3 Aug 2026

The framing of AI agents as an 'overlay' on existing core banking infrastructure is seductive. Drop the clever orchestration layer on top of your legacy stack, automate the messy middle of lending decisions, and watch your operating costs fall. The pitch makes sense on paper. The execution is where it gets uncomfortable.

The real question for UK consumer credit isn't whether agentic AI can automate loan origination workflows. It can, and it will. The question is who owns the decision when the agent gets it wrong.

The FCA's Consumer Duty doesn't care about your architecture. It cares about outcomes. If an AI agent is continuously making credit decisions, adjusting affordability assessments in real time, and routing customers through different journeys based on inferred risk signals, you need to be able to explain every one of those decisions to a regulator. Not in aggregate. Individually.

Most lending platforms aren't built for that. The audit trail that satisfies a compliance team looking at a batch process is completely different from what you need when an agent has made ten thousand micro-decisions overnight and a customer complains they were declined unfairly.

Two things need to happen before UK lenders treat this as production-ready rather than a proof of concept:

• Explainability at the decision level, not the model level. Knowing your AI is generally fair is not the same as knowing why this customer got this outcome.
• Human escalation paths that are genuinely used, not just documented. Governance theatre will not survive an FCA review.

The fintech vendors selling agentic orchestration are, mostly, selling efficiency. That is a legitimate value proposition. But efficiency built on governance you can't demonstrate is a liability dressed up as progress.

I'd rather see one well-governed agentic workflow in production than ten pilots that look impressive in a deck but would fall apart under scrutiny. The institutions that figure out the accountability layer first will have a genuine advantage. Everyone else is just accumulating technical debt with better branding.

The interesting tension is whether the FCA's pace of guidance on AI decision-making will move fast enough to give lenders the clarity they need to actually commit.

The US is stalling on open banking mandates, and that should matter to every UK fintech leader who has spent the last five years complaining about the pace of our own regulatory change.

Because here is the actual situation: while the Consumer Financial Protection Bureau's Section 1033 rules sit in limbo, US banks are still running on screen scraping and bespoke integrations that cost a fortune to maintain and break constantly. Global fintechs building account-to-account payment products have a genuine choice about where to prioritise their next market. The UK, with its established open banking infrastructure and instant payments rails, is structurally more attractive than the US right now.

That window will not stay open indefinitely, but it is open today.

The angle that UK leaders should take seriously is not the threat from the US catching up. It is the opportunity to cement network effects before they do. The fintechs that build customer bases, distribution partnerships, and brand recognition in the UK and Europe over the next two or three years will be far harder to displace when the US eventually sorts itself out.

• Account-to-account payment volume in the UK is growing, and processing costs are a fraction of card rails
• Our standardised APIs mean a fintech can integrate with multiple banks without a bespoke deal for each one
• Regulatory consistency, even when it feels slow, beats regulatory uncertainty for anyone trying to plan a product roadmap

There is also a less comfortable observation here. UK consumer credit specifically has been slow to move payments and data infrastructure away from legacy approaches. Plenty of brokers and lenders are still relying on outdated affordability checks and card-based disbursements when open banking alternatives exist and are cheaper.

The US delay does not automatically benefit anyone. It benefits the UK fintechs and lenders that actually move.