A candid opening interview cutting through the hype to surface what is actually in production, what has stalled, and what the next 12 months will demand. Many institutions are still navigating the gap between experimentation and enterprise-wide deployment

After years of pilots and PoCs, AI is finally hitting the core of major banks — this session goes behind the press releases to show what is really in production today. Where are models quietly moving billions, where have they stalled, and what will be impossible to roll back by end of 2026?

As AI consumption accelerates across financial institutions, senior leaders are facing a new and urgent question: what happens when the cost of running AI outpaces the value it delivers? This session examines how institutions are starting to measure and justify AI running costs before it becomes a board-level problem.

Ignoring legacy data is like building AI on quicksand — yet most institutions can't afford multi-year rewrites. This session explores how banks are turning decades of messy historical records into a genuine strategic asset through smarter data ownership and targeted fixes.

Regulators won't green-light AI strategies, but they will evaluate how banks govern, control, test, and justify them. A senior figure from the Federal Reserve, OCC, or equivalent shares how supervisors are assessing AI risk in practice and what banks must do to make defensible decisions under regulatory uncertainty.

Model performance alone is no longer enough. Organisations must clearly evidence intent, controls, accountability, and decision-making. A financial services firm shares how it built a governance framework that satisfies regulatory scrutiny, internal audit, and board-level review.

How do you build the right guardrails so employees can use AI freely and safely, without the institution taking on unacceptable risk? Attendees work through their own governance gaps in small groups facilitated by a practitioner.

Data is cleaned, governance is in place, strategy is locked — now, with where you are, where can you go? Learn how an institution separates genuinely AI-worthy ideas from the ones better handled by simple automation.

Large institutions must decide when to build in-house, when to buy, and when to partner — and the right answer differs by function, regulator, and risk appetite. Speakers unpack why many promising AI deals fail not on model performance but on governance fit.

Three 10-minute practitioner snapshots delivering concrete, specific examples attendees can take back to their organisations. No high-level framing — just what was built, how, and what happened.

Fraudsters are using generative AI to create fake identities, synthetic profiles, and mimicked communication styles — and banks are fighting back by using AI to detect AI. This session presents a practitioner case study on emerging fraud typologies, the detection models being deployed against them, and why slow AI governance creates outsized fraud exposure.

Structured small-group discussions on themes from the morning sessions, facilitated by practitioners. Consistently rated among the highest-value elements of the programme by past attendees.

An invitation-only pre-programme breakfast exploring how diverse leadership teams build safer, less biased, and more commercially effective AI systems. Association-led.

Macroeconomic conditions, vendor concentration, cross-border regulation, and capital constraints are shaping AI decisions more than technology maturity. This panel examines how and addresses the workforce consequence question: as AI automates at scale in a consumption-based economy, what does the macro picture look like for financial services?

Most financial institutions are still in the planning or early pilot phase of agentic AI — very few have reached production. A practitioner who has shares what working agentic deployment actually required: the infrastructure, the governance decisions, and the failures.

An unsponsored editorial interview with a C-suite leader on the defining AI decisions their institution has made — including what they got wrong and would do differently.

Selecting an AI vendor is only the beginning. The real complexity starts once that model, platform, or agent becomes embedded in day‑to‑day operations, touching regulated decisions, customer outcomes, and internal accountability.

This panel focuses on what happens after a vendor is chosen: how banks operate, govern, and scale AI systems they don’t fully own, don’t fully understand end‑to‑end, and can’t easily switch off without consequences.

AI products are no longer discrete features — they are living systems powered by data loops, orchestration layers, and autonomous agents. A practitioner shares how their organisation bridged the gap between what looks good in a pilot and what can hold up at enterprise scale.

Many AI programmes die after governance approval because value can't be demonstrated — especially when benefits show up as avoided loss or resilience rather than revenue. This session examines which ROI frameworks actually work and how to make a compelling case when the tools will have evolved before budget approval arrives.

Upskilling is not a one-time event — it is an ongoing organisational capability that most institutions are still building. This panel surfaces the most practical models for building AI literacy at scale, from champion networks and Data Academies to dashboard-based adoption tracking and cash prize competitions.

Facilitated peer discussions on what success actually looks like once AI is live — covering measurement frameworks, the conditions that sustain momentum past the pilot stage, and what failure looks like in practice.

How far can AI agents go in customer service, back office, and compliance support — and what must be true before any meaningful deployment? A practitioner addresses the guardrails tension directly: the more you constrain an agent, the less useful it becomes, but without constraints it runs rogue.

The AI vendor landscape is shifting rapidly and business cases go stale before sign-off arrives — making technology decisions harder than ever. This panel addresses how senior leaders are making consequential bets when the ground keeps moving.

A closing conversation on the most advanced AI capabilities being developed today — and the considerations that have kept some of them from broad release. Includes discussion of Anthropic's Mythos Preview and what frontier AI development looks like from the inside.