For the past three years, the conversation around artificial intelligence in financial services has been dominated by technology. Every week seems to bring another breakthrough, another model or another promise of greater efficiency.
Yet amid all the excitement, one question has often been overlooked. Just because we can automate something, should we?
The FCA’s Mills Review is important because it shifts the debate. Rather than asking whether AI will transform financial services, it accepts that it will. Artificial intelligence is no longer confined to chatbots or helping staff write emails. It is beginning to influence underwriting, fraud detection, customer service and operational decision making.
The regulator recognises this reality and, rather than proposing an entirely new rulebook, concludes that many existing principles such as Consumer Duty, operational resilience and the Senior Managers regime remain fit for purpose. The challenge is applying them to a world where people increasingly supervise AI rather than perform every task themselves.
This matters because the next phase of AI adoption is unlikely to be won by the firms with the cleverest technology. It will be won by those with the strongest governance.
For years, organisations have understandably viewed AI through the lens of productivity. How can we process applications faster? How can we automate repetitive tasks? How can we reduce costs? Those remain important questions but are not the whole picture.
Customers will not choose a lender simply because it uses AI, and regulators will not judge firms on how sophisticated their models appear. They will judge them on the quality, consistency and explainability of the decisions they make.
Can a firm explain why an outcome was reached or how it delivers appropriate oversight? Can it show where human judgement has been correctly and appropriately applied? Traditionally, in the mortgage industry, underwriters have spent a remarkable amount of time reconciling information rather than assessing risk. Payslips are checked against bank statements, tax returns against declared income and identity documents against application forms. Essential work, certainly, but not underwriting in its truest sense.
At Ohpen, our Collect & Conclude capability uses AI to extract, validate and reconcile information from multiple documents and trusted digital sources before an underwriter begins their assessment. It is not replacing professional judgement. It is ensuring that judgement starts with better information.
Artificial intelligence is not making the lending decision. It is making better decisions possible. As routine administration becomes automated, human expertise becomes more valuable, not less. The underwriter’s role evolves from processor to decision maker. Their time is spent considering exceptions, applying experience and exercising judgement rather than chasing paperwork.
Ultimately, that is where the future of financial services lies. Not in autonomous systems operating independently of people, but in organisations that understand where automation creates value and where human wisdom remains indispensable.
Technology will continue to evolve at extraordinary pace and so governance must evolve with it. Using it well, will earn trust and that has always been the ultimate competitive advantage.
