Capgemini’s GenPAL: Payments Data Monetization in Action with MongoDB

0:09Welcome to another podcast brought to you by Mongo DB. My name is Luis Padminio Diaz, and I'm the industry Principal for Financial Services at Mongo DB and I'm thrilled to be your host Today. We have a special guest with us, Sourav Kandelwal, he's engagement partner for Financial Services at Gabgemini where he's he's also the leading the go to market for Mongo DB in the UK. So Ralph, it's a true pleasure to have you with us. Could you please introduce yourself? Thank you, Louis, and thank you for having me here. Hi everyone.

0:45I am Saurabh. I'm a part of the Gabgemini Financial Services team, I'm a senior manager with the Gabgemini UK team and I'm also the client partner for the banking and financial services related to business. Yeah, that's about me. Fantastic. Thank you so much. Well, last year during our flagship event, the Mongo DB dot local in London. So now we introduce a green edge solution brought by Capgemini based on Gen. AI for payments analytics and it's called Genpath. It aims to find new revenue streams for financial institutions in the particular segment of SMBs or small and medium businesses.

1:26So without bothered you, Soraf, let's get it started. So first, to set the stage to our listeners, Soraf, what are the latest trends you're seeing in the payments landscape today? And what made these the right moment for a solution like Gent Pal to emerge? Right. So I think payment has particularly evolved a lot over the period of time, right. As we all know, you know, there's a lot of focus on going cashless across the geographies. And what it basically means is that you know there's going to be many folds increase in the data as well, right?

2:05Because when you do a transaction in payments online or in store using your card for example, then you generate a lot of data in in the background. And of late, what we're seeing is that with the onset of AI and particularly agentic AI, there is a lot of focus on how we can derive more value from this data. Now this typically used to the case when we used to have standard business intelligence analytics kind of operations within the payments ecosystem up until a few years ago. But of late what we are seeing is of late what we are seeing is a tactical shift into embracing agentic AI And a lot of a lot of our clients have already started moving to that direction. And that was one of the motivations for us to kind of come with the idea of Gen.

2:55Pal in collaboration with, with with Mongo DB where we could, you know, go to our client and and and help them to derive maximum value from the data. Absolutely. Thank you so much. And, and actually, I was quite intrigued about the solution because indeed, merchant servicing is one of those candidate areas to be disrupted. Traditionally, you know, SM BS has been recognized as the most underserved market. Nonetheless, it's one of the most valuable ones. According to some figures, 60% of all global payments interactions go through SM ES.

3:32Hence they represent a valuable source of data for things like customer segmentations, the behavioral analytics like the chunk prediction or Jio insights, fraud signals, just just for naming a few. And I agree with you with the with finding new avenues of data is the new challenge for banks because they're constantly feeling the pressure for new fintechs. Well pay tax offering these modern cost effective services. And on the other hand, we have the customer expectations, right.

4:05The small and medium enterprises can be substantially enhanced by by using these value added services. So this is the world we're living today, right? And but we have a really interesting third player coming into the picture as you mentioned, which is AI, you know, it's considered at this transformative capital of enhancement these services. And today we're listening use cases like delegating the full shopping experience to agents of the so-called legitic e-commerce.

4:34Maybe it's something that we can discuss right after. So let's let's get into the business capabilities. So when we think about these modern payments, real time transactions, AI driven insights or evolving regulatory demands, where do organisations typically feel the strain in their data architecture? So I think, I think from our perspective, what we are seeing right now is, and that's, and this is something that we have also, you know, published in our latest payments report, right? If you or any of the viewers want to access the payments report published by Capture Me, then that's already available online.

5:18And then I'll really recommend all of you to kind of go through that. But you know, there's a lot of focus, as you said, on merchant acquiring and, and what can the merchants do additionally with their data in addition to that. So that's basically on the card transactions side, right, Where this is a separate, separate division in itself within the payments ecosystem. So there you talk about, in addition to the value that the data brings in, you talk about, you know, are the payments safe? What can we do to mitigate or deal with fraudulent transactions?

5:49How can we reduce false positives? And the same concept basically then translates to the commercial payments and when we're dealing with cross-border payments, for example, particularly with the onset of ISO 2002 two which happened which went live in, in November last year. I think there's a lot of focus now on again, you know, the regulatory aspects. First of all, you know, so again, as I said, reducing the false positives, fraudulent transactions reduction, that is one way where where our clients are seeing a lot of use cases coming in when it comes to agentic.

6:29In addition to that, what we are also seeing is there there's a lot of focus on improving the efficiency of the payment. So how can they process the payments faster? And if a payment fails for, for, for, for some reason, like, for example, you know, I being one bank have, you know, on boarded my systems to ISO 22 two, but potentially the, the bank with which I'm transacting is still in the process and there could be a schema mismatch. And when that happens, typically the payments would fail and therefore you would need a manual intervention to pick up that payment and then process it again. And there's, there's a lot of, you know, this, this additional effort that's needed.

7:07And our idea is how can we use agentic AI to kind of, you know, help the banks and the payment organisations to improvise in these areas and, you know, reduce the manual integration that's needed. So yeah. Absolutely, yeah. In my experience, what I'm seeing in the field is that, you know, a payment order by itself is not enough. You need to it, it lacks context and that's the important. So for example, bringing an operational data layer that can aggregate the complemented data from different sources and to serve as a single data point for serving these different steps that you have just described compliance check, fraud monitoring, AML screen screening. And now that you mentioned the AI and the agentic part, it's also interesting thinking about what are the kind of ingredients or or the requirements that we need in order to put these agents into action, right?

8:11We we need to think about on a semantic layer to read the right contextual information, a memory to keep track of of previous interactions and for the refinement. And of course the the callable API driven tools like query, like search, all those kind of services that data platforms like Mongo DB is able to provide. So how are how are you leveraging all these technical capabilities in the context of payments? That's I think that's a really good question.

8:43And and that is where you know, our role as an SI partner comes in and and and our relevance kind of sits in as well. And if you actually just take this whole conversation a little bit above payments beat any domain within the financial services, One of the things that you know, that has to be thought through is, and that's even before you kind of talk about, you know, embracing the the AI part or the AI journey. Is that how's my data looking at the moment, right.

9:19And the second thing is now that we are seeing that as a lot of migration happening from the legacy systems to say cloud could be any of the hyper scalers. I think the other thing which you also are trying to focus on is that, and when I say we, I mean the clients, our clients are also trying to focus on is, is that what data should sit on premise and what data should move to cloud. If you actually try to put everything on the table, the key theme that emerges is how do I govern the data?

9:51Because now things are evolving from just a proof of concept agentic AI solution, which used to be the case, say about a year ago when all of these LLMS were just hitting the market one after the other and everybody was trying to, you know, trying to experiment, put up some use cases. I think the shift is now very, very real. And we can see that the the clients have a plan and a path towards production. And with when they think about this journey towards production, the typical aspects that they discuss are things around how do I go on the data?

10:28As I said, my data is not in the right form and shape as you would expect this to be. If I were to design this kind of an AI architecture, how can I still go ahead and, and make that first step? You know, so these are the, these are the ingredients or these are the, you know, sort of the modules or the areas, whatever you want to call it, where I think the focus is going to be there, right? So that is the evolution and that's a natural step that we are seeing, right? And, and that's across, not just in the payments, even in other areas, for example, in, if you talk about a risk function within a bank or a finance function within a bank or a treasury function within a bank, it's typically the same kind of the, the, the topic of conversation, which is if I were to embrace and move towards an identic path, which, which obviously will happen.

11:25The thing that I want to kind of focus on is how's my data structured at the moment? And how do I govern the data? How do I make sure that when I go to the cloud, my solution on my platform is scalable, I'm production ready, typical things like observability, for example, is very, very important. So I think there's a lot of focus around these areas and not just the core, you know, architecture around the AI ecosystem. OK. It's true that AI can take care of the heavy burden of the clearing, the settlement, all those manual processes and it can automate things like reconciliation with legacy systems, with payment gateways, etcetera.

12:07And actually nowadays we're, we're hearing that some entities are working on pilots of delegating A gentic AI for exception handling. But one of the key aspects in in payments or as a must have, if you will, is immediacy. You know, as soon as you swipe the car or you type your smartphone, there are a lot of checks, a lot of things that need to happen in a quite short time frame in, in a matter of milliseconds. You need to do fraud checks, screening, etcetera. How do you make sure that in that short period of time a genetic AI can be executed accordingly? Yeah, I think, I think so that's where the capabilities of a platform like Mongo DB coming, you know, which is one of the leading players when it comes to operational data processing. And what we are seeing is that when it comes to submillisecond latency processing and those kind of operations, Mongo DB becomes a natural choice.

13:09And of late, the new features that Mongo DB comes in with, I think they are again a game changer. For example, the MCP server that has been introduced very decently by Mongo DB, right? It allows us to kind of change the game completely. So what we are saying is that design wise we are looking at multiple agents sitting in the ecosystem. One agent would be doing a particular task. For example, there would be an agent that would be focusing on fraud detection, right?

13:41And within fraud detection also you can then have multiple different swim lanes and then you can decide like how do my operation, how would my operation look like? The second agent would then focus on processing and drop picking up a new drop transaction, for example, right. So what we are seeing is this approach of creating multiple agents in the architecture powered by the capabilities that Mongo DB brings in on the table. Absolutely. And and since you mentioned that the cloud hybrid deployment, it's, it's an alternative from the technological perspective, scalability can can be overcome. So let's let's switch gears for a little here sort of now let's talk about the potential market opportunity. So payments data has been traditionally be seen as just operational.

14:25So how are organizations now rethinking it's or could represent the strategic asset? And how does, for example, Genpile help to move the raw transaction data into meaningful business outcomes? So there has been a significant evolution as we as we discussed this previously, but I'll probably just double click on this whole aspect once again. Right about say 2025 years ago there used to be payment systems where you swipe the card, generate the transaction the data sits in, in our database and you know an OLTP system and then you would do your OLTP related operations on that data.

15:08If you wanted to do any analytical operations on, on this data, you would have to create a separate OLAP kind of a mechanism where you would then you know do some kind of analytical processing. From there, we evolved into a more mature model where we started doing dashboards, right? And we said we'll take the data from the operational warehouse. We will probably just create, as you said, a semantic layer or perhaps something like data marts. These will be functional data marts that would be sitting across the payments ecosystem. And we'll derive the data from, you know, multiple different functions.

15:51They'll combine the data, create a functional Mart which can be queried with simple sequels and and then on the on the extreme right hand side would sit your, your final consumption layer, which would happen to be a dashboard, typical, you know, dashboarding tools. And you could see those, those nuances there on the dashboards, right? Those, those questions that you would want to ask from your from your payments data. From there things evolved. And then there was, there was this, this whole idea of can I make these dashboards intuitive? Can I play with these dashboards? Can I, can I, can I, you know, use some inputs if I want to and interact with the dashboards.

16:35And that's where the whole, you know, the whole shift happened. And we thought we started talking about self-service BI, right? And, and now what we are seeing is that the the game has gone one level up where you don't even need to be a dashboard savvy or technical to write a sequel. You can just simply ask a question in natural language, throw up on the console, right? And empower the business with the same high quality data or even better, that can help the businesses to get more value.

17:10And what this basically means is that I as a merchant now no longer need to know these technologies or worry about my technical infrastructure. I can simply, you know, use this service and, and ask a question in natural language and, and I can focus more on what I'm supposed to do, which is excelling at my business. So you see there is a, there's a, there's a complete shift which we have seen and an evolution within the payments landscape and, and things are just going to get better from here.

17:42Yeah, you summarize it perfectly. Yes, indeed that's something that we are seeing in the market as well, these rays of translitical platforms that can cover the transactional aspects, but at the same time bringing real time analytics without the need of advanced or or you know more complex data intelligence platforms. Some some kind of questions that you can solve in the moment. And you're absolutely right that banks are investing in these new market segments when you know, there's some solid business model behind and payments has always been attractive in from the standpoint of volumes. The other day, just I read an article raising that the, the true challenge here is not a technology gap, but rather than an experience gap.

18:31And since you mentioned the natural language processing and, and, and the capabilities of, of agents for interacting in a much better way, I think it's, it's critical for, for small and medium enterprises. Actually in this in this article, it revealed that 40% of merchants are moving away from banks to work with with paytech companies. But yeah, the problem also they mentioned is that however, that 80% of those mergers will be willing to switch back to the banks only if they provide data-driven and value added services like the one you described.

19:12That's absolutely right, right. And, and they, they, I read it somewhere online, like, you know, we are living in a time where fintechs are, you know, becoming banks and banks are on the path of path to becoming a fintech and they're both trying to catch up with each other. And I think that this is where the customer gets the value from the competition, right? It's all about who is going to provide the maximum value to me as a, as a customer, right? And I myself, for example, when I use my bank card or make a payment in store or online, I always kind of look through these things like, OK, what, what kind of offers is my bank providing me?

20:01Or what kind of offers is my card issuer providing me any offer that I'm getting at the checkout, for example, right? These are small things, but they change a lot of things around the customer experience. Right. And, and if you really look behind the, behind this experience, it's, it's all about reading everything about the customer, the demographics, which is basically the data. And, and then use a solution that you can then use to get the maximum value from the data, improvise your business KPIs as well as pass on the same value back to the customer.

20:41And it's a win, win situation for everyone. And therefore this competition is, is only going to get better from here. And as I said, you know, the, it's, it's the customer who's actually going to see a lot of benefits in terms of, you know, this bank is going to provide me the maximum value. And therefore, as you said, you know, all of those customers who are making this move to switch to a fintech would probably just think about moving back to the bank if given a choice, because these banks have been traditionally doing, you know, great job in terms of providing services to the customers, right?

21:16All of these High Street banks, for example, everybody's doing a great job. I mean, they're they're all there for a lot of years. But what is in what is interesting is that now when they go to the customer, the proposition is completely different in this data-driven world. Yeah, yeah. And, and also let's keep in mind that banks has the biggest asset from banks is the credibility and trustworthiness that we have from the regular standpoint. And well, I, I, I cannot agree with you more on the value that needs to be delivered for, for the end user at the end, right? And AI, it's clearly a transformational enabler, but from the technical perspective, it seems to be like a daunting project.

21:59So from your experience, what does actually take to operationalize AI at scale with such high volumes, high velocity environments from the bank's standpoint? I think that's a, that's a, that's a great question because, and it's very relevant in the, in the current scheme of things where, where the whole journey towards agentic is moving. And, and very interestingly, this, this just happened like probably a couple of days ago. And I was talking to one of our clients and we were talking about, you know, you know, just doing a data transformation and basically, you know, the agentic transformation for the area that this client works in.

22:43And, and one of the questions that he asked me is that how do you ensure that we do not fail, right? And I just took a pause. I just took a pause for a few seconds because I really thought that it was a very good question that he asked because when it comes to talking about AILLMS rag, all of us know that I mean, these are things that have that are there for a while now. So I think it would be, it would be, you know, just prudent enough to say that many of us at least who belong to the technical fraternity, they understand these things and so does the customer now.

23:28Now when, when they ask this kind of a question, it basically kind of means that while they are ready to make that move, there are a lot of questions that are there in their minds, right, Whether my strategy is going to work or not. I have been doing this proof of concepts for a while. But now that we are talking about taking this whole thing to production, how ready are we to do this, right? And, and we as partners, how can we support them and, and how do we mitigate that risk of failure, right?

24:00Because these are strategic projects. And as you know, right. And when you, and when you start a strategic project, it sounds very interesting. And then it becomes like a, you know, daunting task if you just kind of stray from the path. And therefore it's important to, to address this in a way where we put enough safety, you know, safety checks in place. Now it's easy as that is done, I understand that. But one of the things that what we are proposing is that we will not take a Big Bang approach.

24:36You know, when it comes to doing these AI transformations. We will identify what are the top burning problems that the business has. You list on those problems sequentially, you know, 1234 and when we know these that these are the problems that business has, we pick up the top one or two challenges, just one or two challenges to start with, right? Because we really want to test the waters. What it basically does is that when you identify these, these two first, you know, the top 2 challenges, it's easy for the customers to kind of take the buy in from their stakeholders.

25:13Because obviously they will also need to, you know, go to their higher ups and their stakeholders and their management and say, you know, we are starting this whole AI transformation program and you know, we want to, you know, achieve this. This is our goal and this is where we want to be, say, in the next two years. Now in order to get the buy in, it is easy for them if they take a small piece of work and identify the top one or two use cases, because those are the use cases that everybody's kind of interested in. If, if you solve this problem, all of the people in their team or the department are, are going to be happy about it. So it's easy for them to take the buy in. It's easy for us as well because we know that we are dealing with, you know, a smaller yet powerful subset of the, the bigger problem or the bigger umbrella of the challenges that the customer has it.

26:02It also gives us a chance to put the right things in place, whether it's about putting the right skilled people, whether it's about putting the right, you know, architecture, whether it's about pairing up with the right software partners, They're all connected right. And, and therefore we do not want to make it more like a monolithic architecture. If I may, where, where you know, it's very, very easy to start Big Bang, but then, you know, six months down the line, you find yourself in trouble and then it's very difficult to explain and, and to kind of make a corrective action.

26:43So my, my idea is we pick up one or two use cases which will give us the, the best bang for the money. And then we start working on those use cases together with the customer. There's a lot of, there's a lot of focus on the quality of these resources as well, because obviously when you're talking about these new age technologies, AI agentic, there's a skill shortage as well in the market, right? And, and that's where we are also telling to our teams to kind of upskill themselves.

27:22And then we are also hiring from, from outside. We are also looking at at at the quality of of the people, because it's very important when you're looking at these high impact critical programs, which look very small, big, you know, to start when you're starting, but then they kind of proliferate and become like a, you know, a major transformation program. So we want to make sure that we don't mess up when we start in the beginning. And that's where we talk about the resources, we talk about the right technology partners, we talk about the right architecture. We also talk about if this doesn't work, how easy for us to kind of go and make the corrective action right. So I'm just kind of going at a very, very high level, but then these are the things that are very, very important for us to kind of, you know, well, I can probably maybe if I can say that ensure that you would not fail.

28:13Maybe, you know, obviously nobody can guarantee that a project would be successful, but we try to do the best we can to kind of put these, put these safety Nets wherever we can. Yeah, totally. Yes, absolutely. And I and I think actually picking the right use case that could the, the risk as you mentioned, or minimize or mitigate the risk involved. And while you continuously iterate the solution, getting feedback from business and stakeholders and making sure it provided value.

28:44We were discussing just previously is what it, it, it could make a successful project in the end. But thinking about the solution standpoint in in Gen. pal at this time, how do you feel that it could become a potential accelerator for a project like this? And the reason I'm asking is because normally my conversations with key business stakeholders, we tend to think about, for example, MongoDB, not just a database, but rather a time to market accelerator. And I believe Genpal it's also a key, could be potentially a key asset for accelerate this time to market from solution like this.

29:26So how, how would you leverage this premium solutions aligning to business stakeholders and making sure the value is delivered as soon as possible? Yeah, I think if you really look at it and if you talk a little bit about Genpal, right, what is Genpal really again, at a very, very high level, it's it's, it's, it's a way we bring in the best of the both worlds. And what I mean is that we as, as, as I partners, you know, we excel in domain consulting, domain expertise and on the other side, and of course, you know, the experience of delivering large scale data migration or, or technology migration programs, right? So that's, that's where we come into the play. And what Mongo DB brings in is this robust, state-of-the-art, highly efficient architecture where it's not just about storing the data, it's about what we can do more with the data, right?

30:29And when you amalgamate the 2, the best of the 2 worlds together, what we come up with is an idea like Genpal. And what started as an idea, you know, probably about a year or a little more than that ago has now, you know, put us on the launchpad to think bigger. We started with card transactions, for example, right? And then when we started taking this idea to our clients, to our partners, what, what made more sense to us is that we can solve a lot of problems with this. Yeah, within the payments ecosystem. So it's not just about solving problems just for the, the merchant acquiring side of the things. But we can obviously target a lot of use cases, say for example, in cross-border payments, right. Eventually it's all about the data and, and obviously when you talk about the data, you cannot ignore the domain part of it, right?

31:31And we have got a lot of domain expertise in the house where we understand these problem statements given we work very closely with the customers and, and, and then when we partner with Mongo DB and, and thanks to the brilliant team that Mongo DB has, and I work very closely with some of the architects myself, you know, from the Mongo DB team, I think it's all about them doing this brainstorming where we, we put the customer in the centre of everything. And that customer could be, you know, a merchant. When I talk about the retail transactions, it could be a, a business user who's sitting in a middle office operations team that is managing cross-border transactions. It could be anyone, right?

32:15And depending upon who the customer is, we devise everything around him, right? What is the kind of problem that he's facing today? And let's take an example here, right? So, for example, if it's if it's cross-border payments and this this user is sitting in the middle of of the system and and he has to deal with a lot of payment failures every day. What could help him is or what could make him like make his life easy is a solution where he has to do bare minimum to fix this, right?

32:48So we try to to vision this and that is what we want to do with Genpal where you know you have a dashboard, you can see everything on the dashboard itself. What went wrong with that particular payment, why it failed. And the system is intelligent enough to let you also know what is it the correct, what is the corrective action that you should be taking to fix this problem, right? You still have human in the loop and the decision still lies with this particular user to then analyse everything what he sees on the screen and then just submit the the transaction for reprocessing.

33:27So in a way what we are doing is we are empowering the business teams. And then if you talk about card transactions, the our merchants, our team sitting in the acquiring businesses, we are empowering all of these teams. And the users were sitting within these teams with a solution where they need to do minimum from their side and they can, they can make use of whatever time they have to do additional bits, you know, and, and, and that is where we see the, and that is how we see the journey of Genpal going forward.

34:03It's, it's, it started with cut transactions, but now we are opening up the gates to commercial payments. And we are also, you know, kind of pivoting as we get feedback from our clients, right. So we when we start talking to our clients about Genpal and we take Genpal to them, we while being very receptive and we have got a very good feedback right about what they have seen so far. There's also some constructive feedback based upon what they see as sort of the additional things that they would want to see on Genpal and that is what we are factoring in now and moving together to bring in the next version of Gen.

34:45Pal. What an amazing journey for Gen. pal. And I'm glad that you mentioned that we need to keep the human in the loop that that is an essential success factor for these agentic AI projects. Because at the end, we all this technology are intended to augment or enhance our capabilities, but we need helpless to keep the the human to make the final call. Absolutely and and I'm glad that you mentioned that our partnership has strengthened over the years.

35:12You know Capgemini bringing the knowledge expertise in Mongo DB as a proven technology for supporting these large scale operations in my experience. Let me bring you a couple examples as well. One of them just recently a story that we have with a well known pay tech company in the UK called Dojo. They have a large portion of the SME market share and their key differentiator is to provided kind of immediate clearing and settlement services. Because one of the important aspects for these small and medium companies that they need to have instant access to cash.

35:53And Dojo decided to run all these real time settlement clearing services on top of Mongo DB, Atlas. The real time capabilities you describe are could be a potential a Kitty Fedexator for these solution providers. And the other example I wanted to bring today is the Nexi group. You know, it's a large payment processor that operates across 25 countries and they account for 65% of the whole European consumption. And they needed to support a considerable number of different payment rails, right? Each one with different schemas.

36:29And they chose Mongo DB because of the flexibility to to be able to support different payment schema but to process them securely and at scale. Absolutely. And I think, you know, I've been fortunate to, to lead the go to go to market for, for Cap Gemini. And our relationship with, with Mongo DB has, has been very strong so far. And I'm, I'm very hopeful that, you know, we, we continue going forward as well and build on this relationship. I know quite a few people from the Mongo DB team in the UK region, in India, in other geographies in Spain as well. So I'm really looking forward, you know, and, and I think there's a, there's a lot of work that we can do together in payments as well as in other financial services areas also with Mongo DB.

37:21Absolutely, absolutely. And and actually brings me to to my last question actually about what's coming, what's the future outlook as we look ahead? How do you see the payments data strategies evolving over the next few years? What would the finish say these organizations to truly succeed in turning data into sustainable competitive advantage? Yeah, I think the future of payments is, is very interesting. At the same time what I would also like to add here is that it's very, very rapidly evolving and and payments is a is a business where the disruption happens very fast, right. If you compare the domain disruption that you would see in other areas, and again, I'm going back to capital markets, you know, any of the functions like risk finance or treasury, there's a significant lead time that you need to invest and to to, you know, to change the technology landscape. But when it comes to payments, things move very fast. And one of the reasons why that happens is because obviously, as you said, right, everything happens in milliseconds here, irrespective of whether it's a car transaction or commercial payment.

38:43When a payment is initiated, things flow very fast in the system. And, and therefore, when things flow very fast, you also need to think that the competition is also catching up and they are also trying to make things faster. And therefore, what I have seen from my experience is that payments is 1 area where the technology disruption is, is quite fast, very, very fast, right? And what used to take 2 seconds, 3 seconds say years ago now takes less than a second.

39:15It's that's kind of shift which we see. And in the future this is going to, this is going to increase. And I also feel that with the onset of ISO 2002, two particularly, which has just been, you know, gone live in November last year, I think there is a lot of focus around how we are going to make use of those additional data attributes that were not available to us before in the previous messaging formats. Banks are still trying to figure that out. When they, when they went live, the problem, or rather the the focus was on being compliant, right? Because everybody had to be compliant, you know, by a given date.

40:09Now I think the focus is going to be that what can we do additionally, now that we are compliant, what is it that additionally I can do with these data attributes, right? And a lot of players still, you know, have done the bare minimum when it comes to embracing ISO 2002 two. And they have just kind of, you know, dealt with some core data fields which are required for their data, you know, for their payment processing. And then the other bits still remain as is and they kind of build a wrapper around it, right. And now they are kind of taking a step back and saying, now that we've done with this compliance, which was sitting on a on the top of her head.

40:45Now I would like to really leverage and build upon what ISO to adhere to, to unfolds in parallel. So this is, you know, the, this is where the regulations and the changes in the payment landscape functionally are kind of moving forward and in parallel. Then you see also see that agenting is moving very fast, right? And things are evolving very fast when it comes to devising an AI strategy as well. And what would happen down the line is that what we are seeing is, and we, this is something that we also gather when we speak to our clients is that this is going to just, you know, just increase from here.

41:24The sole focus would be that I would like to make maximum use from this data that is now available. Otherwise, there's no point, right? It's going to be the same, same thing which we had seen years ago where a lot of data is sitting in my database, but it's doing nothing and everything then just fails. So I think there's going to be a lot of focus on these additional data attributes that ISO would bring in on the car transaction side. As we discussed already, you know, the, the banks are going to catch up.

41:57You know, I think a lot of a lot of these big banks when we, when we speak with what we see is that there, there is a clear clarity now that's coming in and they know exactly that they are competing with very highly efficient fintechs. And therefore their focus is going to be making on, on, on providing more and more value to the, to the end user, which is the which is the customer who's making that swipe either in store or making an online transaction.

42:24So I think All in all, good times, very interesting times. And I think, and a lot of things are still unknown, right? I mean, it's a very fast in rapidly evolving field agentic and AI per SE. I think the whole conversation around the guard rails, you know, that's also going to be very, very interesting because no, no bank is ready to move forward and take the big plunge unless and until you answer these fundamental problems, right? Which is around data governance, which is around ensuring that you don't end up messing around with compliance, which is also around making sure that I know exactly what needs to sit on cloud and what needs to sit on on premise.

43:04And I think that that whole thing is going to also kind of, you know, take a centre stage where, where it's not just about building the AI architecture and, and everything, but it's about the strategy of how I can operationalise and productionalise my my AI assets. So yeah, I think that's that's that's how the feature at least looks to me. Thank you so much for searching inside remarks and and actually I I agree with with all that all everything that you have to just share with us. And I believe also that solutions like Denpal can bring a win to win situations for, for banks, right, Because financial services providers are always looking for new revenue streams and using this kind of as a service business models, offerings like, you know, insights as a service, compliance as a service or real, the real time checks and decisioning as a service could help SMEs to level up when it comes to the modern customer services. We, we were, we were discussing today and, and also more importantly, to delegate this heavy lifting of the technology to the service providers because it can help to, you know, democratize the access of

44:19emerging technologies like the agentic AI. And now the solution providers needs to, to build modern a, a driven solutions. But of course, they need to keep in mind that it requires things like the security, the scalability, the flexibility, the real time that we, we just discussed today that only modern data platforms like Mongo DB are able to provide. So thank you so much, Sourav. It's been a pleasure. I'm looking forward to see you hopefully in the, in the next Mongo DB dot local in London later this year. I'll be keen to learn about the new innovations that you and the team can can bring and can build on top of Mongo DB. And thanks everyone for all listeners. And until the next time.

45:07Thank you for having me.