You can use the search function to find a range of UK Finance material, from consultation responses to thought leadership to blogs, or to find content on a range of topics from Capital Markets & Wholesale to Payments & Innovation.
13 Oct 2026
The opinions expressed here are those of the authors. They do not necessarily reflect the views or positions of UK Finance or its members.What are firms prepared to let these systems decide?
Recent Experian research points to a striking gap. Consumer adoption is ahead, with many consumers in the UK already using LLMs for financial advice. But across 800 of our clients, only 2% are comfortable with fully autonomous AI decisions.
That’s not a technology problem. It’s a trust problem. And it won’t be solved by better models alone.
Compounding this, the announcement earlier this month that the UK is open to regulating advanced AI models has further raised stakes. For financial services firms, these signals suggest that while policymakers remain committed to supporting innovation, they also recognise the risks and need for stronger safeguards.
Instead of viewing potential regulation as a barrier, I’d suggest it’s a reality check. An opportunity for ecosystem collaboration to get advanced AI right.
Agents change the stakes
The first wave of enterprise AI was integration. Success was deploying a chatbot, automating a process, or generating insight from data. A person reviewed the output before anything happened.
Agentic systems have shifted success from outputs to outcomes: coordinating workflows, drawing on multiple information sources, and acting on decisions. As firms grant agents greater autonomy, the consequences of getting it wrong become material.
That shifts where the real work sits. What determines how fast and how safely agentic AI scales is largely invisible: data foundations, identity frameworks, permissioning, governance, audit evidence, and the codified context and semantic layer an agent needs.
Model performance is the part everyone can see. It is not the full story.
No single firm solves this alone
Agentic systems don’t operate purely inside organisations. For example, an agent acting on a customer’s behalf may touch a bank, a bureau, a payment network and a technology provider across one journey. As systems become increasingly interconnected, trust, not just capability, becomes the unlock.
Because of this, successful deployment of agentic systems will depend on common approaches. Financial institutions, technology providers, regulators and industry bodies all have a role to play to set standards together, fast enough to unlock innovation while ensuring consumer protection. This could explain why the policy conversation is shifting and formalising.
Trust is designed within, not bolted on
The firms that lead the agentic era will not simply be those with access the newest models. They will be those that have invested in the foundations for trusted autonomous decision-making. They’ll demonstrate to customers, boards, auditors and regulators that their systems behave safely, transparently and consistently, and produce the evidence to prove it.
Agentic AI will scale at the speed of trust. Whether the UK sets that pace or follows depends on how quickly we build those foundations together.
13.10.26
Laurie Schnidman, Chief Product Officer, AI Ecosystem, Experian UK
Join our flagship annual congress uniting leaders to combat economic crime, fraud, sanctions breaches and corruption on 3 December.
By downloading this document, you understand and agree that any sharing, distribution or republishing of the content, without prior written authorisation from the author or content managers at UK Finance, shall be constituted as a breach of the UK Finance website terms of use.