Decision-Grade Data, AI Governance & the Future of Capital Markets | Vijay Mayadas, CEO of Rimes

Speaker A: Hello and welcome to another episode of fintech Focused TV with me, Toby Babb. Today I am absolutely delighted to be joined by Vijay from Rhymes. Vijay, how are you?

Speaker B: Good. Toby, how you doing?

Speaker A: Very good, thank you.

Speaker B: Excellent.

Speaker A: Listen, lovely to have you in.

Speaker B: Good to be here.

Speaker A: Traveled all the way over from the States. So very good to have you over. Some amazing things happening with Rhymes at the moment. As I said before, I had possibly about a year ago, Theo talking about the AI and capital markets event. Talking Theo Bell from your business talking a little bit about rhymes and evolution of space. I imagine a year on we're going to hear an awful lot of development and how all that's going as well. So we've got a really fascinating conversation to come into. Before we do it, I always ask people to give a brief intro. So uh, if you give a quick introduction to yourself and then the business as well, that'd be great.

Speaker B: Yeah, happy to. Uh, so I'm Vijay Maedas. Uh, I am the CEO of Rhymes. I've been at Rhymes for about seven months now. Prior to that I spent uh, a lot of my career in fintech. I was at a large fintech called Broadridge where I built the capital markets business and uh, really excited to be in the world of Rhymes because so much of what's going on in the industry is about data. It's how do you industrialize data to power not only today's workflows but also the next generation of AI driven workflows. So Rhymes is right in the middle of that. Rhymes think about Rhymes as a two sided network. So we sit in the middle of the 850 plus data partners on one side and about 400 customers on the other side. Those customers comprise some of the world's largest asset managers, several wealth funds, pension funds, asset owners, and we manage a lot of the complexity of a very, uh, volatile data supply chain. A very important but quite volatile data supply chain. Helping these firms uh, understand how to enrich, augment, curate and develop these data sets from these vendors in a way that can power their mission critical workflows. Whether it's rebalancing trades, whether it's running portfolios at scale a trillion dollars in the case of one asset manager, it's really helping them get access to highly pristine cleanse data sets that they can use to run mission critical workflows.

Speaker A: I think that's fascinating, isn't it? Because I remember as you were talking there, I sort of took myself back two years to uh, an event which we held out in the States which talked about is if data is the new oil, how ready is your engine? And I feel like, I feel that that was even probably, it was actually probably before the AI hype really sort of took off, et cetera. But it's become even more of a conversation and we all hit, you know, we've heard on this show many times the whole concept of rubbish in, rubbish out, but data seems to be the, you know, ever increasingly the, the turning point about successful adoption or otherwise in the space. And that's exactly where you come into it, right?

Speaker B: That's exactly right. So, and that the, you have an interesting point around how things are changing around data. So typically when, you know, I spend time with uh, one of our clients, sovereign wealth fund, asset manager and so on, inevitably the conversation focuses AI on AI and there's a tremendous amount of focus on trying to understand what it takes to use these amazing capabilities that are in these frontier models, right, which are uh, you know, accelerating, you know, at a dramatic, a dramatic pace. And inevitably in every conversation ends up talking about the data foundations. You need to be able to unlock the next generation of automation. Obviously in a very regulated industry like you know, capital markets, which is the world we operate in, uh, the need for pristine, curated, well governed data with deep lineage is paramount. And there are a lot of discussions going on in the industry around if we're going to start to enable next generation automation. So increasing the speed, increasing the efficiency of processes across the investment life cycle by 2, 3, 4, 5, 6, 7, 8, 9, 10 times. What are the data foundations you need in place to really trust that agentic workflows are going to do the right thing. Especially in the context of an environment where there's uh, a lot of focus on compliance and a lot of focus on regulation. Um, there are more and more regulations emerging around lineage, around data sets, around ensuring that if something happens in a workflow, you can go back to the audit trail of data and understand where that data broke. And that's only going to become more important as you start to drive more and more agentic type workflows. So the majority of conversations I have with clients around AI become data conversations very quickly. And that's where I see a tremendous opportunity for firms to industrialize data, take it to the next level. Really think of data as infrastructure, you know, versus content.

Speaker A: M. I mean there's a lot to unpack from, from all of that. And I think as you, as you're talking that the word that keeps on coming back in this, this sort of conversation is Trust. Right. And uh, I think back to, you know, the evolution of the adoption of digital assets into the space was again around, you know, how, how far do we have trust, uh, in digital assets following, you know, its links to crypto and everything else around. That's that sort of space, I think in here. I spoke to someone, a sort of fairly leading figure at an investment bank a couple of years ago, uh, who was fairly dismissive as he was about digital assets. Um, you could argue how well he's been proven, rightly or wrongly in this sort of space. He uh, was saying the problem about AI in capital markets is approximately right isn't good enough. And that was the. As I say, that's probably a conversation that's been buzzing in my head from a couple of years now. As we evolve, I think people are recognizing that there is a overwhelming opportunity within AI and there's an overwhelming responsibility, uh, to get the underlying parts of that right, to make that more trusted and more valuable. And as that's happened more and more over the last couple of years, people are able to adopt it and do the things which we want to do in it. I feel that this is, I've said this on the show many times before that I think the real excitement for me is about the increase in productivity in the space as well as the, the last, however many, nearly 20 years of really focusing on efficiency and cost reduction. There's obviously that element to AI, but also the explosion in productivity on a per head basis about what we can do with this is the thing where I get really excited to have that happen in its truest sense. Obviously needs this trust to evolve into it and I feel that's where you're uniquely well positioned for it. Right?

Speaker B: Yeah, I mean, that's right. And Grimes has been in the business of uh, providing trusted data sets to clients for about 30 years now. Right. So we have very deep experience understanding what our clients want, you know, if they have specific things they want to do with data. Customization of index spends, for example, enrichment of data sets. Uh, we have a very deep domain expertise to enable that to happen for our clients.

Speaker A: Right.

Speaker B: And what's happening now, you know, to your point, is that trust has always been non negotiable, but the issue now is that if there's an issue with any of the data and you are investing in next generation AI type workflows, the risk of errors propagating very quickly is much higher. So we sort of come from a world where, you know, machine learning is, is, is, is used quite Prolifically, you know, throughout, uh, throughout buy side firms, but enabling generative AI models with reasoning capabilities to take action with a level of human oversight, but not too much human oversight because otherwise you can't scale. That world is demanding a new paradigm around how data gets managed. The concept we have, which we talk to clients a lot about is this concept of decision grade data. So at what point do you feel that you have enough trust, you have enough explainability, you have enough lineage to, in your data sets that you can start to trust agents to act more and more autonomously across the investment life cycle, Right?

Speaker A: Yeah.

Speaker B: And that's a real white space, right. For discussion with clients and for obviously uh, firms to really lean into and help think through that opportunity. And uh, it's also reframing, it's kind of what I mentioned earlier, reframing what data actually is. Right. This concept of data is evolving from being content to being infrastructure. Right. Infrastructure obviously implies a level of resiliency, a level of robustness.

Speaker A: Right.

Speaker B: Uh, a level of, you know, a level of governance that is at the standard we need in our industry. Right. That I think is the paradigm shift that's happening, you ah, know, in this industry. And you know, what I've seen out there in the market is a lot of really good POCs, um, around uh, particularly, you uh, know, agentic workflows. But they tend to hit a little bit of a barrier in terms of well, what's it actually going to take for us to trust this stuff in production. Right. And a lot of that barrier, a lot of the constraints I think are ah, really driven at the foundational data level. Right. And uh, you know, the analogy, you know, I like to use sometimes is there's obviously a lot of focus on frontier models, on bringing in these capabilities into the house. But it's sort of like, you know, does it really worth keep upgrading the engine if the fuel is contaminated?

Speaker A: Right.

Speaker B: So I sort of think about the data problem, you know, in that context

Speaker A: and when you, when you're talking about that sort of evolution of the market, when you're talking about people's confidence in being able to adopt this, when you talk about the move from pocket to actually implementation, where do you think we are at the moment and how are people winning in that race?

Speaker B: Yeah, yeah, I think there's been a lot of great, great PFCs in the industry. I think people are starting to see the dramatic increases in productivity. Uh, you know, frontier models can provide, albeit in very contained sandboxes. Right. And the obvious one is software Development. Right. And you see, uh, a gentic workflow in whether it's you know, uh, Claude code or anti gravity goals, anti gravity product. You, you see the power of Regentic workflows, uh, in software development. And that maybe is the most sort of visceral experience one can have.

Speaker A: Right.

Speaker B: In terms of understanding the power of this stuff. Right. So I think there's a lot of really good work going on there. We're doing a lot of this stuff obviously within rhymes. I think the next question is, okay, beyond software development, when it comes to the investment life cycle, what sort of areas are best, uh, primed to bring that sort of productivity increase? Right. That we're seeing in for example software development to other aspects of the investment life cycle. And I think there we're seeing uh, uh, a lot of POCs. But it's in that leap where the most amount of challenges arise in terms of these workflows are deeply interconnected. Right. Uh, there's a lot of compliance mechanisms already built around those workflows. Right. And understanding how to take a piece out of that workflow, drive maybe a 10x increase in productivity in a trusted way and then make sure that you're not disrupting your other workflows. I feel that right now is the areas of, call it highest complexity. Right. For, you know, for many firms. Right. And uh, so that's a big question. And then the other, the other issue is, you know, you can sort of look at very specific because every workflow has its own compliance needs in some ways around what uh, you need to get. Right. To run a gentic AI, Right?

Speaker A: Yeah.

Speaker B: Obviously, if it's a, if it's a workflow, you're actually making an economic decision. You're defining the status of a trade which has downstream consequences. There's like a super high bar in terms of compliance. Right. If it's a workflow where maybe you're, you're generating content that ultimately is going to be improved by a, you know, by, by a person. That's a slightly different bar. So I think a lot of firms are going through this process now. We're trying to unpack all of these different workflows, right. Where do they fall in terms of the sort of autonomy threshold, if you will. Right. And then how do I go about sort of, you know, breaking down each workflow, developing a prototype and then understanding what it takes to go into production? So I think we are in fairly early stages of that whole, whole thought process right now. And of course in the meantime, these frontier models, uh, are Moving ahead, you know, even more quickly. And so the gap between potential capabilities and the reality of implementing them sort of on the ground I think is widening. And I think a lot of that again boils down to foundational aspects of data.

Speaker A: Yeah. And people are having to work back on that a lot, aren't they? I mean, uh, I love this catchphrase and I'm going to read it directly so I don't get it wrong because the intelligence fabric for capital markets is where you are positioning yourselves within this. And it feels like that is an incredibly important part to get right. And people are starting to realize that more and more from the excitement of saying, right, we need to bring in AI.

Speaker B: Yeah.

Speaker A: And I always feel it's slightly patronizing to say that, but I feel there's boardrooms that are not far away from that conversation and then having to re engineer that backwards afterwards to say actually what have we got to get right before it actually works in, in practice. And a lot of that comes down to that intelligence fabric piece beforehand. Tell me about some of the conversations you're having there to allow people to get to that sort of, um, I guess level of sage, level of wisdom to be able to bring that together.

Speaker B: Yeah, yeah, yeah, definitely. So intelligence fabric is really a unifying concept around how you build data infrastructure that solves for the sort of mission critical, highly regulated nature of uh, buy side workflows, buy side and sell side workflows. Right. And also reduces the friction for you to run next generation AI on that data.

Speaker A: Right.

Speaker B: Now what it does is it brings together a couple of concepts. Number one, uh, the ability to uh, aggregate a vast range of data sets into a point where you start off with a very high level of trust and you have a starting point for data lineage. So think about it as governed foundation. So I described this two sided network of 850 plus data partners, 400 customers. Right. some point someone needs to agree that this is the sort of source of truth around all the data sets that we're aggregating. Right. If you're a buy side firm, if you're, let's say aggregating 100 data sets. Right. This is the sort of truth, the governed source of truth. And I have visibility, transparency into everything that's happening around that data. So that's the foundational kind of starting point. Right. Governed data, um, with high levels of observability. So you can kind of see what's happening if something goes wrong, you understand what goes wrong. You have high levels of lineage. Again, you can kind of trace things back to this kind of governed source. So that's the foundation of the intelligence fabric. Right. After that you then start to surface AI ready data sets. Right. So these are data sets that have been enriched, they've been customized depending on what a client wants. They are ready to be, to be used to drive the client's workflows. Right. Whether they want to surface the data through an API, whether through an MCP server or directly into a UX for a human. Right. So you're creating that level of master data sets that a client can surface with a tremendous amount of optionality in terms of the end use case, agent, machine or uh, non agent machine or uh, human. Right. So that's the second piece. The third piece is uh, intelligence that you can help the client with on top of that data. So rhymes. And of course many other firms will have uh, analytics use cases that clients can run uh, on top of that data or the clients can develop their own AI use cases of course, or AI workflows on top of that data. And then underpinning all of that is operational operation. So running this fabric at scale, uh, with the right levels of rigor, um, with the right levels of telemetry, observability, disaster recovery, very important in our industry and also operational, uh, processes that enable you to triage and diagnose issues that happen in uh, the data very quickly.

Speaker A: Right.

Speaker B: That's what the intelligence fabric is really four pieces. It's governed data where lineage starts. It's the mastering and surfacing of AI ready data sets. It's uh, you know, use cases on top of that, uh, either yours, uh, as a vendor or client, use cases supported by deep operational uh, uh, uh, you know, deep operational support.

Speaker A: Right. And that sounds a very clear way of putting it. And obviously when you, when we're seeing this huge wave of AI, uh, hype coming through, if want of a better word than hype, I don't necessarily feel like that, but the demand coming through from. It's probably a better word when you've got this AI demand coming through and people wanting to embed it into their businesses and falling uh, back because they're not ready for that sort of stuff. It's a lot of stress around actually implementing it into action. So firms aren't necessarily ready for this at the moment, are they? But it's not a huge leap to get themselves into that sort of position with the right partnerships I guess.

Speaker B: I think that's right. I think that firms are definitely primed for a different way of thinking about how to unify and simplify their data foundations. There's a level of interest, uh, that I've never felt before around this specific topic again, driven by how do we actually sort of think about data foundations in the context of rapidly advancing AI? So I think the need is very much there in terms of what it takes to get there. That does vary by firm. And there's quite a significant, uh, I would say of dispersion, if you will, in terms of the readiness of firms to build data infrastructure that's going to really be the right thing for them in this, in this next generation of AI.

Speaker A: And what, what drives that? I mean, is it, is it, is it about their general, uh, tech adoption? Because I know there's sort of slightly more old school, ah, firms in, in this space and there's people there who are constantly looking for alpha. And I, I can, you know, we probably can both think about the immediate names you think of in both of those categories straight, straight away. But is that what's happening or is it just about the advance of the team and the technology, the appetite of the senior teams and the PMs in those sort of spaces? Where does that come from, do you think?

Speaker B: Look, it's all of the above, right. And the ingredients of the mix really depend on the firm. I mean there are definitely some firms, I would say, that have built up their data foundations over time. Uh, like many firms, they have different systems, disparate ontologies, different ways of describing kind of the same thing. All right. Different technical architectures. And all of, all of this fragmentation kind of becomes more and more of an issue as you start to want to drive, you know, more and more automated workflows.

Speaker A: Yeah.

Speaker B: Because it just becomes hard to kind of, you know, create that level of abstraction, if you will, by that interfaces with the AI.

Speaker A: Right.

Speaker B: So that's where a lot of the problems exist. Right. And there is, I would say, a real, a real sort of mixture, if you will, around uh, around investment and sort of historical appetite to think deeply about data, uh, foundations and designing data foundations from first principles, uh, to solve for the kind of AI capabilities that are coming out right now. I think that's probably the shift right now that a lot of firms are trying to kind of think about. So it's really mixed in terms of where, you know, where you know, where

Speaker A: clients are and how much does the, does the market play into that as well? Because it's um, you know, I think in peacetime it's always, uh, easier to think longer term, um, and strategically we've had some fair volatility for one reason or another over the last few months. Um, probably since the last seven that you, you took seat. There's not been too much at times where everything's been running normally, you know, across, across the globe. And sometimes I feel that that sort of can um, push things onto a back seat. To me it feels like this is something there which is more front seat because people recognize the opportunity to get this right and the huge advantages to get this right. It kind of takes me back, you know, 15, 20 years to the, you know, to the zero latency race and all that sort of thing where people are saying, right, this is our opportunity right now. I think the real use case and adoption of AI becomes an alpha, becomes an advantage for people, um, that feels like people are properly invested to getting that right. Is that something you're saying or are you seeing, you know, the volatility in the, in the buy side space causing a little bit of looking left rather than right?

Speaker B: I think it's, I think if that makes sense. Yeah, yeah. Look, I, I was at a, I was meeting with one of our largest clients up in Boston a couple of months back and they said to me the only, and they are a, an equity, uh, an equity fund manager. They said to me the only priority high that's, that's bigger than generating alpha to them is AI. Right. So I think that sort of sets the stage, stage around, level of priority.

Speaker A: Yeah.

Speaker B: Uh, in terms of focus, you know, in some ways volatility, uh, surfaces more issues with data.

Speaker A: Yeah.

Speaker B: Because in a very volatile market, uh, you know, if you have uh, a problem with a, a data field, you make an incorrect trading decision, uh, based on that your, your economic impact could be greater. Right. And, and, and obviously you're, if there's an issue with data, you need to remediate that faster in order to react to the volatility that's going on. So in some ways I think volatility has surfaced more and more issues with underlying data sets. But I think that's been going on for quite a long time.

Speaker A: Right.

Speaker B: Um, uh, but I think a lot of the challenges, it's resource resourcing and focus, you know, is probably the primary constraining factor in a lot of our clients. Right. Getting really dedicated resources to focus on uh, data foundations on data, strategy on data roadmaps, on AI roadmaps and so on and so forth in the context of uh, a, a volatile market which inevitably surfaces issue with the underlying operating model, which have to be dealt with kind of in real time.

Speaker A: Right.

Speaker B: But, but that I think is like, that's normal now. I mean, I wouldn't expect, you know, things to settle down at all. In fact, it could, it could sort of increase.

Speaker A: Right. I think this is really interesting because as you're talking, I'm thinking to my day job, the recruitment side of things in this sort of space. I think we've seen more and more demand over the last year to 18 months within this whole data world. So when you look at, um, the fusion between, it's a, it's an absolute fusion between AI and data.

Speaker B: Yeah.

Speaker A: Um, and more and more companies, uh, who we're working with, looking for it to build AI teams internally, but also beefing up their data teams alongside that and having the two very closely merged together alongside a much more collaborative world of utilizing expert vendors at the same sort of space that sort of, you know, we've spoken um, many times in the show about the move from buy or build to buy and build.

Speaker B: Yeah.

Speaker A: And to have a collaboration of tools and people looking at various expertise areas within that alongside ever increasing teams. Is that something you're seeing as well? When you're talking about people putting their, I guess, money where their mouth is with regards to their investment into AI? It's about building teams as well as investing in the right technology as well.

Speaker B: Yeah, look, I think, I think it's a, I think it's a bit of both.

Speaker A: Right.

Speaker B: And no doubt what I've seen is that the, the budget firms are allocating to solving data foundations is increasing.

Speaker A: Yeah, absolutely.

Speaker B: At the same time, the ability to, for firms to do more themselves is also increasing.

Speaker A: Yeah.

Speaker B: So I think for certain types of technology platforms, the pendulum is swinging from build to uh, from, from buy to build.

Speaker A: Yeah.

Speaker B: Right. And in certain areas. In other areas it's swinging the other way.

Speaker A: Yeah.

Speaker B: Right, yeah. And I think it's going to take a little bit of time to kind of figure out where exactly the dust will settle on that.

Speaker A: Yeah.

Speaker B: Uh, it's only when it comes to the field, the field that I'm in.

Speaker A: Right.

Speaker B: Which is sort of mission critical data infrastructure.

Speaker A: Yeah.

Speaker B: There is, there is a view that, you know, the complexity of managing data pipelines at scale and dealing with, uh, you know, hundreds and hundreds of data partners is, is something that is challenging to bring that in house. Right. Um, uh, so I feel that the pendulum is swinging a little bit more towards, you know, how can we partner with specialist fintech companies to do more of that stuff for us.

Speaker A: Yeah.

Speaker B: Uh, which Obviously is a good business to be in for us.

Speaker A: Yeah, absolutely.

Speaker B: Uh, at the same time it means that we need to build at rhymes, platforms that scale even more to support even more clients.

Speaker A: Right.

Speaker B: And one with even more unique, sort of even more unique needs.

Speaker A: Right.

Speaker B: One of the things we've done to, to ensure that we are well positioned to drive that next level of demand is our partnership with databricks. Right. We announced this a couple of months back and we're super excited by that. Right. Databricks, uh, is I think a fascinating company. Obviously very deep heritage in distributed computing with Pyspark. Um, very deep heritage in being AI first in terms of designing uh, data platforms that can help with, you know, highly scalable, you know, machine learning, AI workflows. And uh, have invested a lot in a lot of the concepts I talked about. Right. Which is embedding things like lineage, explainability, observability deeply into the data infrastructure stack. Right. So as we think about the underlying platform that will give us the ability to kind of scale into the next level of demand. Given the focus of these clients on sort of AI workflows or AI first workflows. We think the databricks platform is fantastic infrastructure for us to build on.

Speaker A: Yeah, yeah, absolutely agree. And it sort of ties into this, this move that we're seeing from uh, managing data to sort of moving to activating governed intelligence. You've spoken about this before in your sort of pillars, I guess expand on that a little bit.

Speaker B: Yeah. So govern intelligence for us is the ability to many uh, of the concepts I talked about before, the ability to trust an agent and understand what the level of data, uh, uh, decision grade data, what that actually means for you to be able to trust more and more agents, uh, you know, doing more and more across your investment life cycle. Right. That's fundamentally what it is. Right. And a core part of that obviously is governance. Right. Governance meaning you have visibility, transparency into exactly what data is being used by what agent. And if there's an error that the agent made, you can trace it directly to, at a very granular level to the data, ensuring that you have the right permissioning around the data sets. Obviously as a number of agents your firm is using increases, uh, it sort of tees up very complex permissioning conversations. Right. Um, and how you apply those at a very granular level and govern those across your data sets. I mean many, many ready roads lead to the broad topic of governance, uh, which again is best solved at a deep data infrastructure level.

Speaker A: Yeah. And I guess this is, this comes Back to this whole concept that happens in so many areas of life. Um, sport, business, military, wherever you look at it is. It is the, the unglamorous stuff that allows the glamorous stuff to happen.

Speaker B: Yeah.

Speaker A: And that feels to me this is, this is the. If you get this unglamorous piece right, there's no way that you can do everything that you really want to have happen that can be so transformative in this industry at the same time.

Speaker B: Yeah.

Speaker A: And it's a really important bit to get to. You know, you mentioned it before. The plumbing, Right.

Speaker B: Yeah, the plumbing.

Speaker A: Right. Everything else can happen.

Speaker B: Yeah.

Speaker A: Afterwards.

Speaker B: Yeah, that's right. I mean, this is, this is critical plumbing. It's becoming more critical to keep those pipes, uh, clean, I guess, given everything that's going on. But that's the right analogy.

Speaker A: I love this, uh, this piece and again, I'm going to quote it directly because I think it's a, uh, great takeaway from it that intelligence without trust becomes a liability. Yeah, exactly what we've been talking about all the way through here.

Speaker B: Yeah, 100%. And it's a velocity thing. I mean, you're looking for massive speed ups in decision making.

Speaker A: Yeah.

Speaker B: Massive speed ups in processing. And you know, if you get things wrong at the trust level, those massive speed ups sort of, it amplifies issues for you. M. Right. And that's kind of one of the things I think that getting the foundations right really helps solve.

Speaker A: So I've been, um, doing this, this show weekly now for the best part of five or six years. And I've been in the industry for best part of just over a quarter of a quarter of a century, speaking to people within the, this whole ecosystem. And what I'm finding more and more is that over the last year or so, I don't think I've ever seen founders, CEOs, C suite people, in fact, pretty much anyone in the industry who's associated to this world as excited, as personally invested as any time over that quarter of a century. Um, you were talking about beforehand about how you've gone back to do a master's in this. It's something there which has become more than just just a day job. The, the energy and the, um, the, the thought and the creativity and the thinking ahead that people are doing at the moment is like nothing I've ever seen before over that sort of period. It's exciting. Give me your take on where this is heading on an industry basis.

Speaker B: Yeah, look, uh, it's really interesting because we had an innovation forum, uh, you know, a couple of weeks back, actually, here in London, uh, and we've had a couple of other events. I kind of asked this question, you know, um, you know, when you think about the level of change in two years, five years, 10 years, on a scale of one to 10, like, what do you think that looks like? Right. And pretty much everyone is like, in 10 years at 10, 10 being defined as, uh, so dramatic, it's probably unimaginable today. Right. So when you think about, uh, sort of, you know, what it takes to kind of get to that level of change. Yeah. I mean, I think what we're going to see over the next couple of years is going to be, uh, uh, very dramatic and very transformative. Now, we work in a highly regulated industry for very good reason.

Speaker A: Yeah.

Speaker B: Um, so we have to be very thoughtful about how we, uh, implement, you know, these sort of frontier model capabilities. Uh, but I would expect to see a pretty dramatic acceleration, improvable POCs in production, ready POCs in, uh, workflows that are fundamentally reimagined in our industry, uh, that will create a next level of scale and efficiency that, you know, we've never seen in our careers. And I think that's the opportunity. It also, you know, comes, uh, with a tremendous amount of risk. Another conversation I have with folks is, you know, what do you think could go wrong? And many things could go wrong. Right. I mean, if you start to get agents you don't fully understand, start to act autonomously and make mistakes, I mean, that, that could be extremely problematic. Right. And that goes back to the governance.

Speaker A: But do you think you'll be able to be, uh, to get to that sort of stage? Do you think the industry, which had such strong, you know, uh, guardrails all the way through it will allow it to get to a stage where that gets loose, to allow that sort of thing to happen? I think that's a really interesting thing.

Speaker B: Yeah. Look, I mean, obviously I'd like to think that we've got the right, you know, governance in place.

Speaker A: Yeah.

Speaker B: You know, the right, the right compliance mechanisms to ensure, uh, that doesn't happen. But it is new territory for us.

Speaker A: Yeah.

Speaker B: Of course, you know, in many cases, and with newness comes risk.

Speaker A: Yeah.

Speaker B: Uh, so it wouldn't surprise me if

Speaker A: we do start to see a few

Speaker B: things like that happen. Hopefully they won't be at a really dramatic scale. But it all goes back to ensuring that you have the right governance, the right data. Right. The right processes, the right observability in place to trust that implementing that kind of level of workflow.

Speaker A: Get that bit right at the start. Where do you think we'll see the most dramatic impacts? Where do you think the biggest opportunities are?

Speaker B: I think many, uh, aspects across the investment life cycle. I think a lot of areas like sort of reconciliation, uh, a lot of areas around managing trades, trade exception, I think in operations, triaging and understanding data pipelines, for example. We think there's a big opportunity there. Um, obviously machine learning has been around in areas like portfolio management for many, many years. It's uh, not necessarily got back into other areas in the trade lifecycle like middle office work and back office work. But I think we're going to see a lot more opportunity there as well.

Speaker A: Yeah, absolutely. And in terms of, I mean we've spoken there about uh, the opportunity, but the barriers to do that, to avoid it come predominantly in your mind down to trust. Is there anything else that people would think about or should we think about about what stops them to get to that sort of level?

Speaker B: Uh, look, trust is definitely the biggest one. I think there are core, uh, call it foundational architectural constraints that, you know, if called like data architecture, and we're going back to that, uh, is not designed correctly. I think that can be a major inhibitor. So I think, I think it's trust, I think it's architecture. I uh, also think it's the ability for uh, talent or the next generation of talent to really embrace these workflows and figure out how to innovate around those workflows. I think there's a big talent question that's a topic for another day maybe. Yeah, yeah, uh, around, around.

Speaker A: Touch on that because I do think it's a really interesting piece. So, so when, when we're sat there and we're talking about the right talent for doing this, I think people are sort of dancing around what that actually looks like in my, in my estimation at the moment. There's, you know, we were talking before about our kids. Yeah, kids coming through at this sort of age. There's certain people are saying that that's a great advantage and we're going to invest in the graduate end. Um, because those people have no preconceptions about uh, you know, what life was like beforehand. So they can think on a completely different level. Other areas are saying, right, okay, we need to have that experience and you have to see this uh, this diamond as opposed to a pyramid where there'll be a big middle but not an awful lot of investment into um, entry level talent coming into the industry.

Speaker B: Yeah.

Speaker A: Do you have A view on what that looks like. Do you think people need to evolve with experience or. I always worry about this, this, these sort of areas where if I go back to when we first launched the business in 2010, there was this m, this missing generation of people who weren't invested into in because of 2008. So everyone wanted 1 of people. I was doing quite a lot in the fixed world at that sort of stage. Everyone wanted to fix people with two years experience. There weren't any. Yeah, they just weren't in that sort of space. And I wonder about whether companies are starting to recognize that investment uh, at a junior end is important or whether that talent is about bringing the best people from other areas or it suddenly becomes quite a level playing field in my opinion. And that's quite an exciting position to be in.

Speaker B: I think it does. I look, I think that the ability for uh, folks to get up the learning curve very quickly even in very complex areas like uh, sort of investment management workflows.

Speaker A: Yeah.

Speaker B: I think that's dramatically accelerated. I mean you can go to a frontier model, you can generate a course and help me understand the trade lifecycle.

Speaker A: Yeah.

Speaker B: And it'll be really, really good stuff. So I think you can accelerate the learning curve. And as I think about of the next generation of talent, like what do you need to survive and thrive? I think you do need very deep domain expertise because you need to ultimately uh, be able to design and judge how well these AI driven workflows are doing. Right. And you need really deep domain expertise. I also think you need a skill of understanding how to use AI responsibly.

Speaker A: Right.

Speaker B: So I think about domain expertise using AI responsibly. You know I might argue that those are the two foundational skill sets that the next generation of talent needs.

Speaker A: Yeah, no, absolutely. Great. And I think it's really interesting to see people come into that as well at the same time too. So um, let's start to wrap up a little bit about where we're going. Give me some exciting stuff. You've been in the business for seven months now. Tell us what the rest of the year looks like. Tell us uh, what's exciting. Coming up for rhymes.

Speaker B: Yeah, look, definitely, uh, we are um, doing a lot of work with databricks. I think that's super exciting. Where we're going to start to migrate a lot of core data pipeline operations onto the databricks platform. Seeing things like 5, 10x, uh, increases in, in operational efficiency and speed, uh, generating AI ready data sets. Right. That reduce the friction. It Takes for an end client to actually run their own AI, agentic AI and so on, on these data sets. Super excited about that. We're uh, also disrupting a lot of our own internal processes. So looking at for example data implementations which might have taken 12 to 18 months, taking that down to two to three months and then pushing the boundary even beyond that.

Speaker A: Right.

Speaker B: Um, you know, being a physicist by background, I always love to challenge my team. You know, they ultimate rate limiting step of the laws of physics. So once in a while just apply that lens and kind of reimagine all of your processes and we have some kind of great conversations around that. And they're conversations that are in some ways grounded in reality. Because I think a lot of what AI is giving us right now enables us to think differently with that kind of lens and think about a process that might have taken a certain amount of time and really shrinking it down to like 90%, 95% of that time. And that is incredibly exciting for us and obviously for our clients. Right. Um, and so we've looking to bring a lot more, a lot more, uh, I would say sort of disruptive, uh, ways of thinking about data, uh, and data processes which ultimately benefit our clients in a tremendous way, you know, to the market.

Speaker A: You know, it's really encouraging because I've known Rhymes as a business for a long, long time and actually to see it now, uh, it's something I'm hearing a lot about. There's a lot of excitement, a lot of hype around it. So, um, keep up the great work. It feels like, it feels like right company, right place, the timing is fantastic for it. You must be loving it, must be enjoying.

Speaker B: Yeah, yeah, great stuff.

Speaker A: So if people want to reach out and find out a little bit more, I imagine connect with you on LinkedIn. Yeah, go, go to the website, do all of the above. Anywhere else to get in touch with you guys.

Speaker B: That's perfect.

Speaker A: Sounds good.

Speaker B: Great, thank you, Toby.

Speaker A: Thank you having on the show. Thanks so much. Yes, thank you and thank you all for watching. We will see you soon on another episode of FinTech Focus TV. Thanks a.