Banks are increasingly centralising FX operations to cut costs, improve risk management and capture more trading revenue says Matthew Avery, managing director at Integral. Speaking to Trader TV at Trade Tech FX in Amsterdam, Avery discusses how API integration is transforming FX workflows, enabling real-time pricing and risk visibility. He also unpacks how regulatory resilience requirements and AI-driven automation could accelerate adoption, as banks prepare for growing demand for automated execution and distribution.
Interview
Josephine Gallagher – Welcome to Trader TV. We’re here at Trade Tech FX in Amsterdam, and I’m joined by Matthew Avery, Managing Director at Integral. Matthew, lovely to speak to you.
Matthew Avery, Integral – Hi Jo. Great to be here.
Josephine Gallagher – So we’re seeing a big push among banks and buy-side firms to centralize some of their FX operations and desks. What are you seeing from firms in terms of how they’re adapting their FX functions, and what are the leading drivers?
Matthew Avery, Integral – Yeah, so FX is inside everything within a bank. Typically, the FX function within an organization was responsible for the larger corporate and institutional flow, but there was FX happening across the whole organization, inside the equities desk, inside the bond desk, inside the options desk, payments divisions, retail banking, private banking, and all of their overseas subsidiaries. So banks have realized that they are managing complex IT projects inside their organization, as well as potentially multi-vendor solutions that they are paying for across the whole business. So what we’re seeing right now is banks pushing to centralize that core FX function, and that means one source of price discovery, one price being constructed and distributed across the whole organization, and critically having a centralized view of all of the resulting risk and credit exposure that they have across the whole of the organization. So that is what the push is for right now. There are significant cost drivers to that. As I say, there are potentially expensive internal IT resources being tied up in supporting a lot of kind of archaic workflows, and also the potential of multiple vendors being suppliers across the group. But also, and this is critical, there’s a significant revenue opportunity being missed right now because all of that flow that isn’t being centralized with the main trading function is leaking out of the group and going to third parties. So capturing all of that flow internally gives you the opportunity to internalize and capture more yield from that flow, and that’s a critical driver for the banks that we’re working with right now. All of these banks have a significant FX franchise, just not many of them are accessing it right now.
Josephine Gallagher – Okay, so one of the things as well that we’re seeing is that there’s you know advances being made in API integration. So what kind of role do you see the likes of API integration in pushing forward, like auto execution, real time risk management, and how mature is the technology?
Matthew Avery, Integral – The last 10 years have seen a real race to shave microseconds off execution time. We see the next 10 years very differently. There is still going to be a market segment that requires ultra-low-latency execution, and the APIs will continue to serve that market segment, but what we’re seeing now is that utility is more important than speed. Customers are looking for flexible, functional APIs that solve for real workflow problems within their organization, and those APIs need to deliver two things. One, they need to be able to support the service level that the customer and the bank would like to operate between themselves, so that could be streaming, real-time pricing. It could be RFQs. It could be fixed-period rate files. Whatever the service level, the engine behind it needs to support it, and the API needs to support it. And importantly, the API needs to be easy to consume within the bank and within the customer organization, and that means easy-to-consume REST and WebSocket APIs that can easily integrate to core banking systems and treasury management systems. There will still be the fixed binary protocols that service trading desks around the world, but the future is in that type of functional API that solves for real problems in the client business. What that means when those APIs are embedded into the workflows is that now banks have the opportunity to see their risk exposure in real time as those trades start to flow through to the bank in real time, rather than waiting for some of the more manual processes that existed prior to this level of automation. So the technology is mature, adoption isn’t, and that’s where we see a big change coming in the next couple of years.
Josephine Gallagher – Okay, and as part of this as well, there’s a higher level of trust when it comes to API integration. So, like when you’re working with a very heavily regulated bank, their third parties often, so you have to have a high level of trust with them. So, what are the kind of different levels of this kind of API integration that you can maybe explain for us, and what is the standard that a vendor needs to meet in order to be a credible partner?
Matthew Avery, Integral – Yeah, so there are four layers that we can. In terms of the API integrations, one is the pricing layer-the ability to consume market data and pricing from any source and distribute that pricing out anywhere it needs to go. The second is the execution layer; you need to be able to receive orders and send fields back on those APIs. Then you need to be able to send the post-trade STP messages wherever they need to go as a third layer, but the fourth layer is around the workflow and operations, and this is when you start to touch credit and and balances within the the core bank, and this is where trust becomes critical. So trust used to be a adjective on a slide deck, it was a function of relationship. One of the key evolutions that we’ve seen over the last couple of years has been the evolution of the DORA regulations in Europe. What that has done has made trust become measurable. So now, under DORA, trust is data-driven. It’s around your uptime, it’s around your response and resolution times, it’s about your business continuity plans, it’s around your exit plans that you put in place with your customers, and that is now a measurable exercise that banks can see as a function of trust, underpinning all of that is the type of certifications that are just critical now to any vendor supply and regulated institutions. So that’s SOC 2 Type II certifications from independent auditors, underpinned with penetration testing on a regular basis and data residency agreements that you can pin to a region. in the future, you’re not going to be able to service regulated institutions without being able to evidence your resilience.
Josephine Gallagher – Okay, we could talk at length on this topic. I could say, but I’ll have to ask this final question for you. So now, thinking about everything we just discussed, over the next three years, how do you see developments in, you know, the likes of FX embedding into systems or the centralisation of FX systems, API connectivity, and also new developments in AI shaping the way buy sides and sell sides interact?
Matthew Avery, Integral – Yeah, I’ve been fortunate enough to be in this industry for 30 years. I’ve seen the transition from phones to screens, whether they are single-dealer platforms or multi-dealer platforms, and then from screens to messages, which is essentially APIs. We see a world very soon where a corporate treasurer’s AI assistant will be calling for prices and sending back orders in an automated fashion through agentic workflow, our customers are already starting to ask for that level of service. This all becomes an API question. 60% of banks are expecting AI to have a significant impact on their distribution in the coming years. Right now, only 8% of corporates are actively piloting any type of AI workflow. 42% are currently in kind of watch mode. So this evolution from phones to screens to messages is an API conversation. These APIs have been around for a long time. The challenge has been the integration, and we’ve talked a little bit about how we can make that easier. But AI is the greatest enabler that we’ve ever seen for adoption of API workflows. So those customers that have never asked for API access before will leapfrog directly to AI workflows, and the banks that will succeed will be the ones that are ready to meet that demand.
Josephine Gallagher – Wow, lots to come and lots of change expected.
Matthew Avery, Integral – Absolutely,
Josephine Gallagher – Matthew. Thank you very much for speaking to us. Thank you very much for joining us.
Matthew Avery, Integral – Pleasure.
Josephine Gallagher – I’d like to thank Matthew for his insight and, of course, for watching. This has been Trader TV at Trade Tech FX in Amsterdam.
