Now we're going to talk about adoption. And in the financial sector, it's rapidly increasing as well. According to a new report by AI benchmarking platform evident now, the rate of utilization is actually happening twice as fast, though, in the top performers, the top ten banks. And this index has to be in evidence here. Alexandra Muse, a visitor, is with us now. And Alexandra, it does seem to be that if you put a lot of talent and focus and energy in R&D and you're winning and you're going to continue winning, yes, you are. Caroline? Yes, we released the index today. And as you know, it's covering the 50 largest banks in the US and Europe. And we cover all of those aspects, as in looking at the talent stack and the innovation levels that the banks are pulling. We're also very, very carefully tracking what the banks are reporting on in terms of their use cases and their ROI.

And also looking at the response for the AI side. And what we see is that the leaders are pulling ahead and leading more at double the pace of the rest of ten. Top ten ranked banks in the index are moving at double pace from the rest of the banks in the index, which should indicate and does indicate a bifurcation in the rankings. So the leaders are leading more, and the ones that are following or standing still are actually falling behind in this race. Now, Alexandra, I'm pretty certain that social media accounts like overheard on Wall Street and Wall Street memes are listening to this because there are two things here. Banks and AI, and in all world, that is everything. So here's the question. What is it that makes a bank good or not good at AI? Well, we look at the strength of the capability of the banks, all of the aspects that go into a very strong ecosystem that moves the use case from

ideation to production. So a good bank is able to move that through the system. So identify where you have problems, where it can be a solution to those problems, and move them into a testing phase and then into a production of a use case using AI. So a good bank can do that very quickly. And all of the elements that are needed to move the use case into production swiftly and well and with ended are why have those the strengths of those component parts. Okay. So we're showing JPMorgan, Capital One, and RBC. And so based on your report, they are top three performers. Is it that they just hired the best computer scientists, the best engineers only? What did they do differently to the rest of the list? Yeah, the common denominator of the leading banks is that they went out early and they went in big and they reorganized the bank to look like more like a tech company. So an AI-first business that has a

research lab that is attracting the best talent, that is reorganizing the bank from inside, looking at the data infrastructure, looking at this question of build versus buy, looking at the decision-making, whether that lies in the center of the bank or if it's federated, and striking that balance. So what these leading banks did is that they went out early. Jamie Dimon famously laid out the strategy for JPMorgan in 2018 and established a research lab and really went to work on looking at how to reorganize the bank so it could respond to the oncoming technology swiftly and easily and get that time to production down and really infuse the bank with, you know, AI thinking and a mindset and a nimbleness. And that is what they're reaping the benefits of now, not that they're standing still. They're moving at a pace of other banks continuously. And to spend more on I hire more people

and continue the research and the patents and the partnerships that are delivering the impact that they hope to get from their investments. Jamie Dimon said this year that the ROI for him from his investments is that $2 billion. So that's point is making money. Who else is managing to see a return on investment? Because everyone's desperate for that. Everyone's measuring the impact internally only to banks and the 50 to 50 are actually talking explicitly about their ROI. That's JPMorgan and that's DBS out in Singapore. But this is going to ramp up. So we think that in the next 6 to 12, maybe 18 months, really as air use cases move into production more solidly and the pressure to show our ROI of that investment is going to ramp up and the ROI is being measured internally in the banks. And we're seeing many of the banks looking at how is it impacting on non-numerical terms and numerical terms. But we're going to see a cascade of

announcements of ROI. We think probably from about 6 to 12 months on. Well, that's what's so interesting on the timing of the report. We've just had banks' earnings season, so I guess did you see any early evidence that gets you excited for future reports of ROIC? Absolutely. Absolutely. I mean, it was also other factors that were going on in the earnings calls that were around the rates and geopolitics, of course. But there were four banks that were asked about their strategy and spend and use cases and the impact it's having. And so Jamie Dimon was asked, there was also Ted Pick from Morgan Stanley was asked why, and Bank of America were asked specifically about what they were expecting to see in terms of their investments and where they see that having an impact. And Ted Pick actually went into a lot of detail on a tool that they've rolled out across Morgan Stanley, called debrief. That is something that they're trying to

put in the hands of many across the bank and service that helps manage meetings and responses and so on. But we must remember that we're in the early innings. It is a big and monumental transformation, and we're only two years in. So there's a lot to come, and the tools are getting better and better with much more reliability and speed. But we're looking at a in the next year or two, seeing what we call again, TKI come in and being tested out and being a potential tool for the banks to consider and are considering already. And that's going to be a real game changer on many accounts. So agents front and center, what are they using? They're using models. Are they building all on open age largely? And what are the frustrations on the again, I know just more broadly what models have been deployed. Are they building them from within, depending on in large language models deploying from open I and the competitors do that.

Most of the banks are looking at the variety of models out there. So, it could be open. And the topic that you were mentioning before, Mr. Ball hugging face. Every model has that has a uniqueness. And when you're looking at the variety of use cases and problems that you're trying to solve in the bank, there will be different types of models that need to be deployed and fitted. And then the building and the engineering of making it fit on your own data used for your own systems in a particular way that is unique for you. Depending on what type of bank you are, you're looking at. You're looking at these variety of large language models that each have their own features and but then they get bought in and then they get implemented and fitted into the data, into the system of the bank.

Big Banks Tackle AI Adoption

Welcome to the Bloomberg Technology YouTube channel. Subscribe for original series and mini-docs on game-changing technologies, profiles on innovators, live conversations with experts, and breaking technology news & analysis from Bloomberg TV.

Big Banks Tackle AI Adoption

Big Banks Tackle AI Adoption

Scripts