Post lunch on the second day of Optic 2026, five experts roused the crowd with a lively debate on the motion “This house would trust Claude to manage its settlement fails”. The verdict? Claude can probably be trusted – but it comes with caveats.
Peter Tomlinson, managing director of Equities Trading and Post Trade at AFME, was moderator. On the proposition were Kerril Burke, CEO of PayRecAI, and Niamh Kingsley, CEO and founder of Ace and chief technology officer at G Mass. Playing the opposition were Rebecca Healey, managing partner at Redlap Consulting, and Thomas Daniels, post trade industry expert at Digital and Data.
Kingsley opened the debate by pointing out that the workflows for AI agents and operations employees already share an uncanny similarity. Operations analysts have “a rule book that checks the work and a senior manager who plans the outcomes”. When a settlement fails, anything outside of their limits is escalated. “That person doesn’t own the risk; they manage it,” she says. Similarly, “an agent does the work, rules vet it, but humans keep the judgement”.
Some hesitancy
Daniels was next up with his counter-argument. “I will absolutely trust Claude to help manage settlement fails,” he concedes. “But would I trust it to manage them autonomously? Not yet.”
“Capability does not equal authority. I don’t doubt that Claude could eventually outperform many human operators in identifying and resolving settlement fails. But does that automatically give it the authority to act? I would argue not.”
He shared a second point: settlement fails are not always simply an administrative problem. Issues such as insufficient securities, counterparty disputes, liquidity shortages, and collateral shortfalls may require “commercial judgement and an understanding of risk across the institution”.
He emphasises, “the real constraint is governance, not technology. I’d happily trust Claude to identify settlement fails, investigate their causes, and recommend remedies. I’d even allow it to resolve straightforward exceptions autonomously, provided it’s operating within clearly defined controls. But when decisions involve committing capital, accepting financial risk, or prioritising one part over another, that is where I draw the line.”
His final point was, in his own words, a more philisophical one – the need to preserve human capital. “Junior operation professionals learn by investigating fails, understanding markets, and managing exceptions. If AI automates all of that, how are we going to develop the next generation of experts capable of supervising?”
Indeed, AI is not capable of anything that requires judgement, but Kingsley believes that an effective safeguard against that – which is also the crux of the argument for the proposition – are limits and the rule that anything beyond those limits should be escalated. “We know that there needs to be rubber stamping, because humans rubber stamping the work is what already happens today,” she says.
Put the brakes on
For Healey, the speed at which AI is evolving is an issue – she believes it is moving too quickly for the industry to understand its full capabilities. “The speed of innovation is pulling us back to reconsider what we need to look at today in terms of control.” In the future, know-your-agent (KYA) might become just as much a standard practice as know-your-customer (KYC), she predicted.
Healey’s concern with security is one that resonated with the audience. In answer to a question about data resilience, Kingsley pointed out that it is not an issue unique to using AI in settlement fails. Anyone working with cloud infrastructure or market intelligence would face the same challenge. These challenges, she countered, have more to do with data management, governance, and third-party regimes – which form a separate layer from AI challenges.
“I would ask the audience to rememeber that the question is whether we would trust Claude to manage settlement fails – not if we would trust it end-to-end, for the entire process. So would I trust Claude? Yes, I would.”