This session puts the gap between stated ambition and operational reality on the table: what's shipped and adopted versus what's still stuck in pilot, and how leaders are honestly measuring progress when there's no agreed industry yardstick. Going into 2027, the question that matters isn't whose roadmap looks best on paper, but rather who’s putting it into action.

The shift from "AI that chats" to AI that executes specialist analytical work to a defined standard sounds like a trust upgrade. But when a bank's credit analysis, peer benchmarking, or due diligence runs through a third party's encoded framework inside an agent, the accountability question doesn't go away, it just moves. This fireside chat looks at what it actually takes to deploy specialist AI skills in a regulated environment, and who's on the hook when the output is wrong.

AI is only as valuable as the decisions businesses can trust it to make. Yet organisations are discovering that having more data, or even better data, isn’t enough. AI also needs to understand what that data means, who is accountable for it and the rules governing how it can be used.

This session examines how organisations can close that gap, turning fragmented enterprise knowledge into a foundation for trusted AI. We’ll explore how better context can reduce risk, accelerate decision-making and help businesses turn AI experimentation into dependable, enterprise-wide value.

Most institutions can point to agents handling individual tasks, but far fewer have handed one an entire process. Owning a process end to end means managing exceptions, judgement calls, and the handoffs between systems and people, not just executing a step, and that's where most agentic pilots quietly stall. This session gets into what must change, technically and organizationally, before an institution can make that shift, and where it's likely to break first.

As AI agents become more capable and autonomous, trust cannot be something financial institutions assess after the fact. It has to be built into how agents operate. For highly regulated institutions, that means creating a control plane that governs how agents, people and systems interact, and provides the confidence to put AI to work in increasingly consequential processes.

This panel explores how orchestration can provide that control: defining the boundaries of agent autonomy, turning policies into enforceable guardrails, retaining meaningful human oversight, monitoring execution, and creating a traceable record of what happened, why and under whose authority. As AI capabilities accelerate, what will it take to ensure financial institutions can innovate at the speed of trust?

Each table is focused on a different topic and led by its own moderator. After a quick introduction from the Chair, you’ll join your chosen table for a focused, hour-long conversation. There’s no feedback to the main room, so all your time goes into meaningful discussion. Each table has prompt questions, which moderators can adapt as the conversation evolves. It’s an informal dialogue with a purpose, ending with clear takeaways created by your group.

In this session, enterprise executives will present rapid-fire case studies highlighting tangible returns from AI initiatives. Each 10-minute presentation will showcase real-world examples of how organizations achieved measurable business outcomes.

Financial services have invested heavily in AI. Now comes the harder question: what is it delivering, and who ultimately benefits?

This panel examines where AI is generating measurable returns, which use cases are surviving scrutiny, and how leaders are defining value across productivity, revenue, risk and customer outcomes. Can AI simultaneously improve the economics of financial services and make them more accessible, useful and effective for the people they serve? And what separates investments that create lasting value from expensive experiments that stall?

AI features are everywhere; AI products that customers rely on are not. This opening keynote panel focuses on what’s actually shipping: the AI-driven products and workflows that are live, scaled and moving real business metrics. Here we look at concrete examples of what made it from pilot to product, what got killed along the way, and the patterns that separate durable products from proof-of-concepts.

The institutions furthest along have found that the work itself has to be rebuilt. Automating a task leaves the process intact, and speeding up something badly designed just gets you to the wrong answer faster.

This session looks at what changes when AI moves from tool to operating model: which processes get redesigned rather than accelerated, how team structures and spans of control shift, and what a manager does differently when part of the team isn't human.

Risk frameworks in financial services were built on assumptions that AI has quietly broken. Model risk assumed predictable behavior. Third-party risk assumed you knew what a vendor's software did. Operational risk assumed a person somewhere in the chain.

This session looks at how firms are rebuilding the risk function for an enterprise that runs on AI: which categories no longer hold, what new roles and reporting lines are appearing, and how to keep review from becoming the thing that blocks everything else.

Shipping an AI product is only the beginning. Once live, its economics keep moving.

This fireside explores how the cost and value of AI products evolve in production — from inference and infrastructure to pricing, usage and margins. As models improve, costs fall and customer behavior changes, how should leaders build products whose economics can evolve with the technology?

In financial services, AI can only move as fast as the confidence to use it. Unclear mandates, cultural resistance and fear of getting it wrong can stall adoption long after the technology is ready.

This panel explores how leaders are bringing the enterprise along: setting clear direction, reshaping how people work with AI, and defining risk boundaries that give employees the confidence to move from experimentation to everyday use.

As the AI ecosystem evolves, competitive advantage depends less on owning every capability and more on knowing what to build, what to buy and who to build with. But buying AI is rarely just a technology decision: procurement, risk, legal, security and business teams all have a stake, and promising projects can stall long before a contract is signed.

This session explores how financial institutions are making and executing those bets. We look at where ownership still matters, where partnership creates advantage, and how the right mix of internal capability and external expertise shapes the long game. We then go inside the internal journey from interest to approval: where deals get stuck, how stakeholders build confidence in unfamiliar technology, and what buyers and vendors can do to move from promising conversation to successful deployment

Some of the forces that will reshape financial services still look comfortably distant. They may be closer than they appear.

This presentation looks beyond the obvious AI trends to the shifts gathering momentum across technology, economics, regulation and competition; identifying what could arrive faster than expected and the strategic decisions leaders may need to make sooner than planned.

Two days of predictions, investment and transformation plans rest on assumptions about where AI is taking financial services. But which of those assumptions are we least justified in making?

This closing conversation challenges the prevailing consensus: what the industry may be overestimating, what it could be missing entirely, and which decisions become very expensive if today’s assumptions about AI turn out to be wrong.