Kirsten Rastrick: Hi, my name's Kirsten Rastrick, and I'm the Global Head of Fixed Income Financing Sales at JP Morgan.

John Schwartz: Hi, my name is John Schwartz and I'm the Global Head of Sovereign Financing Trading at JP Morgan.

Kirsten Rastrick: So this is such a treat to be able to sit with you and have a quick conversation away from the trading floor. We sit at an interesting intersection of a lot of different conversations happening in the market related to financing and repo, specifically at the largest bank, a security services custody platform offering large corporate treasury securities lender and market making division. You're having conversations with industry working groups, regulators, your peers. We're both having conversations with clients. So we thought this would be an interesting time to just give a little status update on the repo market and hone in on the fixed income space specifically in the US treasury market.

John Schwartz: Yeah, for sure. I mean, there's been a lot of talk recently just around the size and scale of the US Treasury market. I think it's always helpful when you think about the financing space is really get a sense of where we've come from. So if you think about the US Treasury asset class specifically and some of the just meaningful growth that we've seen over the years, going back 20 years ago, outstanding US treasury debt was approximately $5 trillion. That has grown from five up to 30 today. So the growth of the market has clearly led to a lot of discussions around, well, what does that mean for internal mediation? What does that mean for treasury market liquidity? Not only have we seen growth in outright issuance, we've seen really structural changes in the composition of ownership. So if you go back a decade, you had large foreign buyers, both private and public sector.
Clearly there was participation in the bank space, there was participation in the levered fund community, but that composition has evolved over the last decade to where more price sensitive buyers have become a larger and more meaningful component of the US treasury market. The complicating factor of that is hedge funds and other price sensitive buyers require not only dealer intermediation in terms of buying and selling bonds and providing market liquidity, but also financing intermediation. So with that change of composition coupled with the growth in the US treasury market, you have seen growth in reliance and growth in dealer balance sheets being committed to fundamentally supporting that intermediation. The repo market by proxy has grown probably two or three times in size as close to $6 trillion just in the US treasury product today. Hedge fund participation in repo has grown from 500 billion back in 2013 up to close to 3 trillion today.
So that increase in dealer balance sheet pressure has clearly led to, well, how do we continue to intermediate? How do we continue to provide liquidity to the US Treasury product, not just in terms of the outstandings that we're really sitting with today, but no secret that the budget deficit is large, it's growing. So when we think about the $30 trillion figure today, that is expected to grow approximately 2 trillion a year for the next number of years. So the magnitude only increases. I mean, in terms of client participation and what you're hearing from hedge funds, how do you think clients themselves are thinking about allocated capacity? How do you think clients themselves are thinking about their position in the market?

Kirsten Rastrick: I think I agree with you on all those points, and you've really started to see much more reliance from a variety of different clients on the repo market for leverage in this space. So hedge funds in particular, like you said, have almost quintupled over the past decade in their utilization of the market. But that's not to say the other accounts haven't also gotten more familiarized in utilizing the repo market. And this spans not just across the US treasury space, but across EuroGovies, other non-dollar sovereigns emerging markets credit and securitized product financing as well. The hedge fund space specifically has been a bit of a, it's helped fill a void in terms of absorbing some of this supply that I think dealer balance sheets have been a little bit more pinched to be able to absorb with not only increased issuance, but also the increased regulation on the banking sector.
There has been a need for additional participants to take down some of the supply that's coming from the market. So I think the big multi-strat hedge funds have just gotten bigger. The systematic funds have started to emerge with a fixed income strategy, both directional and relative value strategies. You've seen different uses for expression of interest rate views from the real money community. Not only is cash government bonds a viable opportunity to express interest rate risk, but also they're looking at maybe more cash-efficient opportunities in futures or derivatives space. So you are seeing a slight drift into a diversification into other asset classes for expressing that same risk from the real money community, all the more reason for the need for another buyer to step into the US Treasury space. So all of those things combined show that support the fact that this market has grown in not only its size, but it's also its significance.
Now from there, what does that mean in terms of binding constraints and optimization of those binding constraints? The dealers have had over the past five to 10 years a whole slew of new regulations imposed on it and regulations, looking at balance sheet, looking at capital, looking at liquidity. But the growth in at least the hedge fund space as well is also limited by a credit component as well. So there's credit risk limitations that we've started to see. So I think to your original question of how hedge funds and the client franchise are thinking about utilizing the repo market, one really key thing for them to be able to understand is how their different dealer counterparties are managing to the different binding constraints. So what we have seen recently is not only a desire from clients to understand from us where and at what different point in time different binding constraints come into play, G-SIB, RWA leverage in the form of balance sheet liquidity, and then therefore how they can optimize to really best suit flows that fit neatly into all of our appetites across those different metrics.
And from a client's perspective, it's not just the JP Morgan footprint. They're going to have to do this for every single dealer that they're speaking with. So it is it a bit of a project for the clients to understand this, and it's also in an evolving landscape. Everything is actually highly in flux right now with regulatory change and market structure change as well. So maybe that's an interesting segue into some of the work that you've been doing, either with advocacy or understanding of the evolving market structure changes that are in flux.

John Schwartz: I mean, it's certainly challenging when you think about it, whether it's from the dealer side or from the client side. There is some asymmetry there in regulation, whether it's gold plating of US regulations, all adhesive and SLR, whether it's just different governing bodies governing different types of intermediaries. I think the challenge is how do you navigate your trading strategies? How do you navigate where you lean in terms of clients to really optimize your own capital base and your return base? So as a trader, you think you spend a lot of time focusing on risk or market positioning, market views, and I certainly spend a lot of time doing that. But more and more these days is how do you structurally build the appropriate, as you said, a capital liquidity framework to the extent we think there are deficiencies or gaps in how regulations have been brought up, how do you advocate?
How do you align, yes, we think prudent risk management aligned with treasury market liquidity to me is an important thing? So in terms of what we're advocating for, there's a big industry push today to potentially reassess how SLR is calculated and viewed. Certainly if you go back to periods of volatility and stress, certainly during COVID, LCR became a meaningful binding constraint, particularly in treasury intermediation. That is less so today, but I do think there's risk where markets do dislocate if dealers are asked to step in and provide liquidity capacity and liquidity provisioning that that may become binding. So a lot of advocacy work focused on US Treasury carve-outs, focused on fed reserve carve-outs. Certainly we're very focused on G-SIB, not just alignment of G-SIB across jurisdictions so that US banks can compete on the same, really a level playing field to our US or to European peers, but also recalibration at G-SIB.
The fact that the economy has grown approximately 55% since when G-SIB was put in place, G-SIB has not been tailored to account for that economic growth. So I think there are thoughtful ways that both the regulators and certainly the community can approach these things. They don't necessarily introduce or increase systemic risk, more adjust and tweak the capital rules that would support intermediation. I think another thing that clearly there's been a lot of talk on post-CSEC's finalization of the US clearing rule is market structure and market structure change. If you think about the increased leverage needs and liquidity provisioning to hedge funds, unlocking dealer intermediation capacity, how we do that from a market structure standpoint is going to be super important. You've seen natural growth in the clearing of US treasuries over the last decade, and that has been willing clearing, not necessarily regulatory driven, and the incentive function is to manage G-SIB and it is to manage balance sheet for us to be able to do more with clients and service their needs.
That now will go into overdrive as we have to think about clearing the entire US treasury market, both from a repo perspective in a much larger component of the cash market. So how that evolves and how that necessitates change in how we interface with our clients, I think is going to be very important to not just our ability to continue to facilitate client activity, but also impacts the treasury pricing, swap spreads things that ultimately the regulators keep very close eyes on and impact the taxpayer at the end of the day. But with that comes opportunity said. When I think about stronger connectivity to the CCP, you naturally then dovetail into, well, what else can we do that helps streamline the way that we interface with clients? So electronification, that's been a big agenda. So maybe you can spend a little bit of time talking through what you've seen over the last five or six years on that electronification front, the introduction to D two C type connectivity, our own effort to build in-house solutions, I think that would be is clearly a natural additive to that.

Kirsten Rastrick: And I think it just is a natural solution for the growth in not only the size of the market, but therefore the client flow that's come off the back of that. There's been an absolute need for automation and digital solutions to handle all of this increased volume that's going through the system. So there's a multitude of third-party vendor platforms that have been rapidly adopted by the client franchise, and this spans all types of clients. It's not just your systematic clients. This is the entirety of the hedge fund community. It is asset managers, it's central banks, it's real money pension insurance type names that have adopted these dealer-to-client trading vendor platforms, and even our own JP Morgan Execute platform as well. What has that done that has enabled them to more efficiently trade? Not only just the high volume, high number of ticket treasuries and EuroGovie type trades, but it spans across all asset classes.
So some of the emerging market and credit flow is very ticket intensive and it's really helped to simplify some of the financing in those other asset classes as well, so that is a frontier that it is still emerging in terms of its development, but the volumes going through on these vendor platforms are very significant. I think it is going to be also a necessity at the time when treasury clearing comes around for implementation, there will need to be solutions for pre-trade credit checks, et cetera. So that's all going to have to have a venue or a forum for electronic trading capabilities. So it also speeds up execution. So when you're thinking about just the day in the life of a client or a salesperson or a trader, one of the other binding constraints that we didn't talk about in the previous segment, daylight overdraft costs. And so to have an automated system that enables you to trade your entire T plus zero same day trading book in the first hour of the day is very impactful from a cost saving perspective.
Every basis point, every dollar counts in the funding markets. And so these clients are looking to just really optimize and create efficiencies wherever they possibly can. So that's been a notable shift as well where you've seen such a move forward in the timeline of a day to day a client-sales-trader interaction by utilizing these automation tools. Now, with those automation tools can come a whole host of benefits in addition to some cost savings, but it can also create a repository for data to be analyzed. We are getting much more efficient at tracking metrics associated with hit-miss ratios, trends of client behavior across all of these different asset classes. So the data element is going to be quite impactful as we get more and more flows through the vendor platforms and able to be warehoused in a repository. So that I think is another value and it is going to really inform us to help us be informed in how to make better decisions for our binding constraints, the capital liquidity balance sheet metrics.
So a massive lift to the industry by the evolution of these platforms, and we're super excited about it. We continue to think of new innovative ways to contact and to engage with our clients over the platforms. The other thing that I think in terms of, we didn't necessarily go down this path before, but in terms of other ways of optimizing our client flows and behavior, the market is migrating a little bit towards a cross-product margining offering as well. There's dealers out there who have it. We are also in the final stages of a cross-product margining offering for clients as well. And so I do think that will just create even more efficiencies for the client franchise in the coming years.
But between optimization tools, digital and automation solutions for our clients, what that does is I think it really frees us up to focus on the more bespoke long-term structured financing solutions and mirror all of the benefits that we have gotten from some of these more highly developed high-flow markets and developed government bond space into some of the other asset classes as well, which is a boost to opportunity sets in those markets as well for credit, mortgage, emerging markets with more time and effort and tools available to support those markets, it will actually fuel liquidity and general market functioning in those markets as well.

John Schwartz: I mean, you bring up cross-product margining, and I think it's interesting because in my mind it brings together all these things that we've been talking about. On its face, it's really there to unlock capacity and to better align risk within a trading strategy. So you use the treasury cash future basis trades as an example where if you go back to COVID and you had one leg of that trade clearing through the CME and through an FCM, you had the repo leg clearing through a different provider. There wasn't a risk view to the actual trading strategies. So ultimately when margins were increased on the CME side, it led to a massive liquidation.
And I think when we think about appropriate market structures that better align trading strategies, in my mind, cross-product margining is a really good example of that, where if you bring things together in a thoughtful way, not only does it increase capacity and client's capacity to trade that strategy, it fundamentally reduces, in my mind risk that we wear associated with that because we're not sitting necessarily with only one side of that transaction. And it's ultimately, I think it promotes just broader trading liquidity. You don't have to have as much capacity laying around in reserve. You don't have to think about multiples of liquidity buffers if you know that these really tied together trading strategies are going to behave and be margined in a thoughtful and constructive way. So I think it's very important as this market evolves, as more and more of this goes into central clearing, that the market embraces ideals like this.
Now, I do think, not to wrap it back around to advocacy points, but aligning how we think about capitalization of different products in a single netting set goes back to I think a number of the points that we are actively engaged with, both through industry forms like SIFMA and the regulators. Cross-product margining. You can package it together. You can have the risk infrastructure there to support it. You can have clients willing to participate in it, but if the way that we need to capitalize our business isn't aligned with that, then it quickly becomes a very non-commercial offering that I think we can offer. But I do think that that idea really brings together a number of points that we've been talking about.

Kirsten Rastrick: And I guess to wrap it all together in some final observations, the repo market has been extraordinarily resilient over the past couple decades that you and I have been in the business. And I think some of these regulations have obviously been to pinpoint and target maybe some areas that needed a little bit of additional risk oversight. So whether it's leverage rules, whether it is additional capitalization for the banks, whether it is tiering of different sized banks given their systemically important significance to the market, all of these regulations have proven in the past couple crises that we've been through, whether it's COVID, whether it's the regional banking crisis that the market is holding up, it's very, very orderly. And so I think to the extent that there is, we can continue to innovate with tools to just optimize and create efficiencies as much as possible. It will continue to fuel growth in an orderly way in this market and get more creative and innovative with our solutions-based financing on top of the facilitation of client flow and the traditional flow products that we see heavy volumes in every day.

Patrick Whelan: Hi, I'm Patrick Whelan, the global head of Fixed Income Digital Markets at JP Morgan. I'm delighted to be joined today by Andrew Cole, who is the platform product manager for JP Morgan Markets Execute. Thanks for joining me today, Andrew. Maybe you can start us off by telling us a little bit about your career journey through JP Morgan, how you ended up today as we like to call it the chief product officer for Execute.

Andrew Cole: So I joined JP Morgan 2010 in technology. I did 18 months in a support role, most clamorously delivering hard tokens to clients in the FX business. I spent 18 months in credit risk and then I got hired into what was then called the eCommerce team, working as a subsidiary on the trading desk, helping with price distribution. I started to get involved in the Execute platform. I covered our FX Algos business, moved to New York, and then over time grew the role to being overseeing all of our Execute products as it pertains to the clients, internal traders and salespeople at JP.

Patrick Whelan: What's it been like? What's the piece you're most proud of?

Andrew Cole: That's a good question. I think the evolution has been really exciting. I think when I look across the products that we now have and the way in which we attacked growing the platform or tried to grow the platform, we have learned from some of the things that did and didn't work. Obviously, the engine room of the product is the FX franchise where we have thousands of users, but building something from nothing to that is obviously challenging. And I think the thing I could say I'm most proud of is showing or is figuring out the different hacks to making a business grow on the platform. So if you take something like commodities where we have a huge range of different products and a massive, therefore range of different workflows, we've had to figure out how to scale a technology solution to be very flexible.

The commodities tickets on the platform cover milk, metals, gold, everything in between. So building tickets and infrastructure that supports such complexity and variation is hard and is ultimately what's enabled us to scale that business. If you look at something like FX Options where there is a significant amount of complexity to the tickets themselves, figuring out how you scale that is again a challenge. And then onto the distribution side, as well as figuring out a technique for providing technology solutions that allow very flexible tickets and very flexible workflows, you also have to figure out a fairly flexible way in which you manage to appeal to different client segments and to distribute your products. So I think the thing that I'd say I'm most proud of is Execute is a single bank platform for providing execution to sales, internal trading and directly to clients. But the ways in which we do that are really manifestly different depending on the client segment you're going after or the product you're trying to bring them onto.

Patrick Whelan: Look, you've talked a lot there about building scale into solutions that ultimately are developed by a reasonably small number of people. And obviously it's a strong collaboration between our technology teams, our product team, our design teams, and everybody else in between to help develop these solutions. But I think when you talk about the growth and actually getting more clients to use single dealer offering, I think it's interesting to see how the solutions you've built for FX and commodities and how they've been adopted by credit and rates and some of the product and asset classes, they're not necessarily complete in their electronic journey.

I suppose from your perspective, and we've talked about this a bunch of times in the past around platform growth. And if you're in an industry or in a company that is predominantly like a FinTech that's out there today, their entire journey is based on users using the platform, capturing revenues via fees or other commercial models. Obviously for single dealer platform, it's a bit different, but how do you look at the kind of growth profile? How do you set objectives for your team in general? How do you make sure that clients want to come on board?

Andrew Cole: So I think one of the things you said there around the differences between the LOBs, in our management, we often talk about the concept of a platform and building something that as a platform can scale across the pieces within it. And one of the things that's been quite satisfying about taking a role like this where we have a broad range of asset classes that's continually growing is identifying those commonalities between the different user journeys and trying to find ways to scale out and abstract some of those pathways that users are going along so that something you build in FX can scale into rates credit. I think in the beginning of XQ, and again, one of the things that we've started to do is trying to break down the line of business silos. So where you have rates, credit FX, we used to build separate sets of components for each.

We're increasingly trying to combine them. And when you ask about objectives for the team and how we manage that, it's hard because if you're someone that lives and breathes in credit or rates, it's hard for you to appreciate that the journeys that your users are going on and the things that they need are similar between LOBs. You like to think there's a uniqueness to what your users want.

Patrick Whelan: Everyone's special.

Andrew Cole: Yeah, exactly. Some specificities to your product. So when we're setting objectives with the team, we sort of divide ourselves between a core platform group that build these abstracted components that every LOB can leverage, and then you have your line of business specialists and setting objectives so that those two groups can partner and work well together. It's been hard because it is a different model to what we've had in the past. So I think that's definitely something where we've been focused.

Patrick Whelan: I suppose when part of that journey is client's desire for customization in terms of what they want and ultimately we're generally fueled by ideas from their side and how we can actually solve problems from a client's perspective. How do you balance the customization versus what you spoke about about the need for common components across each of these areas?

Andrew Cole: Yeah, it's a great question. And I think it's the main dissenting voice when we started on this journey was the fact that, oh, but what about user ABC wants to customize? One of the ways that we attract users is through offering those custom solutions. Lots of the customization that we do is in the backend and I think a lot of the success of the platform is working tightly with design to figure out how a user can feel that the user journey they're on is customized to them, but doesn't actually involve doing things very differently. Customization is mostly about ease of use. And if you work well with your design function, your user research function, which is what we try to do, you tend to find that you can build solutions that scale even if the user journey that the user is on isn't so different.

So I think for that, yeah, mostly our customization process involves AB testing. We make a lot of use of the fact that sales and trading internally use the same platform as our clients. So by getting them onto a prototype that our design team have built, AB testing that design, prototyping it, that's how we can inform the way in which we build out our tickets. And in so doing, try to avoid a client coming in and saying, "This doesn't work for me. I need that." We just try to make it so that it should scale out. But it is a challenge. And if a big name comes in and asks for something, one of the powers of what we've done with the way we've built the platform is it should be flexible enough that we can accommodate those requests when they arrive.

Patrick Whelan: Look, I'm fascinated by great design and how simple something can be when designed properly and how immediately user-friendly it can be and how quick it can be adopted, but it takes a lot of effort to make great designs into reality. And I think, as you said, that partnership between design, UX, user research and our engineering teams is critical to some of that, to those functions coming out the way they are and being adopted so quickly by our clients. I think as well, like you touched upon it as well, there's benefit from the fact that some of these components are so reusable and that we can quickly go into a new asset class and start streaming prices and grids to clients quickly.

Or when it comes to analytics, being able to design visual representations that can be used again across asset classes for clients to be able to quickly visualize some of the trends they're seeing, whether that's the net flows that we see across certain areas where we have aggregated a bunch of our flows internally and are able to display those to clients to see what is happening across the market more broadly, obviously anonymized and aggregated, as well as in other areas. And I suppose when it comes to some of those components and some of those design features, how do you and the team look at it and how do you actually build those solutions out for clients? How much of the client journey is theirs and how much of it is ours?

Andrew Cole: I think analytics is another area I think we can say we're particularly proud of. Sales and clients started to talk about the importance of data and distributed data. And when you look at the way that JPMorgan's distributed that data, I think we've done a really excellent job of building out products like data query and some of the other products, Fusion within the JPMM family. But I think one of the areas that we had historically been less focused was on trading analytics and point of time execution analytics. And I think that all started for us with the Algos realtime TCA. So we built out those products. And to your question in terms of the workflow, it became quite seamless because as the client was placing the algo, the analytic would pop up and they'd have it available to them in the tool. I think that as we've grown from there, we added net flows, we added market structure, commentary, et cetera.

And now what we're trying to provide to a trader or to a user is all of the data that they could hope to get from their sales coverage in one place in an interactive format where they can tweak it, customize it, select what they want more of, less of, et cetera. And the magic of what we've done there is we obviously have our single bank platform, which you can download. We also have some of those analytics creeping onto our mobile device, but the app that we offer is obviously the secret sauce there because clients can log into that app, it's embedded in the terminals where they already trade. So it becomes more seamlessly part of their workflow. It's in night mode, so it looks like the rest of the platform around it.

We've tried to just build out from there. And the good thing about analytics is that we increasingly see it as a gateway to execution. So a lot of the clients that log in to monitor our axes or see our streaming data will then say, "Oh, can I execute on that data?" And that's a great conversation to be having because that's exactly what we want to be doing, is providing them data that triggers a decision to trade or a realization that there's something there that they hadn't noticed and therefore they want to trade on it.

Patrick Whelan: It was super interesting during the recent boat of volatility during Liberation Day and the subsequent days afterwards to see our clients embrace the analytics, especially around those periods. Obviously the execution days, we had record volumes as well. And so again, being able to see the platform stand up to the heaviest barrage of client users as well as the volumes traded was great to see. But equally to see them embracing things that we had developed with that in mind and being able to say, okay, what really matters real time is different to what potentially you want to look at over a time series in a historical basis and that you want to be able to do some research on when it comes to, as you said, execution, you want to be as up-to-date as you possibly can.

And so to see the uplift in users around that time was certainly impressive. And I think that's just given us more fuel to go and build out more solutions that I think will drive more of that. It also helps us with, we obviously have a single dealer platform, but we also provide API connectivity to the client bases that don't necessarily want to use our single dealer platform for different reasons or have built their own and want to aggregate the information directly in their own EMS or OMS. And obviously it makes sense for them to connect with us directly via API in those instances.

But again, the common componentry behind the scenes make it easier for us to stream that market data to give them the relevant analytics differentiated where they need it at the point of trade, point of execution when they're getting to that point when the execution really matters. So again, the solutions are common. The work behind the scenes is about making sure that they are seamless and ultimately are where clients need them when they need them. So a lot more to go on that front, but I suppose from your perspective in terms of the platform and where it is right now, what's next?

Andrew Cole: So I think just to take back to one of the things you said, like a daylight volatility, the volatility [inaudible 00:13:48] is great because it's a reward for all the effort we put in to scale the platform, to make it cope with that increased load. It then sets a new high bar in terms of what we have to achieve the next time. So it's a massive kudos to all of the technology teams are involved from the backend infrastructure right through to the front end that they work so hard to make things like that work. Because when you have crashes or when you have stability issues, not that we have so many of them anymore, we are aware of them and we were obviously always stress testing the platform when it all comes together in a period of volatility like that is obviously hugely rewarding. So I'm very grateful to the teams that work to make that possible.

I think as we look forward, look, we are always looking for new products, new areas to get into. Looking forward, I definitely see more complicated products as an area of opportunity, both because clients, as we see it, value the ability to input the product correctly and accurately, significantly more than anything else when it comes to that sort of execution. So when we look at the electronification of FX Options, we see that in the third generation exotics, that the most complicated products, the percentage done electronically is far higher because the client likes the comfort that they get from entering themselves, seeing the parameters that they want to select, not going through in a chat and explaining in mega detail how that product needs to be structured for them. And as we get into list products and derivatives in any asset class, we have a new range of tools from the technology side at JP Morgan that makes scaling and distributing those products more easy. And we see the demand from the client to be able to get their hands on a solution that allows them to communicate it accurately.

So I think the future for us definitely looks like a derivatives led offering will be something that will definitely resonate with our user base. We also think as the product complexity increases, the risk increases. And we've seen recently with some of the technology vendors in this space that they struggle to manage the product complexity well, they struggle to synergize the different ways in which different banks have managed to communicate and distribute these products. And also there's the risk element. When you're communicating a product between a client and a bank as a middle agent and there's 30, 50 plus parameters that you need to configure, the chances that something goes wrong is heightened. And so for us being able to say, "This is our platform, we have the ability to flexibly add the fields to customize the ticket to your requirements." That gives us an edge and we believe that as electronification moves from the simpler to the more complex products, that edge will do us well. So I think for me, the future is investing in that area.

Patrick Whelan: I mean, you've touched upon it there in terms of like, it's one part of it is building it for clients and building a solution that obviously works for them. I suppose the other part of our day job is building solutions that work for sales and trading. And ultimately our first clients in a lot of instances are our sales partners and they have to book all forms of products. And so it isn't necessarily ones that have been through the electronic journey, but the reality is we do have to create solutions for them as well. I suppose in terms of the offerings and the differences that come about as a result of that, how do you build solutions both for sales and for clients in one platform and make sure that it's seamless both for our internal and external users?

Andrew Cole: It's one of the things that whilst it's very hard to do, it's really important that we get it right. We talked a bit before about the value we get from internal users, AB testing. We massively value ourselves and trading partners for being the guinea pigs on our tools. We can proudly go out to clients and tell them that this is a tool our internal users are using. When I worked just on the Algos suite, the fact that our Algos are used both by our internal traders and our clients means that all of the feedback comes back to us from the internal desks and it also lends some credibility to the product that you're offering that is good enough for our internal users to be using. So whilst it's hard and we can come onto how we manage it, the reward of it is that you can AB test it, you do find risk issues quicker, you can collect more readily available feedback.

So we do massively value that partnership. In terms of how we do it, the underlying workflows tend to be fairly similar. At the end of the day, if you're a salesperson booking on behalf of a client, you're effectively entering in a lot of the same fields that a client would be. There's obviously a lot of regulatory and credit checks, et cetera, that are done by the salesperson, and so we need to build all of that functionality in, but providing a common entry point and a common set of components is, that's part of the journey of building in a customizable and design led way.

Patrick Whelan: We had a lot of success in areas like fixed income financing in the repo space with that, I suppose. It's been great to see that journey for both sales and for clients happening almost in tandem.

Andrew Cole: Yeah, exactly.

Patrick Whelan: It's getting it out in front of our sales partners, getting them comfortable with the workflows, understanding that there will be issues. And we'd rather discover those issues internally first, fix them, rapidly deploy the bug fixes so that we can actually get the solution to where we need it to. But at the same time, bring clients on that journey with us so that you're not necessarily just building it for sales. You're also building it with clients in mind and you're able to have that conversation both with sales and clients at the same time so that when we are having those design discussions and looking at what the solutions need to be able to handle, then you're able to bring in both sides of the client and the user journey into it.

Andrew Cole: And just to add to that, the idea that our salespeople are out selling the product that they're using, it also all feeds back in, because if it's-

Patrick Whelan: Yeah, exactly. Exactly. They know what they're talking.

Andrew Cole: Yeah, yeah, exactly.

Patrick Whelan: And so we have a lot of salespeople which is good, but at the same time trying to roll it out all over the world is also, it's a scaling problem at times. But I think we're doing a lot better at that too, but it's a continuous work in progress, I think, on that front.

Andrew Cole: So a question for you now, SPG has been something new for us. It's a business that we've not looked at on Execute before and things are starting to click there and go pretty well. Tell me a little bit about how that all got started and how the team came together to make that possible.

Patrick Whelan: Absolutely. I mean, you've touched upon it already in terms of some of the more complex products looking to provide single dealer solutions to their client base, and it is a different client base. In a lot of cases, I think when we looked at Execute in FX and commodities, there was a lot of banks, hedge funds and corporates that were the predominant client base that we focused on and that were early adopters of the platform. SPG is a different market altogether. And I think the clients that have been part of that journey with us have been looking for solutions that ultimately a single dealer platform can provide better than potentially even some of the multi-dealer platforms. Some of it is because of the way the market operates today and BWIC and OWIC, which are bids wanted in competition, offers wanted in competition, is a feature of voice trading that has been around for decades.

And I think everybody was looking for the efficiency that the BWIC, OWIC offerings can bring by aggregating some of that information on a single dealer platform, providing our liquidity alongside the liquidity of our clients on the other side, gives the client who's been who was asking for the BWIC or the OWC in the first interest, the reach of our vast client network, but it equally gives them the efficiency of a digital channel to be able to handle that, which ultimately from an SDP perspective and for some of the other pieces is part of the efficiencies that they're looking to get from a workflow perspective. But equally, they get the pricing and liquidity access to the broad reach of JP Morgan, which is always helpful. And it's just the beginning. So parts of SPG and the TBA market and spec pools have been part of the electronic journey in some cases, but mortgages and some of the broader areas across SPG definitely haven't necessarily gone down that road just yet, but they're certainly headed in some ways towards that.

And providing a single dealer solution, considering the size and scale of our market share in that, as well as our offering in terms of pricing and liquidity and the depth of our risk management within that area, it's always going to be a compelling offering for clients. And I think that's where we hope we can continue to grow. And obviously it's new clients to execute. It provides the opportunity for new analytics, for new offerings to be put in front of a new client base. And ultimately, the clients in SPG, they're also clients in munis and elsewhere in loans and CLOs. And so we do see more synergies across some of those other areas as we grow the footprint in SPG. So again, it's not unique to the broader discussion that you are having in terms of how we grow the platform more broadly, but I do think it is unique in that that client base is in some ways different to the ones that we've traditionally embraced.

Edwina Lowe (00:12) Hi, I'm Edwina Lowe, product specialist within the Data Assets and Alpha Group.

Eloise Goulder (00:18) And hi, I'm Eloise Goulder, head of the Data Assets and Alpha Group, which sits within the prime financial services organization within the trading business here at JP Morgan.

Edwina Lowe (00:27) So Eloise, um, could we go back a few steps and talk about the creation and reason for our team, the Data Assets and Alpha Group?

Eloise Goulder (00:35) Absolutely. So we created the group six years ago, it was in 2019, and really we created it to try and pull together two significant functions within the trading organization. So one of them was the market intelligence function. This is a function that has been around for more than a decade. It predates my life at JP Morgan, and really the market intelligence function is providing on the pulse market dynamics and interpretation of what is going on. Based on political developments, based on macro catalysts, based on results season and what the thematic drivers of the, of results season might be, that is what our market intelligence function does. But in addition to that, and we really wanted to beef up that function and make it a true coordinated global function when we set up the group, we also wanted to add in the data-driven lens. That's so important, partly because our clients demand for data, uh, and data-driven insights has only been increasing over that period, but it's also because it's so important that our markets insights are as unbiased as we can make them, are very tactical, our, um, are very catalyst driven.

Eloise Goulder (01:52) That's so important for our market intelligence function. And it really goes both ways. Our market in- intelligence function is more powerful with that data-driven lens and the other parts of our team, our data intelligence team and our positioning intelligence team, which together make up our overall group, are so much more powerful if they have a market's nous, if they are topical, if we can bring to life those data sets with relevant market insights.

Edwina Lowe (02:18) Absolutely. And where do you see the intersection of those three distinct pillars of the team coming together?

Eloise Goulder (02:25) Well, I do like to think the sum is so much more significant than each one of those parts, and there are so many areas that intersect those three pillars. I mean, Edwina, your role really spans all of those pillars. You're involved in so many cross-group projects. Perhaps I can turn that question back to you.

Edwina Lowe (02:44) Yeah. I mean, I think from my perspective, the opportunity to speak to clients cross-team, so pulling together different parts of our team for our client conversations can be really powerful and certainly appreciated by our clients. But in terms of our products, I think I would single out the podcast as a great example of that team collaboration and bringing the threads of our team together. I mean, we launched our series three years ago, and I think it's a fantastic example of showcasing everything we do as a team, but also the opportunity to bring together other colleagues from the wider markets business. So for example, one of our recurring themes, as you well know, is the retail investor and the rise of the retail investor. So the opportunity to look at that through a number of different lenses and bring in colleagues that can also speak to that theme.

Eloise Goulder (03:34) And it's interesting, we, I think at the outset of the podcast, we just saw it as another means of communication, another means of content dissemination. We have the written content from our market and our positioning intelligence teams. We have the data content, the API content, which of course is an enormous growth area, which I'm sure we'll talk about today. And the podcast, another audio means of disseminating content, which of course has been growing across the world and across our industry. But what I've learned, and I'm sure you feel the same, Edwina, is actually, it's a way of creating new ideas, new insights, which you couldn't create if you didn't put those subject matter experts in a room together to have that discussion. Um, in that so many topics don't have a firm conclusion. It's not as simple as question, three bullets and a conclusion. So many questions are so nuanced.

Eloise Goulder (04:30) You know, we're tackling should, uh, will US exceptionalism continue, will it last, or will other economies start to outperform, and will other market assets start to outperform? We're covering the rise of the retail investor in so many markets, in US markets, in Asian markets, and what the implications of that are. We're covering behavioral biases and the extent to which we can both try to mitigate them in investing decisions, but also how we can capitalize on them to create, to, to attempt to create alpha generating opportunities. These are such complex topics, and it's n- and most of them don't have a clear cut answer. It would be naive for us to assume that there is a, a clear cut answer. So I do think the power of the podcast, the power of the audio conversation is so significant because you can really, you can learn so much, but also you can share and articulate the nuance of the argument.

Edwina Lowe (05:25) And really the beauty is in that nuance. Could we focus on our team name, the Data Assets and Alpha Group? I've been in the team a little over three years, but I'm aware that the team predates me. Could you talk a little bit about that, Eloise?

Eloise Goulder (05:37) Yeah, absolutely. Well, in a way, I think we do exactly what we say on the tin. We are the Data Assets and Alpha Group. What do I mean by that? Well, we are creating and curating data sets from across our market's trading organization. They may be positioning related, uh, data sets, hedge fund positioning, they might be social sentiment related data sets. What are our interpretations of sentiment based on scanning the multitudes of posts on social media? Uh, so we're creating all of those datasets, but why are we creating them? Why are we, why are we in the pursuit of the creation of those datasets? It's because we're using them for predictive purposes within markets. We're using them to help have a lens on which markets and which assets and which stocks will outperform or underperform, and therein lies the alpha term. It's all about alpha generation at the end of the day.

Eloise Goulder (06:32) So we are the Data Assets and the Alpha Team, and we wouldn't want to care about the data assets without that alpha focus. Alpha is the core of everything that we do.

Edwina Lowe (06:42) So Eloise, a question that comes up a lot is what distinguishes us in the Data Assets and Alpha Group from the research organization. Could you talk a little bit about that?

Eloise Goulder (06:52) Yes, absolutely. And we do get that question quite a lot. So we are distinct from research. We have a very strong relationship with research. That is critical. It's so important we understand what research are doing and saying, whether that's from the macro lens or the micro bottom up stock picking lens or indeed the quant lens, but we're distinct. We sit within the trading organization here at JP Morgan. And why are we different? Well, first of all, the very data assets that we create and that we curate, they come from the trading organization. They are very separate from research. Those data assets include hedge fund positioning from our prime business where we sit. For example, they include all of our analysis around social media, social media sentiment as a lens on the retail investor, which has become so significant. Those data assets, they sit within the trading organization and therefore we sit within the trading organization.

Eloise Goulder (07:51) Second of all, this very tactical intraday activity, these insights, that very often sits within the trading organization across the industry, and indeed it sits within the trading organization here.

Edwina Lowe (08:03) Yeah, that's very clear. Thank you. And actually, I'd like to pick up on that point of the data sets. You know, I know as a team, we are responsible for creating data sets, these predictive trading datasets and insights, but could you expand a bit about what that actually means and looks like in practice?

Eloise Goulder (08:19) Definitely. And before going into detail on perhaps one or two of those datasets, it's worth noting that in creating these datasets and in providing insights on those datasets, we're striving to really open the box, open the lid on our analysis so that we're not simply presenting opinions and views, but we are actually fully open source in a sense. If we have a view on a market or a stock or an asset, uh, yes, you can listen to the opinion, you can listen to our podcast, you can read our written content on that opinion, but you can also download the dataset behind that opinion via an API, do the analysis yourself, stress test that analysis yourself, and importantly, critically utilize that in the future so you don't have to speak to us every minute of every day, uh, but you can put that dataset in with all of the other datasets you source in order to see how that view will materialize through time.

Eloise Goulder (09:13) And I think that's so critical. That's why we are creating data-driven insights and we are both making the insights themselves available, but also the underlying datasets. It's all about opening up, being transparent, being as unbiased as possible, we're showing our workings, which I think is so critical.

Edwina Lowe (09:30) My obvious next question is, what do those datasets actually look like? And in terms of the creation of new datasets, how do you go about that process?

Eloise Goulder (09:40) Yes. Well, it's a great question and it's worth bearing in mind that our datasets should be we strive for, for them to be as proprietary as possible. And by that, I mean, we're either leveraging proprietary data assets that sit within our organization or, and/or we're leveraging proprietary techniques on top of perhaps publicly available data sources or commercially available data sources, but we are using the full breadth, scale, and technological capabilities that we have available to us at JP Morgan to create proprietary analytics on those data sets. So let me give one example, prime positioning. By that, I mean, we look at the aggregated and the critically anonymized collation of hedge fund positions from our Primebook. This is a give to get model. Clients opt into this dataset, but to the extent that the vast majority of our clients do opt into that dataset, we create an aggregated and an anonymized lens on the positioning of those clients in given stocks and in given indices and markets at a given point in time, on a daily basis, lagged by three days in this case.

Eloise Goulder (10:55) Now, it's absolutely critical that this dataset is aggregated and anonymized so that our clients feel comfortable that no individual client position can be back derived.

Edwina Lowe (11:08) And that's really critical.

Eloise Goulder (11:09) It's absolutely critical. So what sort of techniques do we use to ensure that that is the case? We use, for example, winsorization. By that, I mean, we chop off the tails of a distribution so that any extreme position is normalized and is taken out. We bucketize, and by that, I mean, we don't provide precise numbers for all of our data sets, but for some of our datasets, we just simply put them into buckets like deciles, and therefore you can only see if it changes if it moves from one bucket to another. We cap weightings. So if a given client or a given client's position exceeds a given percent weighting, we will cap it to ensure that outsized positions aren't showcased. These enhancements are absolutely critical to the dataset, most importantly, to make sure that they are truly anonymized, but also to make sure that we're taking out extreme so that we have a representative series over time for our datasets.

Eloise Goulder (12:04) So that's our prime positioning hedge fund dataset, for example. That's very clearly leveraging proprietary data that we have available to us as a result of the scale of the prime business here. Now, to give another example, which is not at its genesis utilizing a proprietary dataset, but does, I would argue, utilize a proprietary analytical technique, this would be our social media, social sentiment dataset, which you know so well. Here, again, it's a single stock dataset. It has significant history. Going back to 2009, it's provided on a daily basis. Here, we are ultimately developing relationships with social media companies. We have a paid for license with them to utilize it, to aggregate it, but this is an enormous dataset, and it's very hard to disentangle a signal from that noise, so to speak.

Edwina Lowe (12:59) And social media is a fantastic lens by which you can understand or try and understand retail investor activity.

Eloise Goulder (13:05) And it has been quite a feat to ingest that in such a way that we believe it has predictive power. We believe it has use for the investment process for us internally within JP Morgan and for our client base. So what have we done here in terms of the process? Well, even to ingest these millions of posts that are available to us, we have to spike detect. We can't possibly bring them all in to our platform. That alone is just too much data. So we have to spike detect to, to identify which data points we believe are worthy of coming in. They are either data points that relate to specific assets that we care about. For example, we track 3,000 stocks within the US and/or their posts that have significantly increased over time, which warrants their inclusion. So first of all, we do that. We then need to entity map those posts to determine which stocks or assets or entities they relate to.

Eloise Goulder (14:00) For example, if we're looking at posts relating to Apple, do we, are we taking Apple the fruit or are we taking Apple, the company? That in itself is such a critical distinction. So we've had to use our partners in the Machine Learning Center of Excellence, their techniques to correctly or as accurately as we can, entity map all of these posts to real financial assets that we care about. We not only create a volumes dataset, which is really how many posts are there for a given asset on a given day over time, but we also need to create the sentiment related to those posts over time. And sentiment analysis, I mean, it's really evolved over the years. 10 years ago, we utilized traditional NLP, natural language processing, and very traditional, arguably quite simplistic approaches were, for example, bag of words. Let's get a bag of positive words like up, beat, and a bag of negative words like down, disappoint, and let's map those posts to that bag of words to infer a sentiment for those interactions.

Eloise Goulder (15:04) But thankfully, technology has enormously helped us with that job. Um, we can now leverage LLMs to help annotate different posts, which has been transformational, as you mentioned earlier. We also leverage transformer models on the NLPs. For example, Google's Bert model and Facebook's Roberta, a really, uh, an augmentation of Google's Bert model. We utilize that specifically to aid this sentiment generation process.

Edwina Lowe (15:32) And to be able to look at that nuance that's not simply positive or negative. And actually, incredibly topical to be talking about this because we've seen a big resurgence of the retail investor this year, and in fact, the meme phenomena come back. So this is a very topical dataset that's, uh, highly appreciate or highly utilized by our clients.

Eloise Goulder (15:51) Yeah, absolutely. And when I think about the datasets that our clients are asking us the most for, I would say it comes down to positioning flows, sentiment based on as many sources as possible and textual data, as you mentioned before. I think all of those topics come up.

Edwina Lowe (16:07) I agree.

Eloise Goulder (16:08) And so what have we been developing in that space? Well, on the positioning side, we obviously have our hedge fund positioning dataset. We augment that with positioning from as many other investor types as we can track. So we look at retail investor flows, which as you say, have been on the rise, particularly post- COVID and indeed this year. We look at ETF flows, we look at CTA, inferred CTA positioning, we look at mutual fund positioning, we aggregate all of that together to have a lens on the aggregation of that positioning, that's our tactical positioning monitor, which itself is another dataset. That's the positioning and flows lens. We then have an equivalent for sentiment. We obviously have social sentiment, a lens on the retail investor. We now have professional sentiment, a lens on the institutional investor, and we have the aggregated tactical sentiment monitor, which really pulls all of these together.

Eloise Goulder (17:00) And of course, we could talk about the predictive power of each one of these series, which is a whole another debate. Uh, but that is absolutely critical when we're having our conversations with our clients is not only what data do you have, what are the attributes of those datasets, how much history, how much lag, how can I access it? But it's also what, what are the lessons you've learned? What are the predictive analytics that you've learned so far through all of your back tests? So there are obviously so many data sets that we can look at that we can create. There's such a journey in front of us in terms of continuing to enhance this. And of course, we're always client led. And Edwina, you're speaking to the clients, you're hearing from them about what they want to see. Can you speak to what clients have been asking for and also how we determine which data sets to prioritize?

Edwina Lowe (17:48) Yeah, it's a great question and it's a little bit of a balancing act in reality. Um, it, you know, to, to echo your point, the client feedback is absolutely critical in terms of determining our pipeline and where we allocate resource. You know, just to be very clear about this, bringing a new dataset to market is a considerable endeavor and doesn't just involve our team. We work with, as the point you made earlier, our partners across the business at JP Morgan. So we really need to make sure that when we put that effort in and allocate that resource, that we have the client demand to make it worthwhile, but that also has to be balanced with what's actually realistic, you know, in terms of really boring things perhaps, but things like licensing are absolutely critical. And when you talk about proprietary data, that comes with a whole host of sensitivities around it.

Edwina Lowe (18:37) But in terms of what's actually being demanded, or sorry, what's, in terms of what, where the client demand is, I would say that the rise of sentiment as a theme, so wanting to understand through a variety of different sources, what different investor types are thinking about or what the sentiment where the sentiment sits is key. So you talked about social media as a lens for the retail investor, but what we're now trying to do is broaden those horizons to try and have a lens on sentiment for the institutional investor. And of course, when we talk about the institutional investor, there are a variety of different investor types within that bucket, but one of the, um, an area that we're expanding upon is looking at reports or documentation from the industry to try and understand institutional sentiment. And then obviously within that, there's more that we can do.

Edwina Lowe (19:33) So we could look at different types of client or different types of investor, and we could also look at, for example, regional variations. There's a lot we can do there.

Eloise Goulder (19:43) Absolutely. And that demand for institutional, professional, textual content is so rife right now, isn't it? And that demand to consume it in a machine readable form so that it can be put into the knowledge basis, uh, of, of our investing clients. And whenever we present these data sets, it's so important for us to put them in a markets context and to look at them continuously through a markets context. Just to give a couple of examples, when we went through the real market lows in early April, the S&P 500, it dropped around 20% post liberation day. And at that time, our positioning data sets and specifically our tactical positioning monitor, it fell dramatically and it actually triggered an attractive signal based on our back tests. And this is something that our positioning intelligence team regularly write about, but even more powerful was to have an awareness of the market dynamics at the time.

Eloise Goulder (20:36) The fact that developments in the US political situation were very rapid. We had the 90 day pause in tariffs announced, and perhaps as a result of that, coupled with the fact that positioning levels were, were low, which is historically quite a ripe time to buy equities, markets then rebounded. And we've continued to see markets rebound since then driven by the retail investor, which comes back to all of our retail datasets, retail flows, retail net flows increasing, retail sentiment never actually got as light and it has been increasing. So I think putting these datasets into the market's context so that when we're using them real time on an ongoing basis, we can either read the datasets as they are through a purely systematic basis, which has an enormous amount of merit is arguably the most unbiased way of looking at the data, or we can add that quantitative view to more of a discretionary view in terms of how catalysts in the market are playing out to come up with more of a quantamental or a, a combined view, marrying the systematic with the discretionary view.

Eloise Goulder (21:38) That's also so powerful. And really as a team, we strive to do both.

Edwina Lowe (21:42) Well, arguably that's where our edge lies.

Eloise Goulder (21:45) Yes, indeed. So it's pulling together those threads from our market intelligence who, who are tracking markets real time through the day, leveraging these data sets with our positioning intelligence team who are doing exactly that, but specifically focused on positioning data to our da- data intelligence team, which is much more the data science capabilities creating these datasets such as social media. It's, it's combining, it's that aggregation of them all that we believe

Edwina Lowe (22:10) To be so powerful. And really that's at the heart of what our clients are looking for from us, is bringing those threads together to provide that combined lens.

Eloise Goulder (22:19) Absolutely. And isn't it so interesting and such a privilege for us to be facing all of these different investor types from the quants to the discretionary, to everything in between, from the portfolio managers and the analysts, uh, to the traders, to the central data sourcing teams, and to hear about their evolving client needs, which are evolving so rapidly and to be doing our best to service those needs.

Edwina Lowe (22:43) Eloise, I've heard you speak compellingly in the past about continuous learning and, you know, really honing intellectual curiosity. Could you talk a little bit about what that means to you and what you do in practice?

Eloise Goulder (22:55) I mean, I do think it entails listening to the client, the voice of the client. That's absolutely critical. It also entails being willing to accept feedback. It's not always fun learning that you need to change something, you need to create something different, you don't know where to start in terms of identifying a new data point, but at the end of the day, it's only hearing all of those requests that gives us a lens on our clients and how they're evolving and what they want. And it's the accumulation of all of that. It's the iterative process of all of those discussions that really builds our knowledge base over time and that really builds our capability to service our clients over time. And I really embrace that. I love that. What do you think, Edwina?

Edwina Lowe (23:39) I think it's a great point, but in terms of what it actually looks like, I mean, I'll speak for myself personally. I make an effort every morning to start with listening to a 10-minute news briefing so that I have some idea of what's going on in the market and the wider world. I find that 10 minutes a very manageable amount of time and actually the cumulative effect of that is significant. I, I also like scanning the newspapers and understanding what's going on. And then I, you know, without doubt, our colleagues, whether that's within our own team and we all sit together or across the wider business, but maintaining that curiosity and that seeing learning as an ongoing thing is vital. It's vital for continuous development and for being the best we can possibly be. And I'd like to talk to you a little bit about mentorship, both in terms of mentors for you or being a mentor to people who are earlier in their career.

Edwina Lowe (24:34) What does that mean to you?

Eloise Goulder (24:36) It's so important to continuously mentor, whether that's a formal mentee mentor rel- relationship or more commonly, I would argue across the industry, and for me, it's more of an informal thing. It's so important. I mean, if we look, if I look at my mentors, I almost call them our team stakeholders, as you know. And for a team like ours that's somewhat hybrid, that sits within trading, that's very close to research, that's very close to sales, that's very close to the data science and the data-driven parts of our organization, we have many stakeholders. And yes, that's a responsibility because we need to make sure that what we're doing is fully aligned with their businesses, but it's also such a privilege, such an opportunity to have those relationships with those individuals, all of those relationships different. But I really try to take on board the advice that they will give me and give our team in helping to steer our team in the most effective commercial direction.

Eloise Goulder (25:37) I see those, in a sense, as mentoring relationships, because those people are in good faith giving the best advice they can to me and to our team. And I really value those relationships. But equally, the importance of mentoring others, guiding others, supporting others, it can't be understated. I think for individuals in the organization to know that we are all approachable, we should all be approachable and that we all will take the time to hear concerns, to hear, uh, challenges, and to provide our feedback on that, I think it's completely critical for the ecosystem to work well. It's also, by the way, as powerful for the mentor as the mentee. And, and I truly believe that in all, in all senses. You know, for myself, if I meet with someone in a different business or perhaps in a business that's closer to us, and I hear their challenges and their concerns, I always learn something new.

Eloise Goulder (26:28) I love being able to help to the extent that I think I've got experience that, that might be relevant. Um, and I so often am reminded of things which are, which are critical to my own career and to our own team and to our own development. So I do see these relationships, they don't need to be formalized. In fact, overly formalizing them can, uh, can take away from the natural rapport and the natural dialogue and conversation, um, that's required, but I do believe they're really important. Another attribute that I think is, is really important to perhaps underestimating is consistency of delivery and following up. And it's relevant when it comes to stakeholder management, mentee, mentorship. I always love it when, um, people that I've, I've mentored follow up with interesting snippets, interesting articles, podcast recommendations, pieces of news. I love it when they do that because it's strengthening our relationship.

Eloise Goulder (27:24) It's keeping that line of communication open, which means the probability of us reaching out to each other in the future is that much higher. Uh, it's also truly informative and we learn much more from each other through that process. And, and I think that's so important in all directions. So the significance of following up, of consistently doing what you said you'd do, of consistently following through, um, it is, it does ultimately require a lot of hard work. And perhaps that sounds boring, perhaps that sounds dry, but to me, that is the greatest inspiration for creativity. You only really learn new things by doing, and you only really see parallels between different processes by doing, and therein lies so much of our creativity at the end of the day.

Asif Mohamed: Hi, I'm Asif Mohamed from Global Research, head of digital product.

Sam Azzarello: And I'm Sam Azzarello and I lead content strategy for global research.

Asif Mohamed: So Sam, how are you approaching digital research content to make it more impactful for our clients?

Sam Azzarello: Well, I think at the core, it's about meeting our clients where we are. So on our digital channels, we like to think omnichannel. That could be through email, through our portal, JP Morgan Markets, it could be through a podcast on Apple and Spotify, or it could be on LinkedIn or through client interactions. So there's many different ways that we think about getting our content, our best in class content to our client in a form that they want and meeting them where they already are. So Asif, I know we invest a lot. We do a lot to make sure we're at the forefront of what our clients need. Are there any recent tools that we've created to help our clients access our content?

Asif Mohamed: Absolutely. There are lots of technological innovations that are happening across the industry, and we've been leveraging AI, the cloud, and data in many ways to deliver content to clients in the way that they want to receive it.

Sam Azzarello: I know we've been doing things in AI. What can you share in terms of how AI is changing the global research model?

Asif Mohamed: We've been using AI for quite a while. We've been using it in macroeconomics. Our economists use AI to understand what central bank speeches are all about and when they're sentiments that are changing on a daily basis. We use it to identify thematic information in the news that can help our clients invest and generate more alpha. And we use it to understand how companies, sectors, and countries can be affected by thematic news on a daily basis.

We're also using AI in search. Search is a challenge, has been a challenge for the research industry for a while, and it's because keyword search doesn't give you the contextual value that folks are looking for. And so we're using AI to better search, scan across lots of information, decades of research, both written and data to be able to synthesize that information relatively quickly. And so that's a game changer, we think.

Sam Azzarello: Really valuable, I would add, right? AI is facilitating semantic search. And for those who are maybe not as familiar, it's this idea I can search DC or US policy or Washington and get all the content that I need no matter what term is perhaps mentioned in the report.

Asif Mohamed: That's correct. That's a contextual search, semantic, and we're able to generate insights faster, and that benefits the clients. It also benefits our analysts who are looking for information in lots of data.

Sam Azzarello: Also interesting to note that our analysts are using AI, like you mentioned, economics reports, sentiment coming out from central bank governors, different speeches that we're hearing. We're using AI to do sentiment analysis in a very innovative and value add way for clients.

Asif Mohamed: That's right. We've got the central bank product that looks across all of the central banks and their statements and analyzes it for whether it's hawkish or dovish sentiment coming from those central banks. So Sam, how do we meet clients where they are?

Sam Azzarello: Situations arise and we want to be able to bring our insights to clients and meet them where they are. So when tariff news broke earlier this year, we set up a webcast right away for clients, answering their questions. We had a podcast. We pulled together content from research analysts all across the globe, cross asset into one piece. And we had an omnichannel approach to how we brought these insights to our clients.

Asif Mohamed: And how are these research products changing and adapting to market conditions?

Sam Azzarello: Well, I always think our research analysts have to react to the news, but at the core of it, they're providing processes and frameworks and ways of thinking to our clients because our clients are building out a mosaic of a view, and JP Morgan Global Research is one input in that. So to the extent that our analysts with deep expertise and very deep knowledge can provide a way of thinking to clients and investors, I know that that's very valuable and it allows it to be adaptable for clients no matter what really happens with the news and how things change.

Asif Mohamed: I'd love to hear about how you got here and a little bit of your background.

Sam Azzarello: So I think my journey's a little long-winded, like many peoples'. I actually started as a research analyst. So I love markets and I learned cross-asset markets and how to do research. And then I found over time it evolved to wanting to be on the backend and making sure that that research gets to clients in the right way, at the right time, in the right format. So still out of the deep love of markets, but I would say the expression of that has evolved.

Asif Mohamed: And do you really like what you do today?

Sam Azzarello: I truly do. I mean, I'm having a blast and I think it's important that myself and the team have fun because it sounds maybe cliche, but we spend a lot of hours at work. We enjoy the people that we work with. And overall, I do want to make sure that I'm enjoying myself day-to-day. Asif, tell me about your background. How did you get to where you are?

Asif Mohamed: So I started out as an engineer many years ago and I actually became a banker in derivatives and converts and ran business development functions and ended up currently in research. But I think the common theme throughout all of these roles, whether it was in technology or banking, equity capital markets, has been around innovation. And in every step of the way, there's always been a large component of innovation that really helps bring clients in, deliver better product, whether it's pricing or risk. A big part of it is always about innovation.

Sam Azzarello: Amazing. I would add there's something very powerful about distribution. I mean, distribution at the end of the day is how something gets out into the world. So I know we have really good content. It's award-winning content, but I would also add equally important is how we get it to clients, in what form, and to your point about how innovative we are in terms of evolving as the landscape changes.

Asif Mohamed: We deliver at scale every day for our clients. We send 1.5 million emails to our clients each day. We have 96,000 clients visiting our JP Morgan Markets platform on a monthly basis. My research is the client's personal hub. They can find information there that is based on their subscriptions, and we also make recommendations there for things that they may not know that they need.

Sam Azzarello: At scale is the phrase that always either adds the hiccup or in our case, the opportunity. So doing this for thousands of clients, millions of emails a day on all of our platforms, and I'd add we have a thousand analysts across the globe, covering every company, asset class, country, it's at scale.

Asif Mohamed: Yeah. And clients are demanding it. So we tend to want to skate to where the puck's going to be, and that requires sort of lengthy investments over a longer period of time. So it's not a one and done, it's an ongoing investment. Learning, being dynamic, delivering as the markets are changing and our clients' needs are changing. Who are the content creators within the product?

Sam Azzarello: Global Research is an absolute content powerhouse. We produce more than 130,000 pieces of content per year, and then we augment that with data sets and data experiences, lots of analysis. We cover countries, sectors, over 5,000 companies. I'm always blown away at the level of depth and expertise of our research analysts. I know however, they don't do it in a vacuum. They're working with sales and trading to get ideas. They're keeping a pulse on what's trending, what's important to clients. There is a continual dialogue between our research analysts and clients.

I also know they're keeping their eye out for what's coming, what's next, what's on the horizon, and they do that through data, alternative data, talking to different stakeholders and just drawing on their immense expertise that they've built from decades of being in their markets.

Asif Mohamed: So there are no reports without our analysts.

Sam Azzarello: There's a lot of trust and rapport built up over time between our research analysts and clients. And I think it's a show of trust that clients come to us in moments of market volatility or news flow or disruption, and they come to our research analysts to get their view because they trust our analysts and the breadth and depth of knowledge and expertise is there and it's been established. Insights from our research analysts are augmented with data use, and we know there is an insatiable appetite for more and more data. So Asif, what are we doing with respect to data in Global Research?

Asif Mohamed: You're right. Data is exploding, has been exploding for quite a while, and investment decisions are using more and more data. The electronic delivery of research, we call it machine-readable research, has become more and more important. Clients would like to search JP Morgan's research as well as their own proprietary data sets and content and generate insights. So data is a big deal for us, has always been.

We have DataQuery, which is our data platform on the JP Morgan Markets platform that has 13 million time series of data that investors are using to make investment decisions. We've also created industry differentiated data products such as JPMaQS, which is a macroeconomic quantummental data set that is used for making investment decisions and then traders use it for backtesting their trading strategies. So there's a wealth of existing information and we're creating new products as well to meet our clients' needs.

Sam Azzarello: And that's a powerful combination, augmenting our research analysts' insights with new and proprietary data sets to lead to faster and smarter insights.

Asif Mohamed: So how do we keep delivering for clients?

Sam Azzarello: We keep delivering for clients by making it as easy as possible for them. So we know the amount, the volume of information is only growing. So helping clients get to the key insights they need as efficiently and effectively as possible, that's the way forward.

Asif Mohamed: And what are some of the challenges that you're seeing today in terms of the need for information?

Sam Azzarello: Cross-asset is a space where I think we're going to see more of an evolution because many market events now, they ripple around the globe. They impact all the asset classes and the ability for us to pull together the breadth and depth of our expertise, light up that theme for clients and showcase it in a cross-asset format is going to become more powerful.

Asif Mohamed: And that's a space where we've been accelerating in the use of AI to generate insights faster by looking across lots of information, volumes and volumes of information, both written and data to generate those insights. So looking for a needle in a haystack in a very structured way is really the focus of some of the modern products within Global Research. So where do these ideas come from?

Sam Azzarello: Idea generation in research is a very creative process, I would say, and it's going to vary analyst to analyst, but it could be a tidbit overheard at a client event. It could be a news article. It could be a piece of data or a chart. It could be inspiration walking on the street. And from there, I know that our research analysts are all about collaboration. So they're bringing together their team, bringing together related asset classes and researchers across the department.

And from there, they're refining the idea, they're doing analysis, they're using data. And there's a lot of work and intellectual creativity that has to go into this process before a research note gets published. Asif, let's talk about personalization, the next frontier. So what does personalization look like for Global Research?

Asif Mohamed: So personalization for us is really about ensuring that the thousands, hundreds of thousands of clients that we have, that we have information about how to serve them best. And they've given us a lot of that information and we use it to give them a very bespoke experience and understanding obviously the research reports that they're consuming, but also delivering potentially data and other products and services that may be relevant for them.

And so in my view, personalization really tries to serve each client individually based on their needs, whether it's during a call with an analyst or through consumption of their reports. More importantly, we do listen to our clients. And so listening to clients, getting feedback helps us deliver better products and grow our products and our services for those clients.

Sam Azzarello: And I know there can at times be a little bit of FOMO, right? Clients want to know what they might be missing out on, what the next new idea is, who's a new analyst perhaps that they haven't engaged with before. So to the extent that recommendations and personalization does that, we know we're serving clients through those channels.

Asif Mohamed: That's right. We're serving clients through personalization and recommendations, both in thematic as well as process type products that they may not be aware of, we're helping them to find those proactively. Again, we create a lot of content. How does a client really identify something that they're not focused on every day? We're using technology, AI, and the client's feedback to get them the content that they may not be aware of, that they absolutely should look at for their investment decisions.