By Andreea Pleşea, PhD, Co-Founder and COO, Druid AI.
After billions invested in AI, the biggest challenge facing UK banking isn’t adoption. It’s knowing where AI actually creates value.
Despite five UK prime ministers in the past five years, the message from the UK government, technology providers and financial institutions has been remarkably consistent with the commitment to making AI a core part of the UK’s economic and digital infrastructure. Investment is being directed towards AI infrastructure, skills development and enterprise adoption as the UK forges ahead to establish itself as a global leader in AI.
The UK banking and finance sector has spent the last few years building AI strategies, identifying hundreds of potential use cases and experimenting with everything from copilots and chatbots to customer-facing assistants and productivity tools to understand where AI could be used.
A costly mistake is assuming that the value of AI comes from the number of use cases financial institutions can identify, rather than the operational problems they can solve.
What the leaders of UK financial institutions need to identify as we enter the next phase of AI adoption is where AI should be used.
Recent data from enterprise AI deployments revealed that in live environments, almost 90% of AI interactions are concentrated around just three workflow categories, with access requests, authentication and account servicing activities accounting for the vast majority of usage.
The data suggests that customers and employees repeatedly encounter the same operational friction points, and the biggest gains often come from solving the most common problems.
Where is AI creating real value?
This is particularly important at a time when banks are under pressure to balance efficiency with customer experience. Accenture’s Banking Trends 2026 report argues that an overemphasis on cost optimisation can create fragmented experiences that ultimately undermine customer value. It’s not as simple as deploying AI more widely, leaders of financial institutions need to ensure it improves the journeys (or workflows) that customers and employees rely on most.
AI agents create value not by replacing entire departments or automating everything in an attempt to transform banking overnight, but by removing friction from high-volume, repeatable journeys that occur thousands of times every day.
The institutions most likely to see strong results from AI deployments will take the approach of starting with operational demand, identifying the journeys that generate the highest interaction volumes, create the greatest friction or consume the most employee time. They should focus on solving those problems first before expanding into more ambitious use cases.
This approach is relevant as AI begins to move from experimentation into operating within core business-critical infrastructure. The Bank of England has highlighted the growing importance of understanding how AI is used within core financial processes and the operational risks that emerge as adoption scales. Success will increasingly depend on governance, resilience and the ability to integrate AI safely into existing systems rather than simply deploying more tools.
Just as banks would never deploy critical technology without understanding the underlying business process, they cannot expect AI to deliver value simply because it has been implemented. Financial institutions need a clear understanding of where work happens, how decisions are made and which interactions genuinely matter.
AI Orchestration vs Automation
The workflows that generate the greatest value to the business rarely sit within a single application. Customer servicing journeys often span core banking platforms, CRM systems, contact centres, digital channels, authentication services and compliance processes. Employees frequently move between multiple systems to complete what appears to be a simple request. Adding an AI tool into an already fragmented environment does not necessarily solve the problem – in fact, it can add another layer of complexity.
The organisations seeing the greatest success with AI agents are focused on orchestration rather than automation alone. Instead of using AI to answer questions, they are using it to connect systems, complete workflows, and remove unnecessary steps from customer and employee journeys. Indeed, research carried out by the IBM Institute for Business Value showed that across finance processes, the strongest returns appear where AI is embedded into redesigned workflows – not layered onto existing processes.
Customers do not care whether a bank has deployed the latest AI model. They care whether their issue can be resolved quickly and accurately.
Employees do not care how many AI pilots are running across the organisation. They care whether they can complete tasks without having to switch between multiple applications to find an answer to a problem.
The rewards of AI adoption for UK finance institutions will belong to those organisations that recognise this difference.
Where finance institutions will see value
The finance institutions that get the greatest value from AI won’t be the ones running the largest number of pilots or experimenting with the most use cases. They will be the institutions that identify where customers and employees experience the greatest friction and use AI to remove it.
That requires intelligent orchestration across core IT systems, processes and data that already exist inside the organisation. As AI becomes part of the UK’s digital infrastructure, AI in banking will not be defined by how many interactions can be automated – it will be defined by how many customer and employee journeys can be completed securely, accurately and from start to finish.