Nicholas Mendes warns the industry risks sleepwalking into unsustainable AI dependency
Mortgage brokers rushing to adopt AI to cut costs may be storing up a far larger bill once platform providers raise their prices when dependency is locked in.
Nicholas Mendes (pictured top), mortgage technical manager and head of marketing at John Charcol, argues that brokers adopting AI without scrutinising long-term costs are making a short-sighted calculation – one the industry may come to regret.
What role does AI play for brokers today?
Mendes acknowledges the genuine short-term utility of AI within the broker market. Used well, he said, it reduces time spent on administrative tasks and helps manage the client journey through automated trigger points.
"AI certainly helps when it comes to more of the mundane tasks, such as keeping clients up to date with regards to different stages of the application," he told Mortgage Introducer. "As an example, you could use AI to act like a client relationship manager for trigger points, whether it’s solicitor landmarks, valuations, and things along those lines. It just gives brokers a bit of time back, especially when you're dealing with a lot of clients."
He also pointed to emerging platforms that go further, performing product sourcing and recommendations once fact-finding is complete. Yet he is cautious about how far that capability should reach. "My hesitation with all of that is that, while that might be good in one respect, for certain brokers that might see it as a way of saving time, I always worry that what you don't want to do is sort of take away the essence and the art of brokering," Mendes said. "You don't want it taking away what a broker is fundamentally there to do."
Is AI a threat to the broker's role?
Looking further ahead, Mendes believes the competitive dynamics of the mortgage market will intensify. He sees disruptors emerging who will use AI to capture consumer relationships before a broker – or a rival lender – can.
"There is no doubt that there'll be disruptors in the market in the future that will look to use this space to really dominate it and make it harder," he said. "I think the idea that we're going to be in a comfortable place and just have the benefit of AI just supporting us is a really short and narrow vision."
He points to the changing nature of client enquiries as evidence of a shift already under way. Consumers are increasingly going to platforms first, inputting their scenarios, and being matched to a broker, rather than seeking one out directly. In that environment, he believes guardrails and regulation will be essential to prevent platforms from recommending lenders outright, bypassing broker advice altogether.
The broker's response, in Mendes's view, should be to evolve rather than resist. "I certainly see the role of a broker is going to be more of a role of an adviser than a broker, and we'll be talking more about pensions and other areas – equity release, as an example. It just means that the relationship with a client is more in-depth than just being standalone in one area."
What happens when the Uber effect kicks in?
The sharpest part of Mendes's argument concerns cost. He draws a direct parallel with the ride-hailing market, warning that AI platforms are following an identical playbook – price low to drive adoption, then raise costs once dependency is established.
"You do it really cheaply and as more and more people use it, you then increase your costs when people are dependent on it," he said. "There's only a matter of time as more people become dependent. It's like the Uber effect."
His concern is that brokers and firms replacing headcount with AI tools may be trading a known salary cost for an unknown – and ultimately larger – technology bill. "In five years' time, the cost could be so much more greater because they've got to profit," he said. "By which point, how much have you gained as a result of adopting AI?"
The answer for Mendes is not to avoid AI, but to approach it with discipline. He invokes the second world war concept of survivorship bias – the tendency to make decisions based only on visible data, while missing what is absent – as a framework for thinking about where AI genuinely fits, rather than forcing it into a process because it is available.
"You have to look outside of that," he said. "I think a lot of people try to find the solution first. It's actually trying to find where AI could fit, rather than where it doesn't fit. And that's what I think people just need to do."
Want to be regularly updated with mortgage news and features? Get exclusive interviews, breaking news, and industry events in your inbox – subscribe to our FREE daily newsletter. You can also follow us on Facebook, X (formerly Twitter), and LinkedIn.
