Most buy-side firms know how to find an opportunity. What an increasing number are discovering is that the gap between identifying one and acting on it has become a structural weakness. And that the weakness sits not in the front office but in the operating model underneath it.
That was the recurring argument of a recent A-Team Group webinar entitled “Agility as Alpha: How Trading Infrastructure Determines Who Wins in Volatile Markets“, which set out to examine whether trading infrastructure has become a genuine source of alpha amid heightened market turbulence. The discussion, sponsored by FIS and moderated by Mike O’Hara, Editor at A-Team Group, brought together Richard Bell, Head of Execution Technology at Nickel Digital Asset Management; Oskar Wantola, Managing Director and Head of Execution Technology at Man Group; and Ronan Farrell, Global Head of Buy Side for FIS Cross-Asset Trading and Risk. What emerged was less a story about speed for its own sake than about the cost of operating models designed for conditions that no longer hold.
Short volatility by design
Farrell offered the sharpest framing of the problem. Many buy-side firms, he argued, have effectively become short volatility in the way they are built, with operating models designed for efficiency and optimisation in a world of low rates and low geopolitical risk. As that world has reversed, he says, firms find themselves needing that level of agility to know what their risk is so that they can execute on it.
The consequence is that years of redesigning operations around scale and cost have, in a more turbulent environment, started to work against firms. The pressure now, Farrell argued, is a return to more fundamental questions about whether a firm can see its risk in real time and act on it the same day, rather than waiting for a position to be reconciled the following morning. That demand for same-day certainty ran through much of the discussion.
That demand is not abstract. Bell, whose firm supports 70 external trading teams and processes 400,000 trades a day, described the practical reality of operating in markets that never close. He pointed to the night in April last year when Iran launched a barrage of missiles at Israel – a Saturday, with traditional markets shut but crypto trading live. Automated circuit breakers and proactive risk controls, he says, were what kept the open markets running through the event. The lesson he drew is that agility is not only about capturing upside.
Agility protects alpha, not just produces it
Bell pressed the point further, challenging the webinar’s own premise. Agility, he argued, is as much a defensive discipline as an offensive one: it does not only produce alpha, it protects it. At 400,000 trades a day, every one carries value and every one can go wrong, and a single execution failure can erase a day’s edge. Identifying a strong alpha signal, he argued, counts for little without equal attention to the detail of how it is executed.
Wantola, who oversees a single central execution system at Man Group trading around $8 trillion a year, reinforced the point from the perspective of system resilience. At scale, he said, the risk of a real-time system going down is itself a material exposure. The answer is not only a sophisticated risk framework that can pause or halt trading when it detects irregularities, but the assumption that systems – internal, vendor or broker – will at some point fail. The firms that cope best, he argued, are those that can switch to a backup and keep trading in a second-best mode, which almost always beats stopping altogether.
The data problem underneath the risk problem
Asked how firms should maintain a unified risk view across an expanding set of instruments, the panel converged on data rather than risk engines. Wantola made the case that the analytics are rarely the constraint. The harder task is extracting and unifying data from systems that were never designed to talk to one another. Running data from multiple sources through one central risk component, more frequently than firms have historically managed, is the direction of travel.
Farrell tied this to a broader convergence. As private and public markets blur, the universe of assets a firm cannot see is shrinking. The old model of a separate technology stack per asset class is giving way to a single question: what is my exposure to a given theme or sector across everything I hold, right now? The challenge, he framed, is one of timeliness rather than calculation – not whether risk can be calculated, but whether the data arrives quickly enough to act on.
The poll results gave that thesis some empirical weight. Asked how long it takes to go from deciding to pursue a new opportunity to placing a first trade, the audience split evenly – 33% answering weeks and another 33% answering months, with just 22% able to act within days. The figures genuinely surprised the panel, who found it striking that adding a first trade in a new asset class, geography or strategy could take weeks or months at all.
Where the panel disagreed
The liveliest exchange came over electronic execution and how infrastructure should handle the migration of historically voice-traded and OTC instruments onto screens. Wantola set out a model in which an execution system aggregates both electronic and off-market liquidity internally, then lets the algorithm decide where an order should go – routing electronic flow through standard connectivity and pushing voice orders to traders, increasingly via semi-automated integrations with chat-based tools.
Bell disagreed. For some markets, he argued, internalising and routing that flow is precisely the wrong instinct, because the value of the voice channel lies in discretion – the ability to transact in size without signalling intent to the rest of the market and moving the price against the client. His point was not a defence of opacity – in digital assets, he noted, settlement remains visible on-chain – but an argument that some markets are better served by keeping voice and electronic execution deliberately separate.
The divergence matters because it cuts to what ‘agility’ actually requires of infrastructure. The consensus that did emerge was not that everything should be electronified, but that an execution platform must accommodate the full spectrum – dark, lit, voice and chat – and increasingly use AI to parse unstructured channels and route from them. Farrell extended the thought to execution strategy itself, noting that the range of ways to take on exposure, including synthetic and derivative routes, is itself a capability firms now demand of their platforms.
On the related question of consolidation versus best-of-breed, the panel landed closer to agreement. Wantola argued for a pragmatic middle: reusable, API-driven core modules for risk, data and connectivity, with specialist point solutions reserved for cases where they add genuine value – held together by interoperability. Bell reached for a term he has used before, the ‘multilith’, to describe the architecture most firms actually need: neither monolith nor fully modular, but a tested, secure core with lighter, faster components layered on top.
The undercurrent: AI
Although AI was not a billed topic, it surfaced unprompted in answer after answer – scenario generation for risk, the ability to “vibe up a trading solution in an afternoon,” the automation of data cleaning and integration. By the closing question, on the single most important shift firms should prioritise over the next 12 to 18 months, it had become the explicit answer.
Bell’s warning was pointed: firms that treat AI as a reason to shrink front-office engineering teams are misreading it. The opportunity, he argues, is to clear a backlog of revenue-generating work that previously had to be triaged by return on investment – not to cut capacity. Wantola agreed, framing the shift as much about people as technology: AI-ready systems will accelerate some processes while exposing others, particularly human decision-making and vendor integration, as the new bottlenecks.
Farrell closed by bringing the discussion back to people. AI and tokenisation matter, he says, but alpha ultimately comes down to empowering people to be creative, find ideas and act on them. For all the focus on infrastructure, the panel’s final word located the source of alpha not in the platform but in the people the platform is meant to free up.
For most firms, the binding constraint is no longer the ability to spot an opportunity, but whether the operating model lets their people act on it. On the evidence of the poll, for many it still does not.
Subscribe to our newsletter
