Modern Money SmartPod
Modern Money SmartPod
ServiceNow's Ryan Clare on Capital Markets Tech: What's Now and What's Next
Technology has long been a key component of any operations strategy in financial services. But the recent rise of AI is changing the operations landscape at a rapid pace. Ryan Clare, Global Head of Capital Markets - Go To Market at ServiceNow, outlines how technology is powering today's markets and how AI is being leveraged to help operations evolve today - and in the future.
Key highlights (with timestamp)
What are the hottest topics in capital markets tech right now? (0:50)
What kind of impact is AI having on market infrastructure? (1:57)
How is AI reshaping trading? (3:01)
How is AI playing a role in risk management? (4:15)
What is the future of AI when it comes to compliance? (6:26)
How is increased automation changing the operations landscape? (7:54)
What are the biggest concerns around the deployment of AI and other technologies? (9:32)
Are there enough developers with the skills needed to create all these products? (10:37)
Aside from AI, what other technologies are you keeping an eye on? (12:11)
Do you have any bold predictions for what we're going to be talking about when it comes to technology and capital markets two or three years? (16:03)
How can operations professionals prepare for all this rapid change? (16:58)
(Note: This transcript was created using AI. It has not been edited verbatim.)
Sean McMahon 0:10
Hello everyone, and welcome to the Modern Money SmartPod. I'm your host, Sean McMahon, and today we're going to spend some time talking about technology in capital markets. I know that sometimes it can feel like there's a never ending fire hose of news about all the latest and greatest technologies that are available in the markets. So today to help discern what's possible now, what might be just around the corner and what is perhaps a few years away, I'm joined by Ryan Clare, the Global Head of Capital Markets - Go to Market at ServiceNow. Ryan, how are you doing today?
Ryan Clare 0:46
I'm doing great. Sean, thanks for having me on
Sean McMahon 0:50
Great. I'm excited to have you join us in a minute. I'm going to ask you to dive into some specific areas where technology is changing, the way that markets operate. But for now, I just want to step back and ask you a broad question, what are the technologies that you see are commanding the most attention right now?
Ryan Clare 1:06
Well, I think the markets are taking over at the moment. And if you could have an AI, a model that could process every decision being made around the world and perform your trading on top of that, I think you'd be a millionaire. So unfortunately, I'm gonna have to carry on my day job for a little bit. So in the industry, I see AI is just normalizing data at such a rapid pace, I think there's an opportunity to increase fraud detection, anomaly detection could be deployed. I think everything could be more real time. And the rate of the implementation of it, I think it can help just generate more revenue for firms make better trading strategies. And I think with the infrastructure, a lot of the exchanges can do real time trade surveillance and even do risk profiling, which I think is going to be very interesting in the future.
Sean McMahon 1:57
Alright, now I'd like to drill down into specific aspects of capital markets, and the role AI is playing in those, we'll call it sub sectors. So I'd like to start with just market infrastructure. What are you seeing there?
Ryan Clare 2:10
So as I touched on, I think it's the trade surveillance and the ability to just process enormous amounts of data and to get real insights. I think that's always been a challenge for most systems, to be able to process the data and then to be actually out, to get information out of it, you have to run a model for hours before you could actually see the results. I think being able to see what's going on in the market, whether you're at the exchange or even if you're in an organization, and change your strategies, I think it's going to be really powerful in the future. In the end, it's going to help with all your back office functions as well. It's going to help to make smarter clearing and settlement decisions, funding decisions. A lot of platforms are out there going to have to upgrade to include this as a part of their offering. People are going to expect it when they're purchasing products.
Sean McMahon 3:01
Okay, now let's talk about on the trading side of things. I think a lot of people, when they think of AI, they might think it's just all about speed, speed, speed. But what's your view on on how AI is reshaping trading?
Ryan Clare 3:14
So I think the most visible is the algo trading, this ability to do high frequency trading strategies all now being powered by AI rather than waterfall logic. I think the most interesting part of it, or the most cool part that's going to come into play, is that those models are going to be out to adapt as market conditions change. They're going to be able to consume non structured, unstandardized data and be able to flex the model. A good example is the model will be able to scan negative news or look at articles being published real time, digest that and change the strategy and make decisions based on that which I don't think we're you know, we've ever seen anything like that before. Could you imagine, you know, the trading changes, because there's a new bill been passed, it consumes that information. Is able to adapt and provide insights and different trading strategies for a hedge fund. I think it's going to be enormously powerful.
Sean McMahon 4:15
Alright. And I want to kind of pull on that thread for a second. Here you talk about models being able to process negative news. And the next area I want to dive into is risk management. And I kind of want to look at it through the lens of, you know, I think it was a week or two ago we had just, like, a false news, or fake news, and it actually hit the markets and, you know, caused a quick downturn before participants realized it was fake. So what role can AI play when it comes to risk management?
Ryan Clare 4:40
I think, in risk management, being able to do exposure tracking and be able to take in real market events and run them through an algorithm to see if they're correct, like you said, but also take client information and other statistical information and be able to do. Do scenario based planning. So typically, when you're doing risk management, you're looking back historically. You're not taking stuff in presently, you're taking stuff maybe over six months. Then with that information, you're then sort of trying to lay out a pattern for the future. Like you said, would you take in that negative news? You could compare it to the previous day's news, or what markets or client activities going on, and then be able to make a more realistic decision. And by having that, you can do stress testing, continuous risk assessments of Counterparty exposure. But you can also look into playing, sort of like the Monte Carlo events in real time scenarios where typically you're downloading that data, crunching it, trying to analyze it, but it's too late by that point, just as you said in the markets, the market's already taken a hit by the time people have realized it's not real news. I think you need AI to police AI. The good example of that is my daughter wrote examination paper last year, and her professor said that basically accused her of using AI to write it. She then had to use an AI package to prove that it wasn't written by AI so and that's just that the college. Could you imagine that's going to be needed throughout the market as well, and you can start to use that information to basically check that the AI is performing correctly, so multi facet of AI being used.
Sean McMahon 6:26
So I love that story about AI having to beat AI or keep track of AI, even in academia. So let's talk about how that applies to the markets. Right? It sounds like your daughter's teacher was accusing her of, quite frankly, not being compliant with the rules. So how does AI factor into compliance for capital markets?
Ryan Clare 6:44
I think this is going to be a real tough one. The whole compliance industry is going to have to be looked at. It will help in the lot of the compliance processes. So if I think about where AI is going to change compliance again, back to trade surveillance. Did a lot of remediation projects on trade surveillance. It's quite a hard topic to do. You only have small amount of time to make sure you report your trades and monitor them correctly and demonstrate that to the regulator. But also, there's a Herculean lift that goes into KYC and AML checks. I think that's where AI is really going to help within compliance, you could have AI agents across the full life cycle of those processes. You know, as we mentioned, talking about negative news, but looking at company profiles, scanning documents, just the amount and ability to process that information and come up with a decision. That's where AI is going to be able to help a lot. It's going to be able to take a lot of that compliance burden away of where it's doing, busy work versus complex analysis to make sure that you're meeting regulatory requirements, more just processing information because you're being buried in data and documents.
Sean McMahon 7:54
All right, now you've used the phrase busy work, and I think a lot of folks when they think of busy work, they are always seeking ways to automate that. So how is increased automation changing the operations landscape at financial services firms?
Ryan Clare 8:08
My opinion, it's going to increase the job satisfaction and the quality of life for the operational people. It's going to allow them to focus on the high value tasks and be able to focus on the important exceptions and be involved in more change type work, because the way AI is changing so much, there's going to be a requirement from your individual users to be involved as part of that program, whether they're testing the output of it or get, you know, applying logic or their experience of how to make it better. So I think it's going to adapt and change the way people actually operate in the operations department, and reduce the gap between them and technology. So I believe automation is no longer a cost saving. It's about scalability, resiliency and flexibility, from traditional RPA to agentic AI, I think we've got to realize that we have to integrate fully across all types of processes and workflows, and people should be like I said, being more worried about the high value exceptions. We should be able to increase straight through processing and generally provide a better quality of life for the employees, but also get a give a better client experience. I think that's the way operations is going to change. It's going to become an added value department that helps you gain market traction and provide better services to your client, and that's what AI is going to help do.
Sean McMahon 9:32
And now we touched on compliance a few minutes ago, but I want to dig a little deeper there. So what are some of the biggest concerns that firms have about how to navigate the deployment of AI and other technologies.
Ryan Clare 9:44
I think, especially with AI, it's black box technology. It's like, how do you actually validate that the model is doing what it should be and the inputs are correct, and it's being governed and controlled like you've got to have a really good, strong data governance to make sure that the. Outputs and the prompts going into your model aren't being abused or incorrect, because if people are making decisions, especially if they're making trading decisions off the back of the output from the AI, that's where compliance are really going to have to be strong. Back to our saying completing AI with AI, I think there's going to be a new market for compliance type technology to come into play, where you could use AI to validate, you know, similar to like, you know, we've got aI at the moment summarizing your emails or summarizing data. It's like summarizing the models results and outputting them in a digestible way that a human can validate and say, Yeah, okay, that makes sense.
Sean McMahon 10:37
And then what about talent? I feel like there's so many folks focused on building out these AI models that can speed up trading decisions or increase data management, but now you're saying there might be a market for AI for compliance. So do we have enough developers that are skilled enough to create all these products?
Ryan Clare 10:56
I'm not so sure. I feel like as a starting point that the compliance teams are going to have a lot of competition for getting talent, so even non technical ones, having people come in where they're moving away from knowing the regulatory governance of the market to more now understanding how AI needs to be governed. I think it's going to create a new type of compliance. Well, bit like years ago, you used to have the head of it, and then the CISO was created with all the cyber threats and all the you know, as technology started to grow, I feel like there's going to be a similar transaction in the compliance team, where you're going to need semi technical people in those more regulatory type roles being able to understand AI and how it works and how it all comes together, to even then, be able to help the developers to create those large language models that can be deployed. There's going to be people moving from task driven activity to more added value type processes, similar into compliance. I think it's going to go from more of the technical understanding of how rules work to also being technically understanding how technology works and how that interacts with those regulatory requirements and rules.
Sean McMahon 12:11
Alright, now we spent a ton of time talking about AI. You know, we're doing it today. Everyone I hear at conferences and around the industry is doing the same thing. But what are other technologies that we should be keeping an eye on when it comes to capital markets? Well,
Ryan Clare 12:25
I'll be remiss if I didn't say workflow coming from ServiceNow, you've got to have good foundations. So you've got to have your workflow, you've got to have your data, you've got to have good governance around all of that as a must. But for me, the two big things are, one is cloud. You know, everyone thinks that cloud has already seen its day. It's calm, it's done. Everyone's aware of it. Everybody's using it. But you just the amount of processing power that the cloud infrastructure is going to need to be have available to the end user is ginormous, in the sense of, I don't know if you saw some stuff around chat DBT recently that the electricity and the CPU usage for people saying yes, please and thank you when they're requesting chat DBT to process something is just it's gone off the charts. So you got to think about as we're requiring this real time processing to happen, the computing power behind it and the cloud infrastructure is going to have to keep up with it. So I feel like Microsoft and Google, they're going to have to restructure their strategy. The more we give AI to the individual consumer, whether that's for an organization or as an individual, the energy that's going to acquire to run those models is going to need to be reviewed and expanded.
Sean McMahon 13:46
Yeah, I saw that bit about common courtesy when it comes to the models and please and thank you. And Sam Altman mentioned that, yeah, it's costing millions of dollars to process all that data, but that it actually is is significant, because the models detect courtesy, you know, rather than rudeness or things like that, and it adjusts accordingly. So, but to your point, it's a ton of processing power that's needed for it. Yeah? Any other technologies that we should be looking ou tfor?
Ryan Clare 14:11
Yeah. And I think it's privacy preserving technology. I read an interesting article recently about federated learning. I think the challenge we've always had with AI is how much you put into the model. You know, as you work for a financial services organization, you're not allowed to use ChatGPT, you're not allowed to upload sensitive information, you're not allowed to put preparatory information into those models. So unless you have it encapsulated in your organization, you can't really get the full benefit of the data. I think as federated learning comes in, where you can just listen to the activities and not have to consume that data, but you can have enough information around it to be able to process in it and enrich the model, I think that's going to become very important, especially where you have rules around. Sharing sensitive information around people, telephone numbers, addresses and things like that, elements like marketing, especially for myself, in the go to market team, you'll be able to use that information, but it won't be stored in the in the model. I think as companies start to crack federated learning and using that sensitive data without sharing it with the model. I think that's going to be the next big one, because that opens up a whole new world of AI, where you'll be able to reuse the AI, large language models as an organization from customer to customer. So what I mean for that is, if you deploy a model at a customer and you use their information, you can't then take that proprietary knowledge and go somewhere else, because it's got all of their data in if you're able to federate that and only take out listening points to that information, in theory, you'd be able to expand your capability of deploying AI quicker and using all that history Without taking that proprietary data just just as one example.
Sean McMahon 16:03
Alright now when we talk about technology markets and AI, you know, we weren't talking about AI as much as we are now three or four years ago. So I want to kind of ask you to look into a crystal ball and see, do you have any bold predictions for what we're going to be talking about when it comes to technology and capital markets two or three years from now.
Ryan Clare 16:23
I'm waiting for the first firm to come out with 24/7 trading desks, fully automated through agentic AI, giving them the ability to hedge, potentially even negotiate pricing. I'd love to see a model do that, but also rebalancing the book, resolving issues and breaks. Obviously, would have to be some of the more simplistic products, but being able to do that and offer that across the world, I think it'd be quite exciting. Might be a bit of a stretch, not sure it would get approved by some of the registry bodies, but I think that would be really exciting. Having a fully automated trading desk, front-to-back.
Sean McMahon 16:58
It'd be dynamic to see that for sure. So last question, you know, with everything evolving so rapidly, if I'm a business leader operating capital markets right now, what steps should I take to set up my operations so they're prepared for the future?
Ryan Clare 17:13
So coming from a transformation background, I've got two points. The first one is data, data, more data. You've got to have a good, high quality data for any of the technologies we've been talking about to work. It's garbage in, garbage out. It has to be governed, maintained. The reference data has to be supported around it, and that's where you're going to create new roles, like data stewards are really going to have to be on their game to make sure the inputs and components are being used are correct, otherwise, you're just going to create outputs that no one can use, and it's never going to produce the results that you've invested for and give you the ROI. I think the second one is enterprise innovation. You know, I touched upon it a little bit earlier, with the operational people becoming closer to technology. I think the days of technology driving an initiative on their own have gone. I think it's a joint effort. And at least when I was back in the industry, I found that the best programs delivered the best results. Where you had enterprise engagement, we had all the different levels, from a function and department all collaborating together, coming up with ideas. Some of them aren't great ideas, and some of them are wonderful ideas, but least they're creating that innovation, and then the adoption is quicker, and in the end, you'd end up getting better results. The point I'm making is the innovation journey isn't just text responsibility. I think it's everybody from the top down, from the bottom upwards. I think some of the young individuals coming into the market, they're already coming in, skilled in writing Python, potentially even written some of the AI models themselves. They're using configured dashboards in reporting. Instead of using Excel, one of my interns was able to build Power BI dashboards during a 10 week internship. You know, if you went about five or six years ago, that probably wouldn't have happened. So I think some of these young individuals coming out are very highly skilled, so I think there's a lot of quality thoughts and requirements that can be gathered from them.
Sean McMahon 19:20
All right? Well, yes, I guess the future might be in the hands of the interns, as they say. Well, hey, listen, Ryan. I really appreciate your insights today. Thanks for joining me for this podcast.
Ryan Clare 19:30
Thanks for having me on.