Why governance will define agentic AI era success for foreign banks in the UK

The contents of this blog are for general information purposes only and do not constitute legal advice. Association of Foreign Banks disclaims liability for actions taken based on the materials. Readers should consult their legal advisers.

For some firms, AI now sits at the core of foreign banking operations; what began as experimentation has evolved into everyday integration. The same is now true of agentic AI, with 70% of financial services leaders reporting that their institutions are already deploying or exploring AI agents.

Where previous generations of AI recommended actions, agentic AI executes them. But autonomy without accountability is a risk no foreign bank with headquarters outside the UK can carry; after all, they face a uniquely demanding multi-jurisdictional governance equation, as well as a raft of emerging frameworks across every market they operate in.

Oversight and accountability matter. No wonder our recent AI in Banking Benchmark revealed that the majority of bankers (89%) see a future defined by a dual workforce of AI agents and humans.

A regulatory landscape moving at machine speed

UK regulators have been clear that AI will be supervised through existing frameworks rather than bespoke AI rules, with accountability for outcomes remaining clearly assigned to human decision-makers, regardless of the degree of automation. And scrutiny is intensifying. In a June 2026 speech, the FCA set out how it’s rethinking supervision for the AI era. For foreign banks, add to this the EU AI Act and the diverse range of supervisory models globally, and governance becomes incredibly complex.

Embedding governance by design

At nCino, we believe the answer is not to slow agentic adoption but to embed governance into its foundations. So, what does a best-in-class compliant agentic AI deployment look like for an international bank?

• Role-based design: Rather than deploying a multitude of generic assistants, mature banks are building agents around specific roles – Analyst, Service, Processor, Client, Executive. Each plays their role in augmenting contextual outcomes, as well as the humans they support – giving more time to focus on judgment, oversight, and accountability.
• An orchestration layer: Coordinating agent activity around defined outcomes. Passing context from one stage of a process to the next automatically, so that information gathered during onboarding informs underwriting, and underwriting informs ongoing risk monitoring.
• A permission framework: Explainability is crucial to deploy in a regulated environment: policy-as-code guardrails, a full audit trail, chain-of-thought logging, and explainability on every agent action. An orchestrated system is built with guardrails to ensure every recommendation an agent makes can be traced, interrogated and, where necessary, challenged.
• A human core: The emerging model is a “dual workforce,” treating AI agents and humans as complementary participants in the same work AI takes on data-heavy tasks like income verification and compliance validation, while humans retain the judgement calls: credit decisions, client relationships, and the nuanced local market knowledge that sets successful banks apart. The result: underwriters spend less time assembling the file and more time analysing it.
• A tiered framework: Governance must be risk proportionate. A blanket policy that treats a document-summarising tool with the same rigidity as an autonomous credit agent will simply stall innovation. We advocate a tiered framework that distinguishes between predictive, generative and agentic AI to streamline low-risk automation whilst maintaining rigorous oversight, testing and human sign-off where decisions carry regulatory or customer impact.

For international firms, this combination of orchestrated role-based support, built-in guardrails, a dual workforce, and proportionate oversight turns governance from a brake on innovation into the very thing that makes safe, compliant, multi-jurisdiction agentic AI adoption possible.

Compliant foreign banking in a new era of intelligence. Join nCino in London on 30 September 2026.

Governance in the agentic era will be front and centre at this year’s nCino EMEA Summit. Hear how your peers are moving agentic AI from pilot to production – safely, compliantly, and at pace.

If your bank is wrestling with what AI governance really looks like in practice, you won’t want to miss our keynote speaker, Clara Durodié, AI Strategist: Risk, Governance & Geopolitics of AI in Financial Services. A globally recognised technology strategist, Oxford-educated author, EU AI Alliance member, and CEO of Cognitive Finance Group, she’s spent over two decades at the intersection of AI and financial services. Advising the boards of major financial institutions, governments, and think tanks on the strategic, governance, and geopolitical dimensions of AI adoption.   At the Summit, she’ll cut through the noise and deliver a clear view of where AI in financial services is actually heading, and how to lead that journey responsibly.

Register for the EMEA Summit here

Authored by

Ben Ussher-Stanley

Principal Commercial & Mortgage Lending Strategist, EMEA, nCino

Ben spent eight years inside banking before moving to the vendor side. That experience – managing KYC programmes, navigating credit decisions, and living inside the operational reality of a major institution – shapes everything he shares. Now nCino’s Principal Commercial and Mortgage Lending Strategist for EMEA, he works with C-suite leaders across 40+ countries on commercial and mortgage lending transformation.

Sponsored by

nCino is the platform of choice for financial institutions navigating digital transformation on a global scale. Founded to help banks digitise and reengineer their most critical business processes, nCino serves over 2,700 customers worldwide — including many of the largest and most complex financial institutions across international markets. Its intelligent, cloud-based platform is built to meet the demands of cross-border banking, supporting multiple languages, regulatory environments, and local market nuances while delivering a consistent, enterprise-grade experience. By bringing together people, AI, and data, nCino helps international banks consolidate legacy systems, strengthen risk management, and make faster, more confident decisions — wherever they operate.

You can find out more about nCino here.