the current products that uh that exist. It's letting them happen on a much larger scale and managing that through.

SPEAKER_00 28:13

We actually have been thinking about new product that will that AI will enable, and we haven't found anyone. Any any any anything that is that is so intricate intricately linked to AI that you know that you couldn't do it with without it. There is nothing. AI is not in it by itself not a product, it's just a tool. It's just like the industry is like just fall when you change from manually doing something or putting it in a production chain.

SPEAKER_02 28:42

Yeah.

SPEAKER_00 28:42

That is AI. It's it helps you make things much quicker and be more maybe more precise. But that's it. So you can't say use AI to do a new product. No, you can't use AI to be more efficient, and so maybe you can scale your product, as is the example I just gave, or you can do more of the same product because you have more free time because part of it is automated. But an AI product in finance that doesn't exist. And I don't think nothing exists. There is no AI product except for the AI company that are selling their solution.

SPEAKER_03 29:13

That's a really I really like how you phrase that. It's a really interesting lens. It's there's no AI in itself, it needs to be attached to something else to supercharge or optimize. It doesn't exist in isolation, doesn't exist in the in the vacuum. It needs the it needs the eye to bring the, I don't know, something like that.

SPEAKER_00 29:29

It needs the you need, yeah, and you need, I think the um the thing that we are struggling with, not not just BBVA, not just not just the banking sector, is the integration part because it's going so quick, it's going so fast. We're not used to having to integrate those solutions to so many of our so many of our internal systems so quick. So we are not used to that, we're not prepared to do that. AI IDs arise everywhere in in the company. Many, many people have very good IDs, but to implement that always have to, at the end of the day, it's it's just AI gives you an answer to a question, and then you have to use this answer somewhere. So that's what I was I was saying before when I was talking about user, the user, the end user use of AI or the integrated use of AI. Let's say, for example, if you want to create a chatbot, a very you know, CD chatbot when the client can go into your app and say, what is my balance of my account? Then you would have to integrate this AI with the database where the account is registered. So you have to do this integration and you have to take care of all the security, blah, blah, blah, etc., privacy. So it takes a lot of time. We are used to do that very for very specific needs and very specific proof with a specific rhythm, which is pretty slow. And now it's going so fast that we, I mean, we have many, many AI products and not that many people that are able to integrate all those IDs into our systems. I think it's exactly exactly the same for a hospital, for an accounting company or an airline. If you put some AI into Amadeus, for example, the you know, the booking system, they will have to integrate everything with Amadeus. And all the company, the airlines would have to do the same, so it will be you know total mess at the beginning. So it is uh it is something which everyone is trying with. So we we have to be more responsive for that on the part.

SPEAKER_03 31:17

I mean, is that one of because you mentioned risks at the beginning of this conversation? Is that one of the risk areas that it doesn't get integrated either consistently or in the right way?

SPEAKER_00 31:27

It's one of the risks, yes, probably, because since you don't have the resources to do everything, you have to choose, and you can you can be mistaken, you can choose the right the wrong one. You can use your resources to develop something which isn't at the end not useful to anyone, and well, we all made mistakes.

SPEAKER_01 31:42

Yeah.

SPEAKER_00 31:43

This is one of the risks. But the other risk is misuse of AI, which I think we are all very, very, very aware of. Yeah. And also, well, misuse is one thing, is the other thing is for the first time we have to deal with computer-generated solution that doesn't guarantee that it's accurate. Until generative AI, every time you put a machine somewhere, used to be in 99% of the case, it used to be a deterministic answer. So to the same question, you will always get the same answer. So you know that you know one plus one is always two. With AI, it's not like that. It's non-deterministic. So it can hallucinate.

SPEAKER_03 32:22

Yeah, true.

SPEAKER_00 32:23

It can avoid some data because it thinks in between quotes. So you in most in many processes, you need to have a human in the loop. And this is a bottleneck.

SPEAKER_02 32:33

Right.

SPEAKER_00 32:33

So one of the risks is okay, let's do that with a with AI. And you have two solutions. You can say, okay, I try to maximize my output, putting AI to deal with all those questions from our customers, for example. But since you need to have a human in the loop, at the end, you're not doing anything new, and you create a new bottleneck.

SPEAKER_03 32:52

Yeah, yeah, yeah. It's true. It's a new it's a new thing to check, a new problem to exactly.

SPEAKER_00 32:56

And the other in the other path is okay, we don't put the human in the loop and we take the risk that to give wrong answers. And then, well, it happens many, many. I mean, remember if you remember, no, there's not so many, but overbooking with airplanes. We all knew that it exists, and we just all cross our finger that you know it won't be us, it will be someone else. But, you know, they do it.

SPEAKER_03 33:18

Yeah.

SPEAKER_00 33:19

So you could you could assume that in some businesses. I don't think that in education, health, or banking or finance, I don't think those industries or maybe car manufacturing, you are not really able to sacrifice the accuracy and precision of the answer to your clients. It's too risky.

SPEAKER_03 33:37

Yeah, interesting. How do you, I guess with that level, I assume all technologies carry some level of risk. How do you assess when it's, you know, when it's worth the risk versus when it's not? Like, how do you manage that? I guess I'm particularly thinking about when there's innovation ready to scale that increases the risk. Like what's your kind of going to be?

SPEAKER_00 33:58

It depends on the gravity of giving a wrong answer, the importance of giving a wrong wrong answer. You could assume that if the client is requesting uh a quote from a transaction, you can't you can't allow the bank to give the wrong number, even if it's too cheap or too expensive. If too expensive, you lose the client. If it's too cheap, you're losing money. Not it's not a banking stuff. Whatever product that you have to price for the client, you always want to give the correct price, the most accurate and and most competitive price. But if it's just about IDs, for example, what can what could I do to do this and that? And you give IDs, maybe you could afford to make some mistakes because they are not that concrete. And you're not risking uh neither the balance sheet of the bank, you're not risking the financial position of the client either. Uh, you're not taking so much risk giving this kind of advice. So I would say this is the trade-off. When you want to give some main guidelines, or if you want to advise some of the product, uh the example I said before, that you know the client will is it was an imaginary example. The client was just telling you his story, and then you come up with a solution with AI. Maybe this solution is not perfect. But it works more or less.

SPEAKER_02 35:16

Yeah.

SPEAKER_00 35:17

So what is the inconvenient or the uh disadvantage to propose that to a client that maybe it's not, you would say, well, that's not really a good solution. It just won't do the deal with you, but no risk. Maybe you love to deal. Okay. But you know, you're not putting your clients in a very awkward situation and you're not losing money on your side, or you're not even doing something which is not correct regulatorily speaking.

SPEAKER_03 35:41

Yeah.

SPEAKER_00 35:41

This is this is basically where you can try draw the line. Everything else, every time you have to give a number, an actual number, you have to check it.

unknown 35:49

Okay.

SPEAKER_00 35:49

And I would say because of regulation, in many of those cases, you are obliged to check it. You can't do it by it's forbidden by the regulators. For example, we can't do credit scoring, I think, if I'm not mistaken, the from the Gen AI Act from the EU. You can't do credit scoring of people with AI.

SPEAKER_03 36:07

Right. Okay.

SPEAKER_00 36:08

So you can't say, okay, you give me your financial data, your personal situation, I put that in the machine, and the machine says, Well, okay, Louise is whatever note, whatever score it is. You don't have the you can't do that. So maybe you can use AI to get a first score, but you need to have a human check that and you know, put the checkbox and put his name on it. Yeah, we did before.

SPEAKER_03 36:29

Yeah. I mean, do you think so we're still in a world where we need kind of humans to check the AI first? Do you think that is gonna change in in five, ten years?

SPEAKER_00 36:39

I think probably at some point it would be so accurate that you know the percentage of error will be so small that. Well the AI is checking the human. I don't know. Yeah, that's the it is a discussion that I've I've had a lot of things, a lot of time with uh management in the engineering team or security team, is that we are all scared of doing of putting AI in the middle and let AI do the stuff because we all say it it usually makes could make some mistakes and hallucinate, but people also do mistakes. And we have a lot of a lot of time, you know, we do the mistakes. All the companies, you you know, you fail and something goes wrong, and you have to correct it. Time for if if you are sure statistically speaking that AI makes less mistakes than humans, then why shouldn't you use it? Why shouldn't you use it?

SPEAKER_03 37:26

Yeah, that's true.

SPEAKER_00 37:28

The problem is that in in some aspects the human sometimes are aware that they've you know made a mistake, right? I mean, if it's if it's automatic, nobody pay attention to it. So you have to do some background checks and so there's there's ways to make make sure that you know you're not forgetting anything, but I think yeah, it's gonna be sooner than later anyway, for many, many things.

SPEAKER_03 37:50

Yeah, okay. What are you what are you working on at the moment?

SPEAKER_00 37:52

Are you able to give a sneak peek into well that's tricky without entering into uh privileged information? What we uh what we've seen, I would say, is some aspect of the investment banking products. There are there are some products that are still anchored in in the 20th century. There's not even 21st century.

SPEAKER_03 38:13

Right, okay, wow.

SPEAKER_00 38:15

So it is innovation, but it's it's very far from from AI or whatever. It's just trying to get rid of all the actual paperwork that we have to do, which is signing contracts physically. It is it is really cumbersome and really takes a lot of time and it doesn't add any value. Just try to comply with local regular regulations and laws, and also the way people work around that because all business, all those businesses are very standardized established this way. So you have to transform that, and it's it's still innovation because you have to go forward, you have to try to improve that, but it is not high tech. It's just trying to go from very far away to more or less right now. So it's one of the big uh it's not really, it's not really tech oriented, it's more transformation of of the journey of the product for from all the institutions, from the legal parts, from the consultants, regulators, so put everyone to talk together and try to improve that. And probably one of the main uh the the most guilty part is that we don't have any standard. In many businesses, there is no standard. And if there are standards, there are like several standards, which account like you don't have any if you'd have several standards. So difference between the US, the UK, and Europe, not the same standard. So you do something in one place and it's not actually really valid valid in another place. People understand it, but according to their law, it's not that valid as in yours. So you have you have an advantage because if you go to a court, it is according to your court, it's okay. According to their court, it's not okay. So we deal with those kinds of stuff. Uh, this is one thing, which takes a lot of time. Uh the other thing is I would say something that it is not one project, it's a many projects, it's classification of information. I think this is something that many, many companies are working on. It's dealing with the enormous amount of emails we're receiving. And usually those emails are for uh from companies that we have transactions with, and they are telling stuff. I can't pay, or maybe I could pay earlier, or maybe I won't change my calendar, or maybe I don't know, whatever. Changes in their transaction. And literally there are hundreds of emails every day. So you have to prioritize which ones are really urgent, which one are not, which ones are expected, and which one are not. So everything is done by hand, like in any company.

SPEAKER_02 40:40

Right.

SPEAKER_00 40:41

People just receive those emails and try to classify them and prioritize them, and it takes a lot of time and sometimes they miss something very urgent. So we are trying to use AI to do that. Right. Okay. Okay. People be more efficient, not lose track of what is important every day, focus on the most urgent relevant things, pay attention to maybe more sensitive clients.

SPEAKER_03 41:03

Yeah.

SPEAKER_00 41:04

Even if not, if it's not urgent, maybe if you've done something wrong in the past with its customers, maybe you should pay more attention. Even if it's not urgent, you know, to give back a little bit of what you messed up in the last time. So trying to categorize all those emails and all those requests, it's is is very important for us. So this is one of the big parts of what we're doing. Okay. And uh, as the bank, I would say the most relevant thing that we have now in innovation is what we call the global AI adoption, which I'm stuff, it's it's a unit in the bank which is pushing AI into our DNA DNA everywhere. All the users on everyday work, we all have access to some kind of Chat GPT or Gemini interface. We everyone can use it in the bank, everyone can standardize processes and try to share this knowledge with other people. So it is a big, it's not I'm part of it, but I'm not leading that. This is something higher up in the bank.

SPEAKER_01 42:01

Okay.

SPEAKER_00 42:01

And it is it has a lot of international recognition, which is this is why I'm mentioning it. So it's uh I haven't seen that anywhere else as soon as we took it.

SPEAKER_02 42:10

Right, okay.

SPEAKER_00 42:11

And something I'm doing doing personally, for example, that could be maybe some of some interest for other listeners is since we all have access to this kind of tool like ChatGPT or Gymni, all of us use it personally to improve our day-to-day. And since many of us are doing the same job, teams of several people doing the same job for different clients, for example, there are many processes that are the same, but everyone is doing his own prompting. So I'm trying to build a network of, I will call them ambassadors or stuff like that, or something like that. To try to find that in each of the teams we have everywhere, key person, which is usually the one that does most of those kinds of work, to gather the information from his team and then we'll try to meet regularly, share all this information, centralize it, and then give it back everywhere. The point is try to spread this knowledge to everyone. Some people have more time sometimes. Maybe one month I have more time than you, and you have more time than me the next month. So you can dedicate more time to that. So it's a shame that what you've done during this month is is lost. It's just you that is using it because it could be worth a lot of time saving for other people that are doing a very similar job to yours somewhere somewhere else, in another unit, or maybe another country. Same unit in other country.

SPEAKER_03 43:29

Fantastic. And that's something you're doing outside of the BBVA?

SPEAKER_00 43:33

No, it's internally to BBVA. It's internally to BA. So I'm traveling, I'm going to play from breasts to places and check, you know, talk with people, understand what they're doing, how they're doing it, and try to gather all this information to try to see if there is a common core, or maybe try to standardize it a little bit and then get it back and say, Well, you know, remember when you told me you did do that and this and this? Now you can do it here. And everybody can do it in the same way. And if someone comes with a new idea. or improvement, then we improve it for everyone, not just one. So it's kind of standardizing the prompting for some tasks.

SPEAKER_03 44:07

Yeah, I love that kind of the continuous learning and bringing something together. Yeah.

SPEAKER_00 44:13

Much better to have 20 brains than just one.

SPEAKER_03 44:16

Yes, absolutely. Fantastic, Patrick. Coming to the end of time, I just want to ask kind of one wrap-up question. And then we'll close. But if, you know, I think maybe, maybe we land on this point that you just talked about. But if if you if you know there's one key takeaway thing that you'd love listeners to to take away from this conversation, what would that be?

SPEAKER_00 44:34

I would say if because I'm, you know, I'm always reading and and and listening podcasts and watching YouTube videos, etc. Be very aware of the hypers and don't just try to innovate just to innovate.

SPEAKER_03 44:48

Yeah.

SPEAKER_00 44:48

It doesn't make any sense. So innovation is great. Innovation is a tool to improve your business, to maybe fix something that are not working well, to improve the efficiency or the profability, profitability of some business. So it's trying to apply new systems, new processes, new way of working or new tech to something already in place or to create something new with new tech, but because there is a gap or there is something wrong or there is something you can improve. But just trying to substitute something that is working for something which is newer just because it's newer, you have to think really hard about that. Sometimes you have to do it because the technology is got is becoming obsolete. Blah blah blah. Okay, not I'm not talking about that. Don't change a winning team. So I would say be very be very careful with with hype. We're seeing that with AI. Every day there's something there's a new revolution with a new model every day. So it's not a revolution at all. It just lasts for two days.

SPEAKER_03 45:48

Yeah.

SPEAKER_00 45:49

You know sometimes stop trying and just do it. You know that your solution is going to be outdated by two days. In two days going to be outdated. It doesn't matter. If it works and if it brings you some added value then it's worth it. And this is the first thing. And the second thing in terms of innovation is if you do innovation just for uh just based on technology could do that but it's the hard way. The best way is to be user driven and to be business driven. If you try to impose the technology on the users it's going to be very hard. Usually the users they are they're not they are not focused on the tech. They don't care about the tech. The tech is a tool that they use the tech to go from A to B. So if they can go from A to B on an easier or quicker way, they will buy the the technology the change right away. If it takes them more time than before, even if it's cooler because it's a new tech, they won't use it. They would throw it away. So you you would have lost your time and lost money. So if it's not user driven very be very careful. Not saying you can't do it but be very very careful on how you invest your time and your money just you change the technology just because there's no new technology now that is it looks so much better than before.

SPEAKER_02 47:05

Yeah.

SPEAKER_00 47:05

It's like changing your programming language from the same application because you know it's better. I don't know if it works why should you change it?

SPEAKER_03 47:14

Yeah. Yeah no I I love that I love that. I think especially coming from you when you're you know at the forefront of disruptive technologies that reminder to keep the user and the human and the property always don't get carried away with the hype when there's no you know something to root it in.

SPEAKER_00 47:29

Exactly. So yeah hype is okay because it gives you a lot of info of what's new and what's coming but I mean in a company you're there to you know there are two goals make make money and pay your your shareholders but making money means giving people the tool to be more efficient and to sell more. Yeah so the tech is the tech is a tool the tech is not the tech is not the goal. And in many companies and in in many departments of engineering they usually make the mistake that thinking technology is the goal by itself and it is not except for tech companies obviously but in any other company tech is a tool it's like a knife or a wrench or whatever. It has to be the best one but it has to be useful and better than the previous one. That's it. That's the way you innovate.

SPEAKER_03 48:14

Fantastic I mean great note to end on really really simple but really powerful thank you so much Patrick. Thanks for listening to Codex Futures. If today's episode sparked new ideas we'd love to hear from you. Email me at louise.newson at triptych.com or click on the link below. Follow Triptych on LinkedIn and subscribe to our Codex newsletter as we explore more perspectives shaping tomorrow