For brokers facing AI competition, adverse credit is proving to be a corner of the market worth knowing well
The rise of artificial intelligence (AI) in financial services is forcing small mortgage brokers to rethink their value proposition. For some, the answer lies not in competing with automated tools but in moving deliberately beyond their reach — into the complex, nuanced cases that algorithms struggle to resolve.
Stuart Phillips (pictured top) of AALTO Mortgages is one broker taking that approach. Faced with growing awareness among clients of AI-driven mortgage tools, his firm has pushed further into specialist niches, including adverse credit lending — a corner of the market where human judgement, lender relationships and case-by-case analysis still determine outcomes.
"More and more of our customers are referencing AI tools in their discussions with us, which is a concern, and it's clear that large firms and aggregators are looking at ways of leveraging this," Phillips said. "As a small broker, for us the answer is to move further into more specialist niches and stepping outside our comfort zones."
The strategy has produced results that surprised even Phillips himself. Working through cases that initially appeared straightforward but carried hidden complexity, the firm found that specialist lenders — particularly smaller building societies — were offering terms significantly better than expected for clients with adverse credit histories.
The gap between what AI tools tell clients and what a specialist broker can actually achieve is, Phillips argues, where the profession's future value lies. Clients are arriving at appointments having already consulted automated platforms, often with a distorted picture of their options. Many have then been turned away by high-street banks, leaving them discouraged before a broker has had the chance to assess the full picture.
"For clients that have received vague and generic information through AI portals, and have then had an uncomfortable and disappointing appointment with their high street bank, picking up the phone to a broker who can acknowledge the challenging nature of the situation but is comfortable and confident in navigating to a solution it is going to be worth a great deal to those clients," Phillips said.
The limitations of automated advice are most visible in adverse credit cases, where outcomes depend on variables that resist easy categorisation — the age of a default, whether an arrangement to pay has been honoured, the conduct of an account in the months preceding an application.
Phillips cites a recent remortgage case in which a client appeared, on paper, to be a strong applicant: a thriving business, good equity in the property and an active credit profile. Closer examination revealed an unresolved default with a catalogue lender, the result of a credit limit being withdrawn without warning — a practice that, Phillips notes, generated widespread complaints in consumer forums.
The high street was closed to the client. Phillips placed the case with Hinckley and Rugby Building Society at a rate approximately 0.75 percentage points above comparable mainstream deals — a workable outcome that a generic AI tool, focused on surface-level eligibility, would have been unlikely to identify.
That gap between algorithmic output and specialist knowledge extends to loan-to-value (LTV) thresholds. Phillips found that several smaller building societies were prepared to lend at 85% LTV to borrowers who had recently completed a debt management plan or carried defaults older than 12 months — figures well above what his firm had previously assumed was achievable for clients with that level of adverse.
The commercial case for homeownership, even at above-mainstream rates, also runs counter to the picture many clients form from generic online sources. With private rents forecast to rise by between 5% and 9% annually in some areas, Phillips argues that securing a mortgage now — even at a premium — leaves clients materially better off within a few years, as equity builds and rates are expected to come down.
For smaller brokerages, the AI challenge is therefore also reframing what specialism means in practice. It is less about retreating from technology and more about identifying the territory where technology consistently falls short. Adverse credit, with its dependence on lender-specific criteria, account conduct narratives and individual circumstance, is one such territory. There are likely others.
Whether the broader profession moves in the same direction remains to be seen. What Phillips's experience suggests is that the brokers best placed to withstand automation are those willing to take on the cases that automated tools are most likely to mishandle — and to be confident enough in their own knowledge to push back against the low expectations those tools can leave behind.
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.
