Whitney McDonald 10:15:35
Hello and welcome to the buzz of fin AI news podcast. My name is Whitney McDonald and I’m the editor of finaI news. Fin AI news has rebranded from bank automation news, marking the next step in our mission to lead the conversation on innovation and Financial Services Technology. Joining me today, February 24 2026 is Ravi kokar EVP and head of cloud at Capgemini. During this episode, Ravi discusses cloud based AI agents, multi cloud and multi agent ecosystems, and the importance of governance, trust and training when deploying agenda. K i Thanks for joining us.Ravi Khokhar 10:16:07
Sure Whitney, I’m Ravi Kochar. I’m part of Capgemini financial services. Capgemini is a large system integrators. We are headquartered out of France, out of multiple businesses, financial services, which I belong to, is specifically oriented upon banking, capital markets and insurance. And my role is to run the cloud business for financial services in Capgemini worldwide, so wherever we have interactions with the customers with respect to usage of cloud, usage of cloud based technologies, and now AI is taking the center stage. My team takes point on that for financial services across the globe. In this role, I look at delivery architecture, relationship with hyperscalers, go to market functions, and also, most importantly, usage of AI.

Whitney McDonald 10:16:55
Yeah, that’s what everyone wants to hear about right now, right? Is the use of AI. But obviously there’s a couple stepping stones to get there, so I’m eager to talk with you today and explore your expertise before we get into that. Maybe let’s kind of start with the state of AI and financial services from a bigger picture perspective. How are leaders managing the AI shift? What are you hearing from your financial institution clients? Questions they’re asking? Kind of just give me a big picture outlook.

Speaker 1 10:17:25
Absolutely, it’s the most talked about term. So Financial Services has adopted AI widely and widely is to be to be consumed a little bit more carefully, because at some level, AI technology is implemented by financial services in production. That’s about 10% of our customers that we see have moved beyond pilot and ideation stage, but largely, 90% or so customers are still in that initial phase of exploring the technology, exploring its usage, and investing into doing some sort of dipping the toe exercise with AI. What we have found across financial services is that lot of customers are prioritizing customer facing processes like onboarding services fraud detection as an early use of AI and agentic AI in particular. And there are some customers who are also investing into supervisory roles so that AI can be used little bit more carefully with with auditing, with adoption, with trustability and elements like that. And more importantly, the role of cloud has now become an enabler for AI across financial services. That’s the theme we see across the board of how AI is being used. And lastly, I would say that cloud has not only become the enabler, but also becoming an innovation platform for adoption of AI and financial services, whether it is real time, workloads, analytics, orchestration of processes across the board. Cloud is really creating that, making that happen for most of our financial services clients.

Whitney McDonald 10:19:12
Now on that note about the role that cloud is playing in innovation, in, you know, adoption of AI as a whole, maybe we can talk here about the role of cloud based agents. Obviously, agentic AI seems to be the theme coming into 2026 you kind of can’t have a conversation about AI these days without talking about agentic AI, let’s talk through cloud based agents, benefits, opportunities. Kind of where that sits right now? Absolutely.

Speaker 1 10:19:44
So I’ll touch upon the first part of the element of use your role of cloud. And then I’ll very quickly go on to the actual theme of the question, which is use, use of applicability of agents in the landscape of financial services. So in the world, cloud report for financial services that we published for 2026 the first section of the report talks about how the cloud has evolved from the role it was playing for either infrastructure provisioning mechanism or a software as a service provisioning mechanism or a platform mechanism to really an enabler of AI. And we call that adapt theme. In the world, cloud report for financial services. There’s a big section that we have introduced there, and in that we talk about, you know, across multiple segments, like customer service, card. And payments, fraud detection, loan processing, or even customer onboarding. How Cloud has really shifted the conversation, and the center of that, if you start peeling the onion, is going to be usage of agentic AI. So what we have seen is Gen AI made this whole concept of inference much more practical, but agentic AI really harness the power of orchestration across agents, and what that enabled us is with is unlocking the innovation level where there are multiple agents working in the context of a business process, let’s say onboarding and working in harmony and almost intelligently and autonomous function, where the multiple agents work towards streamlining a business process. In many cases, these agents, which I mean agentic AI agents, are also sort of reimagining the business process ground up, and that is unlocking a lot of stuff for our financial service customers. To be very specific, what we are seeing is high impact processes are really being able to reimagine using agentic AI, then they get paired with the cloud orchestration platform. So let’s say a customer has invested into Microsoft ecosystem. Microsoft Azure, the same open AI ecosystem from Microsoft is now able to change the landscape of how agentic AI works. And then there is this concept of within the same ecosystem, how do you go on? How do you become a little bit more intentional for managing risk, compliance, explainability, and that has also taken a step up with respect to usage of multiple client agent, multiple agent ecosystems, where agents are not only being used for reimagining business process, but also to govern things better, and that is where it has really shifted the conversation from changing or automating one business process to holistically look at business processes being reimagined ground up.

Whitney McDonald 10:22:51
Maybe you could explain or give an example of what some of those business processes, or what a business process is that has been reimagined with an agent.

Speaker 1 10:23:02
Absolutely, there are many. There are many of them in the world cloud to put for financial services. We actually took very specific samples of business processes across insurance, banking and capital markets, to name a few, which are very common across all three is what we call high impact business processes like customer service or customer onboarding. These are use cases and business processes that has been existed for decades. Customer onboarding for KYC function is there for ever. Since we have known banking as an industry, Agent AI and multi agent ecosystems have really helped bring that entire elapsed time down from months to weeks, and that has happened not only at the back of condensed time, but also to create a specific experience for a customer. For example, if you’re a large French international bank and you’re taking 100 to 150 days to onboard a corporate customer, usage of agentic AI and multi agent implementation has helped us bring that from 100 plus days to a matter of 10 days, 20 days, in some cases, one week. And what that does is unlocks a lot of innovation and creates value and value realization early in the life cycle. Similarly, in case of insurance, which is the the art and science of identifying and pricing risk. Underwriting has been one such example where there have been numerous implementations that we have seen in the customer spoke to us in the world, cloud report interviews, where they have significantly changed the way underwriting happens today, with multiple agents working together to profile the risk, to identify the risk differently and then package that so that not only the insurance carrier has a better understanding of the risk which they underwrite in the underwriting function, but the customers also get a far better premium the insurance premium. So not only on one side, the risk gets accounted for accurately. Customers end up paying lower premiums for what they were doing earlier, like for like, and that creates a win, win for both sides.

Whitney McDonald 10:25:25
Yeah, that’s That’s great. Thanks so much for kind of providing some examples and some context around that. Obviously. You can quantify savings. You mentioned the hours to minutes, or, you know, days, months to days, the things that can be streamlined, that’s, that’s that, you know, ROI, or that efficiency that you’re looking for. Now you kind of started talking about, you know, how to select a cloud based agent. You talked a little bit about that, that Azure as an example. When it comes to financial institutions, you know, going with a cloud based agent, is it mostly leaning on those existing cloud partners? Do you look outside the box? What’s kind of the strategy there?

Speaker 1 10:26:06
What we have seen Whitney in the world cloud report write up is that most customers are prioritizing business value creation as against going with the traditional way of I’ve only invested into Azure, and I’ll stick all my decision onto Azure. However, there are some themes that clearly emerge in this process. The simplicity of technology stack usually is a factor in selecting the AI technology, because rest of the organization, if, let’s say, invested into Microsoft or AWS or Google or similar technologies, then some of those choices are also based upon that investment, because you are. It’s all about ecosystem, and not just one small function on the side. And the ecosystem has to transform, which means that you have to layer in the ease of change across the ecosystem, the adaption of change from a skills perspective, across across the enterprise. And that’s where your prior investments, existing skill sets, really make a difference. Having said that, there are some customers who are not getting constrained by them, and that’s why I keep going back to fit for purpose business value creation, and that’s where we have seen that while some investments have happened into one specific area, like AWS or Microsoft for let’s say, real time decisioning, some specific decisions were done by our customers, where they chose a Different tech stack because it made sense from a large language model selection perspective, or from the ability to reuse some investment that have already happened. So those are some themes that we have seen emerge as priority or preference. Is usually about capitalizing on the investments you already made. But there have been some examples where clients have not gone constrained with that investment, and also taken a leap of faith and invested into some unexplored territories, yeah,

Whitney McDonald 10:28:03
depending on what they’re trying to accomplish with said agent. Absolutely, that makes sense. Now, something that we’ve definitely covered in the past is, you know, the multi cloud approach. We’ve all heard of the multi cloud approach. So it sounds like that’s going to kind of bleed into maybe a multi agent approach. It all goes into that ecosystem. Is that true?

Speaker 1 10:28:24
Absolutely true. Multi cloud thing is real. And we see that in the prior editions of World Cloud report, we have sort of reported on within segments how multiple clients have multiple clients have adopted more than one public cloud provider, and we see a very similar pattern for multi agent adoption. So what we are seeing is customers are experiencing experimenting with different agent types, and it has clearly evolving into building agents that and doesn’t matter whether it’s in house, built or built with a partner, but the approach, or the sentiment with which this is being approached, is there’s a multi cloud strategy and there is a multi agent strategy at the same time, sometime There is an overlap, and sometime there is no overlap, but it’s all grounded with what business function you are trying to reimagine in the process. But multi agent ecosystem is a reality. Good news is that customers have openly talked about how the industry is evolving with multi agent ecosystem. There are protocols available now with MCP and a two way that really make it happen. So we have seen the same business function getting implemented. We were talking about KYC and onboarding with a different set of multi agent parameters on one side in insurance, and a similar function getting implemented with different set of multi agents on banking. So, so multi agent is a thing. It is real. Most of our customers are, are actively experimenting with it. Some of them also had the luck of putting into production for real life use cases.

Whitney McDonald 10:30:05
Yeah, I think that that’s super helpful, and it makes sense. And it kind of all goes back to what you were saying about first identifying what are you really trying to solve for? You know, not selecting the provider first, not selecting the agent first, but, you know, working from the problem first and then identifying what will help you solve that. Now we talked about the benefits, we talked about the multi cloud approach. Maybe we can talk here about risk. Involved with a cloud based agent, what do you need to have in place to ensure that that you’re applying the right agent at the right time, or that you’ve done your due diligence?

Speaker 1 10:30:43
So there is this whole section in the report towards the end which details out what are the constraints and how our customers addressing those constraints in short term, what you call as risks. There are many flavors of it, and I’ll touch upon the top four or five that we have observed while writing this report. One is governance and trust. 90 plus customers who we surveyed spoke about this being top of the mind, and it gets amplified for specific geographies, like like Europe, and within Europe specific geographies where sovereignty is taken very, very seriously. So governing that and trusting the decisions that the multi agent take as part of the business function is front and center. So that’s one. Financial Services is one of the most heavily regulated industry, so security and compliance is the second, which is top of the list, where, how do you expose multi agent ecosystems to have the right amount of auditability, compliance and forensics for production issues to get investigated upon. And third, in multiple different ways, is what we call operational risk, which is, at what point you will have multi agent ecosystems, have autonomous decision making, versus at what point you’ll have always a human in the loop, what we call deterministic and non deterministic decision making mechanisms, and that has taken the front and center seat for a strategy example of where you will have non deterministic implementations and where will it where will you mandatory have deterministic implementations. Then there are some softer sides of risk, which is skill gap. Institutions has struggled to uplift and upskill their teams to be able to implement it. A good technology with poor skill set is very dangerous, and there has been a realization of that. Therefore, there are programs that our customers are running which is about actively upskilling their leadership, their engineering teams to be able to use, maintain and operate agent ecosystems with precision. There is a section in the report towards the end where we talk about specific examples of banking and financial services, where and how customers are addressing these risks with taking measured steps on their side, I think 2026, is where agentic AI really moves from experiment to widespread implementation in production. And the reason I talked a lot about cloud is because it becomes a vehicle to deliver that mechanism, however it’s, it’s, it’s a fact that more and more use cases and business functions will get transformed. We use the word reimagine and transform multiple times, because we see that happening across the board, where something that has happened traditionally in a certain way for decades, is not completely being reimagined and reshaped, creating a very different value for not only the provider, but also the consumer. And that is what we’ll see happening. There will always be this backdrop of whether this is trustworthy, is it? Is it unbiased? Whether it is, it can be governed well, and those are the standards that will evolve as they’ve evolved in the past with every single technology that has come up this curve. So that is what we see happening in 2026

Whitney McDonald 10:34:27
you’ve been listening to the buzz of an AI news podcast. Please follow us on x and LinkedIn, and as a reminder, you can rate this podcast on your platform of choice. Please be sure to visit us at finaI news.com for more. Ai News, thanks for listening. You.