The engine room of finance is being rebuilt. Exclusive insights from Sibos experts reveal how the shift from traditional automation to autonomous, agentic AI systems is mastering operational complexity, reducing settlement failures (T+1), and defining the new gold standard for efficiency, resilience, and operational risk management.
The engine room of global finance operations, post-trade, and reconciliation, is being fundamentally rebuilt. For decades, firms relied on legacy batch processing, siloed data, and armies of human staff to plug the gaps. That model is now obsolete.
The consensus at Sibos was clear: Artificial Intelligence is not just an add-on; it is the systemic architecture for future resilience, speed, and efficiency. Financial institutions are rapidly graduating from simple Robotic Process Automation (RPA) to complex, agentic AI systems that don’t just follow rules, but learn, adapt, and act autonomously.
The most significant shift in operations is the move to agentic AI. This model goes beyond prescriptive, rule-based automation (traditional RPA) to create autonomous systems that can handle the messy reality of financial data.
James Maxfield, Chief Product Officer at Duco, defined the term practically for capital markets: “Think of an agent as an autonomous system that combines models and AI components to complete tasks often roles a human would perform. The autonomy matters: agents can adapt, learn procedures, and handle variations rather than stopping when a scenario isn’t rule-perfect.”
In the core operations of reconciliation, agentic AI offers a critical advantage: instant scalability. When market volatility strikes and volumes spike, “you can scale agents instantly, adding more in minutes,” Maxfield noted. This allows firms to maintain control and reduce stress on human capacity.
Mick Fennell, Business Line Director for Payments at Temenos, confirmed this integration, noting the use of generative and agentic AI across their platforms. Temenos has seen immediate results, including a tier-one bank using agentic AI for processing that achieved a 20% reduction in repair rates a direct efficiency gain.
AI is now being directly embedded into core financial engines to tackle the industry’s biggest pressures: the need for speed (T+1), operational resilience, and cost reduction.
The shift to T+1 settlement (settlement in one day) is not just a technology upgrade; it’s a massive operational challenge. Danny Green, Head of International Post-Trade at Broadridge, stressed that the “time to settlement is cut in half,” eliminating the “luxury of ‘we’ll deal with exceptions tomorrow.'”
AI is the essential tool for survival in this accelerated environment:
Even in fundamental control functions like reconciliations, AI is speeding up the process. Jack Niven, VP North America at AutoRek, noted that while AI can’t “magic away” a true break, it dramatically helps with workflow triggers and, critically, with faster configuration—spinning up new products, desks, and their corresponding reconciliations in hours or days, not months.
In payments, where volumes are climbing and delivery time is shrinking, AI is indispensable for resilience. Fennell (Temenos) highlighted its use in payments repair, where AI learns patterns and suggests rules to improve Straight-Through Processing (STP).
Despite the buzz, many executives express frustration over vague AI Return on Investment (ROI). Andy Schmidt, Global Industry Lead, Banking at CGI, advises firms to stop simply piloting existing processes and start being outcome-driven.
“Don’t just ask, ‘how many more payments can I process?’… Instead, ask, ‘how many more loans can I process?‘ since you know the yield of a loan,” Schmidt advised. By targeting outcomes with clear value—such as automating loan onboarding or optimizing call center routing—banks can prove concrete ROI.
With agentic systems holding the power to act, governance is the bedrock of adoption. The industry agrees: a complete “black box” is unacceptable.
Trust, as Schmidt (CGI) notes, means transparency, explainability, and keeping humans in the loop. The operational risk of the “rogue agent” is often lower than the risk of the “rogue human,” Maxfield (Duco) pointed out, but buyers must demand a responsible autonomy framework that includes:
The biggest barrier to adoption, however, isn’t the technology, but the human dynamic. Maxfield cautions that “The biggest blocker is people impact, not models.” Success hinges on change management, job design, and creating a human + agent team where staff are incentivized to focus on higher-value work analysis and policy, while agents execute repeatable operational tasks.
The future of financial operations is here now. It is a cloud-native, API-driven, and intelligently automated ecosystem where machines handle the operational volatility, freeing humans to master the strategy.