Money20/20 Europe 2026 and the governance question finance can no longer avoid.

If 2026 is the year AI moves from support tool to autonomous actor, finance faces a sharper question than productivity. Beyond AI’s ability to accelerate decisions, the real challenge is preserving accountability, auditability and trust when machines transact, select and execute at scale.

What happens when AI acts, not asks?

That question, placed at the centre of Money20/20 Europe’s 2026 agenda, captures a real shift in the market. AI is moving beyond drafting, summarising and assisting. It is entering workflows where it may initiate actions, connect systems and influence outcomes before a human has reviewed every step. For financial institutions, this is not simply a technology upgrade. It is a strategic governance challenge: the ability to evidence control must evolve with the speed and autonomy of the systems being deployed.

Many control environments assume that systems wait for human instruction. Agentic AI follows a different operating logic: it proposes, selects, routes and acts. This shifts the board-level conversation from how the organisation uses AI to how it can demonstrate control when AI becomes part of the decision chain.

From assistance to agency

As AI takes on a greater role in decision-making, financial leaders must ensure their organisations can explain what happened, why it happened and who remained accountable. When automated systems influence payments, onboarding, fraud monitoring, customer support or credit decisions, leaders cannot add accountability after deployment. They must build it into the system from the start.

That requires clear lines of authority, strong evidence trails and a precise understanding of which decisions teams can reverse and which may cause material harm. Calling agentic AI a digital intern may make adoption feel more approachable. Treating it as a decision-making system creates a more honest foundation for oversight.

But when machines transact, trust becomes the real battleground

Trust as transaction architecture

This is one of the most important ideas emerging from the 2026 agenda. Trust has moved beyond brand promise or customer experience objective. When AI agents shop, pay, authenticate or act on behalf of users, trust becomes a transaction design problem.

If the actor initiating a transaction is software, identity, authorisation and intent must be verifiable before the transaction moves. In that world, trust is not what an institution says about its systems. Trust is what can be tested under pressure by risk teams, auditors, supervisors and, when necessary, courts.

The better question for leaders is therefore not “Is this AI impressive?” It is: “Can this action be explained, challenged and reconstructed?”

Fraud in the age of believability

Money20/20’s language around AI-enabled deception is particularly relevant because it frames fraud as a systemic risk. It points to attackers industrialising deception at scale: deepfake audio, social engineering, automated impersonation and more convincing digital traps.

This changes the economics of fraud. When deception becomes cheaper, faster and more believable, controls designed for a world of expensive impersonation start to weaken. Voice, face, familiar behaviour and trusted communication patterns require stronger supporting evidence.

The response must be smarter than adding friction. It should strengthen protection without undermining trust, inclusion or customer experience. More checks may reduce losses, but they can also create barriers for legitimate users. The stronger path is to make transactions more self-describing, with clearer provenance, clearer authority, clearer boundaries and evidence that can be reviewed after the event.

Regulators are no longer lagging innovation; they’re accelerating it

Regulation becomes part of strategy

The 2026 Money20/20 agenda places regulation beside AI, not behind it. That matters. In financial services, regulation is no longer only a constraint on innovation. It increasingly shapes the market conditions in which innovation can scale.

For AI, this has a direct implication: governance has to move from policy language to operating model. Organizations need accountability maps, approval logic, model oversight, incident pathways and evidence trails that are meaningful outside the innovation team.

A useful provocation for any executive team is simple: an AI system should be able to produce the artefacts a supervisor would reasonably ask for. Until then, it remains closer to unmanaged risk than strategic capability.

The underestimated risk: convergence without clarity

The most interesting part of the 2026 debate is AI operating inside rebundled ecosystems, rewired infrastructure and faster-moving regulatory environments. That convergence is where many failures will occur.

The industry likes the language of seamlessness. Yet seamlessness often means that risk is spread across more parties, more dependencies and more opaque decision chains. If an AI agent acts across several services, who is accountable when something goes wrong: the institution that deployed it, the institution that received the transaction, the infrastructure provider, or the ecosystem that enabled the behaviour?

This is where leadership needs to become more disciplined. Momentum is not strategy. A credible AI strategy should be able to say what must be provable, who is responsible and how evidence will survive across the full chain of action.

The financial world isn’t changing; it’s already been rewritten

Why Amsterdam, why now

Money20/20 Europe 2026, taking place at the RAI Amsterdam from 2 to 4 June, brings AI, payments, regulation, fraud, infrastructure and executive decision-making into the same conversation. Its value lies in the environment it creates: a place where leaders can test assumptions before those assumptions become embedded in systems, policies and market infrastructure.

For those attending, one practical question may prove more useful than a long checklist: what evidence would convince a sceptical auditor, regulator or public stakeholder that an AI-driven change improves outcomes while protecting rights, safety and resilience?

When leaders can answer that question clearly, ambition begins to take the shape of strategy.

As AI systems gain greater agency, leadership will depend on the ability to move quickly without losing control. The leaders who matter will understand how to remain accountable when actions compound, decisions move across ecosystems and trust must be demonstrated through evidence.