In this episode of Caplin Connects, John Ashworth, CEO of Caplin, is joined by Graham Moss, Global Head of FX IT at BBVA, to reflect on the evolution of FX technology from the vendor side to the banking front line. They discuss the early success of platforms like Reuters RET, the shifting role of voice trading, and the challenges banks face when building their own FX stacks. Graham shares insights on the changing demands of tier-two and tier-three banks, the impact of commoditisation in trading tech, and why GenAI is now a must-have, not a nice-to-have. A compelling listen for anyone navigating the future of electronic trading.
Read the articleArticle
In this episode of Caplin Connects, John Ashworth, CEO of Caplin Systems, is joined by Graham Moss, Global Head of FX IT at BBVA. Graham brings a distinctive perspective to the conversation; part technologist, part market practitioner, and shaped by experience on both the vendor and bank sides of the FX industry.
Having started his career at Thomson Reuters during the formative years of electronic FX trading, and now leading FX technology at a global bank with strong European and Latin American franchises, Graham offers a grounded view of how FX trading applications have evolved; not just technologically, but organisationally and culturally. The discussion moves fluidly from early automated trading tools to today’s conversations about AI, digitisation and what it really means to modernise a trading stack.
Graham joined Thomson Reuters in 2007, just as the acquisition of Reuters was announced and shortly before the global financial crisis. At that point, electronic FX trading was no longer experimental, but it was far from mature. Many banks relied heavily on vendor platforms, not because they preferred to outsource, but because there were few credible alternatives.
Tier-one banks were already pushing the boundaries of what the technology could do, particularly beyond vanilla FX products. Smaller institutions, by contrast, were often still discovering what electronic trading meant for their business.
This asymmetry shaped the early vendor-bank relationship. In some cases, banks guided vendors; in others, vendors effectively educated banks on what was possible. What unified both sides was a growing recognition that voice-only models could not scale.
One of the most significant shifts Graham reflects on is the rise of automated trading. What began as a way to handle low-risk, high-frequency trades quickly became a foundational capability within FX trading applications.
The original proposition was simple: automate predictable flows, free up human traders to focus on complex or high-value transactions, and reduce operational risk in the process.
Over time, however, automation moved from assistance to substitution. Algorithms could manage flows faster and more consistently than even the most skilled voice desks. Today, Graham notes, even large franchises operate with far fewer voice traders than in the past, “What started as a way to automate very predictable flows ended up changing the whole shape of the desk. You’re not removing people; you’re changing what they spend their time on.”
This evolution mirrors broader changes in trader workflows. Human expertise has not disappeared, but its role has shifted from managing volume to overseeing exceptions, risk and client relationships.
Drawing on his experience on the vendor side, Graham highlights a persistent challenge in FX technology: no two banks are the same. Differences in geography, operating model, governance and culture all shape how technology is adopted.
What works for a tier-one global bank in London may be entirely unsuited to a regional institution operating across Latin America or Southern Europe. This makes scalability difficult for vendors and raises hard questions for banks considering whether to build, buy or outsource.
Graham explains that for many tier-two and tier-three banks, running a full FX stack in-house is rarely economical. “You need to think very carefully about whether you want to do it in-house because it's going to take a lot of time. It needs a huge amount of expertise.” While FX products themselves may be relatively simple, doing them well requires proximity to liquidity venues, specialist skillsets and significant operational investment.
In this context, outsourced trading solutions and managed FX platforms become pragmatic choices rather than strategic compromises.
Asked whether FX technology has become commoditised, Graham offers a nuanced answer. Connectivity and venue access, once complex differentiators, are now largely standardised. Latency has improved dramatically, pushing parts of the market into microsecond territory.
Yet end-client demands, he argues, have not fundamentally changed. Most FX volume remains concentrated in vanilla products. The real progress has come from better digitisation and packaging; making trading simpler, faster and more transparent for users. Executable streaming prices, one-click trading and improved user interfaces have reshaped expectations. Competition between banks has accelerated this trend, driving continual refinement rather than radical reinvention.
In this sense, modern FX trading tech is less about novelty and more about reliability, performance and integration into the broader trading lifecycle.
An often overlooked dynamic in FX markets is that banks themselves can be clients. Smaller institutions face the same build-versus-buy dilemma that larger banks confronted years earlier. Graham is candid about the risks of in-house development for these firms. Without sufficient scale, expertise or geographic reach, running a proprietary FX platform can quickly become a distraction.
Handing responsibility to a specialist provider may lack prestige, but it often delivers better outcomes; allowing banks to focus on client relationships, balance sheet usage and risk management rather than infrastructure maintenance. This reality reinforces the importance of robust forex trade management software that supports end-to-end workflows without demanding disproportionate internal investment.
Like many seasoned technologists, Graham approached AI with caution. His earlier experience with natural language processing tempered expectations, particularly in highly regulated environments such as banking.
That scepticism has now softened. “If the right focus is put in, if smart people are put in the room with the right investment, it’s going to do fantastic things.” And large language models, backed by serious investment and computing power, are beginning to demonstrate tangible value. Within BBVA, AI is no longer an experiment at the margins, but a capability being embedded into platforms and processes.
Crucially, Graham frames AI not as a replacement for expertise, but as an amplifier of it. Used well, AI tools act as research assistants, coding aids and productivity enhancers; extending the reach of skilled professionals rather than diminishing it.
... CONTINUED
The implication for FX trading applications is significant. Platforms must be architected to support agentic workflows and AI-driven decision support, even if the full benefits are still emerging. “If in three years’ time we don’t have a platform that supports agentic calls to make decisions and changes, then I probably failed, to be honest. We need to be ready for that.”
Looking ahead, Graham is less concerned with delivering a specific feature set than with building adaptable frameworks. In his view, success lies in preparing platforms to absorb change; whether driven by AI, organisational transformation or market consolidation.
European banking, he notes, is entering a period of consolidation. Institutions that survive and thrive will be those that combine scale with efficiency, and ambition with execution discipline. For FX technology leaders, this places a premium on clarity of purpose. Modernisation is not about chasing trends, but about enabling growth while managing risk.
FINAL THOUGHTS
Despite rapid technological change, Graham ends on a note of continuity. New generations of developers and traders will arrive fluent in AI-assisted workflows, just as previous generations adapted to electronic trading and automation.
What endures is the value of experience; understanding market structure, client behaviour and the consequences of failure. The most effective teams will be those that blend new tools with hard-won judgment.
For banks investing in modern FX trading applications and broader electronic trading capabilities, the message echoes across the Caplin Connects series: technology is essential, but it is only ever part of the story. The real differentiator lies in how well it is applied, governed and aligned with the realities of the market.
In this episode of Caplin Connects, John Ashworth welcomes Bob Tull, Global Head of Fixed Income, Currencies, and Commodities at Fifth Third Bank. Bob shares insights on the evolution of trading floors, the rise of electronification and the pivotal role regional banks play in the U.S. financial system. They explore the future of stablecoins, blockchain, and post-trade automation, the importance of relationship-driven banking and how emerging talent can blend technical expertise with interpersonal skills. Bob also reflects on leadership lessons, market changes and his vision for Fifth Third’s growth in an increasingly digital, relationship-driven future.
We're always excited to connect with listeners and industry participants alike, and invite you to be part of our podcast discussions. If you would like to appear on the show, know of someone who would be an engaging guest, or would like a particular topic covered, we want to hear from you!
Get in Touch