Christopher Concannon
Analyst · Alex Blostein of Golden Sachs
Great. Great question. And obviously, an area that we've been spending a lot of time on. I think everybody has been spending a lot of time on. This is one that I'm super excited really for our position in the market when it comes to data. And there's really three reasons why I'm so excited about the data opportunity. One, we have the best source data in fixed income. When you think about the breadth of our product the protocols that we use, RFQ is a phenomenal source of data because you have both the inquiry from clients around the planet as well as the response from both dealers, alternative dealers, hedge funds and clients now. So the data source that we have is the raw data and very important. The other key ingredient is we have not sold what I call the good data. We have sold things like CP+. We have sold Access All. There's really good data underlying the platform. And we've been for now several years on these calls saying we will not sell all our data. It's too important to the execution solutions that we are building. The other reason why I'm excited is because we also protect the data that we sold. We have very tight restrictions, derived right restrictions as well as AI use restrictions on the data sources that we are selling. So we feel like we're in a very good position to take advantage of the footprint of data we have. So also the other opportunity that we have is we've been investing heavily in AI when it comes to our date for a number of years. It's not a new topic for us when it comes to data and data products. So within the use of AI to produce data or produce product, I'll be more clear. We have AI-derived real-time data called CP+. We've had it for quite a number of years. It is now across all products. and certainly well regarded in the market, certainly viewed as real-time benchmark for trading U.S. credit, and we've been getting awards with that product as well. We are now able to predict using AI real-time block pricing based on your direction, sell or buy. We are predicting using AI liquidity levels in the market. We were also using AI, leveraging AI to predict what we call counterparty selection or dealer selection as well. And then obviously, one of my favorite use of AI is being able to predict when it is advantageous to provide liquidity. This is a very important component. It's one of the drivers behind our algo solution, which is the theory that clients who have historically crossed spread for most of their fixed income experience may present with the opportunity to avoid crossing spread. And we are now leveraging AI to actually help with predicting when is that beneficial for you to have patience and wait. So again, AI from a product and execution solution is really what's the most exciting part of our day. We are also, just to be clear, piloting a new AI solution with some of our clients. And the areas of exploration that's super exciting for us is, first, AI-derived real-time market intelligence. As I mentioned, our market footprint is quite broad. So we see the markets from the start of APAC through the trading hours of Europe into the U.S. hours, and we were able to leverage AI to look at the market intelligence know what kind of direction certain sectors are experiencing volumes, volatility, spread volatility, all of the market intelligence can be derived leveraging AI sitting on top of our -- quite broad source data. The second area of exploration where we have dabbled already is what we call portfolio optimization. Given the market intelligence that we have, given the levels of liquidity that we see in the market, AI is a wonderful tool in interpreting selection of underlying bonds when building a portfolio. And then the last area of exploration is really -- and one that we hear from our clients the most is leveraging AI to help us suggest protocols. There are times when a portfolio trade is an optimal way to trade a list of bonds. There's times when going direct to a single dealer for a large block is the right protocol for that bond. So using AI to actually suggest protocol selection is a key ingredient to some of the areas of exploration. Now the other area of AI, I think, which is store exciting for us, it's certainly in the way we're leveraging AI in our technology footprint. If you look at MarketAxess for the first 20 years of our history, there's probably a large underinvestment in modern tech. We have been, over the last number of years, making sizable tech modernization investment and we've been doing that both organically and inorganically. If you recall, our acquisition of Pragma was really a tech modernization acquisition. We are now leveraging that technology and things like auctions and and our automation suite. What AI is doing for us today is what's most exciting. We're now accelerating that tech modernization, including our core trading stack. We're currently actively engaged in AI solutions that can look at refactoring our legacy code. This is an exciting component that a number of people have deployed in the market space that we're in, where they can refactor legacy code. We're also seeing AI can increase our time to market on new capabilities and new functionality. So we're expanding our use of AI across our engineering footprint. All our developers have access to all the latest and top models and that is slowly having an impact on how fast we move. As I mentioned in my opening remarks, we brought in our CTO, Will Quan, he has both cloud and AI experience. So that alone has accelerated our deployment of AI, and it's quite exciting. UI design is another area where we see AI development accelerating what clients can have in front of their desktop and how quickly we can turn around those UI designs. The last point on AI, I'll make, and it's really around the M&A space that I foresee. I think AI is going to change the M&A landscape. Most of my career was -- I was involved in M&A that was -- you typically would see sizable tax synergies in M&A and some M&A was really designed either. I either had acquisitions that we were doing that were tech accretive, where technology synergies were one of the drivers of the M&A or I had companies that were acquired because of tax synergies where, again, technology synergies were a big component in the synergy analysis of the acquisition. AI is effectively reducing the value of tech synergies when it comes to M&A. AI has the ability to, in a faster way, refactor, older technology. So it's really changing the math of deals when you actually look at what AI is capable of making some deals less attractive because they're less synergy accretive. And that's an important thing. The other important thing is really to recognize and think about the MarketAxess over 25 years of sales effort and network effect. AI does not sell and distribute services. AI can't build client networks, it can't do KYC onboarding, but they can do KYC work, but it does not onboard thousands of clients. It doesn't take clients to dinner, and it doesn't make source data. So what really becomes valuable in this world of AI when it comes to M&A is companies that have golden source data and companies that have broad networks. And I do think AI, not only is it changing our product set, but it's also changing how fast we can deliver product. And ultimately, it changes the value of the data that we sit on and are mining today. So sorry for the long-winded answer, but it's really reflective of the excitement that we have and the advantages that AI give us.