Innovation Day: London brought together industry leaders and expert speakers to learn, discuss and connect. It was a valuable afternoon of insights and conversation on the biggest challenges and opportunities ahead of the industry.
Duco CEO Michael Chin welcomed our guests, and highlighted the rapid pace of change the industry is going through.
“A lot has happened in the 12 months since I welcomed many of you to our Innovation Day last year. This is a fantastic time for Duco, for Operations, and for the industry as a whole. I hope you already feel that a big change is brewing.”
Indeed, back at Innovation Day 2024 our AI panel concluded that agentic AI was the future. Similarly, at last year’s Sibos, the conversations were largely hypothetical, focussed on how agentic AI could help financial firms.
Jump forward a year, and at Sibos 2025 the conversation was around how firms are using agentic AI, and the results they’re seeing, Chin said. Meanwhile, we launched our agentic rule builder at the conference, which we followed up by announcing agentic workspace at Innovation Day: London.
“For Duco, AI is not theoretical. It's not conceptual. It is actually in action, fully integrated in our platform and, most importantly, delivering value to our clients.”
This surge in AI innovation and adoption is bringing with it new opportunities, while reinforcing the need to focus on strong foundations - in particular: data. The data being created is growing exponentially, driven by surges in innovation. And much of this data needs to be transformed, reconciled, validated and reported.
“One of the reasons now is such an exciting period for the industry is because we’re seeing the convergence of technological possibility and strategic thinking in a way that is opening up a world of potential for the future. People are rethinking what Operations is for.”
We then kicked off a packed agenda full of insights and experts to excite the imagination and explore what’s possible. Read on to discover some of the highlights from the day, and subscribe to our newsletter so you don't miss our in-depth takeaways in the coming weeks.
In this article:
- Duco’s agentic AI journey
- AWS: Structuring your teams for an agentic world
- Top talking points from your industry peers
- Oliver Wyman: 2026 wholesale banking outlook
- Panel with HSBC and Nomura: Innovating in times of uncertainty
- Key takeaways from Innovation Day: London 2025
2026 is all about agentic AI
James Maxfield, our Chief Product Officer, shared Duco’s agentic vision at Innovation Day: London.
He explored how Duco has evolved from a standalone reconciliation tool into a comprehensive, modular data automation platform. One designed to integrate seamlessly within complex financial architectures.
Duco, Maxfield explained, now addresses the full data lifecycle. The platform can ingest and normalise structured, semi-structured, and unstructured data before moving through reconciliation, exception management, and downstream publishing.
Enhancing data control with agentic AI
Maxfield then dived into agentic workspace - a centralised environment within the platform that harnesses generative AI to fundamentally change how users interact with data controls.
Agentic workspace empowers users to rapidly perform three categories of tasks:
- Build: Agents can not only build rules based on a user's requirements, but even decode complex legacy macros or configuration files from other platforms and propose a natural rule language (NRL) process in Duco.
- Optimise: A simple chat interface allows users to interact with agents who interrogate rules and processes to provide detailed analysis, identify data quality trends and issues, and suggest ways to improve performance.
- Manage: Agents operate in the background to investigate exceptions, validate reference data, and identify specific risk patterns (such as low credit ratings). Beyond detection, they can orchestrate resolution by drafting emails or chat messages for a human to validate and send.
The right tool for the job
Agentic AI may have been the headliner at Innovation Day: London, but Maxfield reinforced that when it comes to development, our ethos is always about using ‘the right tool for the job’.
We’re not throwing generative AI and agents at every problem. It’s an expensive and over-complicated solution for many tasks. Other data automation approaches can often do the job better than AI.
The roadmap Maxfield presented therefore balanced high-speed AI innovation with robust, non-AI enhancements that are critical for delivering value to Operations teams.
Putting you in control
Our agents can take actions independently, such as drafting emails to custodians or creating chat messages for trading desks. But our AI ethos is that there must always be a human in-the-loop and that agents should operate in a contained, controlled environment.
Maxfield explained three key aspects of our governance strategy for agents:
- Containment: The agentic workspace stops at the edges of the Duco platform. Our agents can’t access any of your internal systems.
- Secure protocols: Instead of integrating directly into your network, our agents harness specific protocols, such as APIs, to pass information to other agents in your network.
- Human-in-the-Loop: The system is designed to facilitate action, not execute it. We use AI to propose a course of action, such as a draft message or reconciliation process, which a human must review and accept.
Maxfield covered lots of ways that Duco is making it possible to bring the power of agentic AI into your organisation. It was only fitting, therefore, that he handed over to Jonathan Allen from Amazon Web Services (AWS) to explore how firms could structure their teams to thrive in an agentic future.
How to prepare your teams for an agentic future
Agentic AI has prompted a revolution in team structures. In fact, this is the number one topic for business executives right now. This was according to Allen, Executive in Residence at Amazon Web Services.
When it comes to agentic AI, he said, "what got you here, won’t get you there." He highlighted several key considerations for leaders looking to successfully prepare their organisation for an agentic future.
Time to retire the 'change vs run' model
Most financial organisations are typically set up with a ‘run’ function, that focuses on business-as-usual, and a ‘change’ function, that drives innovation. Keeping the lights on often accounts for 80% of the budget, leaving little space for meaningful transformation.
Allen argued that siloing teams in this way is fundamentally incompatible with the speed of an agentic world. To succeed, firms must dismantle these barriers and reorganise into small, cross-functional teams focused on end-to-end business outcomes.
Building an AI-augmented team
So what should these cross-functional teams look like? Jonathan outlined a ‘hybrid squad’ that brings together people with a mixture of skills and expertise, augmented by AI. For example, a typical team may comprise a product manager, a business analyst, and one or two developers, supported by multiple specialised agents.
Critical to this model is the rise of the ‘expert generalist’ - a role capable of working across multiple domains to orchestrate these resources.
The need for ‘Zero ETL’ Operations
Data is the foundation of success in an AI world. Yet most financial institutions remain hampered by information locked within legacy architectures, Allen said. Firms often rely on an enormous number of extract, transform, load (ETL) processes to put data into a usable shape. This creates bottlenecks that are stalling AI innovation. The ambition, he says, is a ‘Zero ETL’ future, where information flows seamlessly between systems.
Building trust boundaries
Just like Maxfield covered in his talk, Allen highlighted the need for clear boundaries for agentic AI, given the speed at which they can operate. He suggested that firms identify where they trust systems to act autonomously and where they require a human-in-the-loop. Firms also need to think about ‘circuit breakers’ that can stop an agent when certain conditions are met, just like the circuit breakers that suspend trading when stock markets drop too sharply during a session.
Innovation breakouts: Your industry peers share their thoughts, experiences and top priorities
Our breakout sessions gave attendees the chance to explore the latest industry developments with others in the same position. Later, moderators Dele Yussuf, Justin Hingorani, Steve Walsh and Rob Gates discussed some of the key takeaways from these valuable conversations.
The daily challenges of data
Yussuf was up first, sharing his key takeaways from the operators session, which gathered together people involved in the day-to-day challenges of managing data. The group explored three key questions:
1. What are the main challenges teams face on a day-to-day basis?
The group identified data governance and consistency as the primary hurdles. External data often lacks standardisation, while complex internal lineage obscures the 'single source of truth'. Participants shared that they also struggled with incomplete audit trails caused by offline investigations and found it difficult to distinguish between high-volume noise and genuine high-risk exceptions.
2. What initiatives are already being championed within these organisations?
While OpenAI use remains limited, Microsoft Copilot is seeing significant adoption for troubleshooting, triaging, and translation. However, a cultural gap exists, as operational staff rarely propose these new technologies to senior management. Yussuf suggested this indicates a disconnect or lack of awareness regarding available solutions.
3. Where do they see the opportunities for this technology in the future?
Participants want deeper integration between reconciliation platforms and communication tools like Teams or Slack to automate audit trails. A key opportunity is dynamic risk management. For example, rather than following a standardised schedule of risk meetings, meetings are prioritised based on high-risk areas which have been flagged by an agent.
AI desire versus delivery
Next, Hingorani and Walsh recounted their conversations with Duco business sponsors, exploring the strategic ways they are approaching AI adoption, and the barriers they’re facing.
The primary barrier to adoption is change management, creating a "hurry up and go slow" dynamic where executive pressure clashes with legal hurdles and staff resistance. As a result, Hingorani noted, many firms feel "stuck" in the same position as last year.
Firms are rethinking operating models to balance cost and speed with strict controls. Regulatory differences see US Operations moving faster than their European counterparts.
Observability in an agentic world
Finally, Gates shared the top concerns and thoughts of CIOs as we move into an agentic world.
Agentic tech remains theoretical, with implementation currently stalled. As vendors build isolated AI silos, an "orchestration gap" is forming around how to integrate them. A new framework for cross-system observability is required to track data lineage across the entire chain of interacting systems, rather than just within single tools.
2026 market outlook from Oliver Wyman
Investors continue to price in efficiency gains for the industry, even though greater volumes and inflation are pushing up the cost of doing business.
The outlook for wholesale banking has shifted from cautious to positive, characterised by strong revenue recovery and improved valuations. However, this growth brings challenges regarding cost management and the need for structural transformation in technology and Operations.
That’s according to the latest market research from Oliver Wyman, which Partner Harriet Roberts presented to our Innovation Day: London audience.
She shared several key trends and developments firms should be aware of.
The efficiency paradox
Strong headline metrics for efficiency are masking rising costs. The recent improvement in cost-income ratios is largely a byproduct of surging revenues rather than genuine operational tightening. Underlying spending on compensation, technology, and compliance continues to climb.
Investors expect firms to continue pursuing efficiency. This will only be achieved by material reductions in their underlying cost base to offset inflation and the rising expense of doing business.
Dismantling silos
Leaders are starting to focus on “horizon 2”, redesigning their operating model to deliver innovation across the group, rather than being isolated to specific divisions.
Success in this new phase depends on consolidating data into a unified view and deploying scalable platforms that serve the entire group, rather than allowing individual divisions to fund and build fragmented tools.
The accelerating AI arms race
AI development is central to how banks maintain their 'right to win' in a competitive market. However, there is a key question to be answered regarding how much of the AI potential will be delivered by banks themselves innovating internally, versus what vendors can offer.
Many leading firms are partnering with tech innovators to accelerate their innovation, but each firm needs to decide what is the optimal balance between external solutions and proprietary capabilities.
Overall, the squeeze for firms continues, with even greater expectations for efficiency and lower costs, in an environment where doing business is getting more expensive.
AI stands out as a path to successfully walking this tightrope, but first the questions of how to break down silos and create an operating model that supports widespread innovation need answering.
Innovating in times of uncertainty
Innovation is difficult at the best of times. How can you ensure you drive innovation forward during periods of heightened volatility, while still meeting the day-to-day needs of the business?
HSBC's Daniel Wright and Nomura's Chris Wells explored this with Chryse Coaching Founder and CEO Cressida Hamilton.
Here are some of the key insights they shared.
1. The impact on budget and strategy
While volatility unfortunately doesn’t result in extra budget, the good news is that it provides clarity. Our panel highlighted the opportunity presented as higher volumes place extra strain on longstanding technology and process issues. This shines a spotlight on critical areas that need fixing. Volatility does, however, shrink the time horizon for innovation - instead of focussing on three to five-year projects, leaders need to be thinking more short-term.
2. Accelerating controls fosters resilience
"Innovation without resilience is decoration". Projects that do not solve real operational problems are often the first to be cut during volatile periods. One of the priorities the panel shared was shifting toward simplification and moving controls from end-of-day (T+1) to real-time (T+0). This ensures that volume spikes can be handled without adding incremental costs or waiting overnight to spot errors.
3. Data quality is essential
Our panel noted that the industry appears to have turned a corner regarding data strategy. Leadership are reacting to external pressures like T+1 settlement and the rise of stablecoins with a renewed focus on data quality. There is also growing understanding that AI and automation solutions are effectively useless without a foundation of trusted data. The old adage of ‘garbage in, garbage out’ remains true.
4. Democratising change
Our panel suggested that volatility demands a new approach to change management. Firms can federate the responsibility for change out to smaller teams, rather than having one central group causing bottlenecks. Providing guardrails helps firms maintain control while delivering speed and agility. This approach also motivates the team, who have more chance to make an impact.
We may be living through a heightened period of uncertainty and volatility, but by keeping these takeaways in mind, you can still push your innovation agenda forward.
Key takeaways from Innovation Day: London 2025
And with that we wrapped up the Innovation Day: London sessions and moved on to networking over drinks. This gave our attendees the chance to strengthen their industry connections, while discovering the best and latest innovation from Duco firsthand in our demo theatres.
The presentations, panels and discussions covered a lot of ground, but a few key themes emerged again and again.
Data quality is essential
Innovation Day: London highlighted that the key to harnessing the exciting power of agentic AI is to focus on the basics. Trusted data is the foundation of any automation efforts. Firms can’t jump straight to deploying agents – first, they must set themselves up for success.
Stymied by silos
Many firms operate in a fragmented way, with responsibilities, insights, innovation and data confined to departments or teams. Our panellists and attendees expressed a need and a desire to bring skills, perspectives, resources and leaders together from across the organisation. This will enable firms to unlock the full end-to-end benefits of agentic automation.
Governance and guardrails
Control was an obvious topic. No one wants to let an unchecked agent run amok throughout their Operations. But leading firms are already thinking about how to solve this problem. It requires built-in guardrails that limit what an agent can do on its own. The most powerful countermeasure is to keep a human in-the-loop; agents can only suggest, not action.
Changing change management
The world is getting faster. Yet traditional methods of managing change are holding firms back from innovating to stay competitive. Change needs to be controlled, but teams need to be empowered to affect more change themselves. Only then can firms realise the benefits of the latest technology innovations.
We still have plenty more of the great strategies, insights and best-practices shared by our expert speakers and industry practitioners to share. Make sure you’re subscribed to our newsletter so you don’t miss out as we bring you essential understanding on the hottest topics in Operations.
