Artificial intelligence (AI) will help the mortgage industry, but there are questions around what the regulator will accept, securitisation and entry-level jobs.
Speaking at the British Specialist Lending Senate, Steve Cox, chief commercial officer at Fleet Mortgages, said: “I think there is a natural limit to what AI will do in our industry, and that will continue to evolve. It will be: how do we all use our data, whether you’re on the advice side or the lender side, so the data you’ve got, what does that tell you, for example, about arrears? What didn’t we do?
“Is there anything we didn’t do when we underwrote the case for those arrears to manifest themselves? How do we change products? What products don’t we have? The AI says there’s a gap there. Maybe you should think about that. Whether you do it or not is a completely different decision, but it will support, not take over.”
He questioned how far the regulator would “allow automated decisions to be made by something that many don’t really understand. I don’t know how far that would go. I’d say there’s always going to be a limit to that”.
Cox noted that another potential consideration for AI in mortgage lending was with securitised markets, which are often used by non-bank lenders as a source of funding.
“With securitised markets… if you were trying to get a book away and were asked: ‘Well, what’s the back book profile of all these landlord customers? We don’t really know. It was underwritten by AI.’ I think it’s unlikely markets would accept that scenario,” he added.
Shaping the mortgage market around today’s first-time buyer
Sponsored by Newcastle for Intermediaries
Tony Hall, head of business development at Saffron Building Society, mentioned speaking to tech firms that said AI can underwrite vanilla cases at the moment and there is the possibility that it could underwrite specialist cases in six months.
“I think that will happen and I think it is all about that efficiency… So if AI can support getting quicker decisions and getting better outcomes for people sooner to allow the specialist underwriters to underwrite the right cases and have the time and make the right decisions, then that’s something that we should embrace, because you’ve got to do it carefully,” he said.
AI can lead to better customer outcomes
Buster Tolfree, director of mortgage at United Trust Bank (UTB), acknowledged that AI was a “buzzword and zeitgeist thing… that everyone’s talking about, but it’s just to facilitate us doing more”.
He continued: “A really good example of where we’ve deployed it at UTB is all of our calls are listened into via AI. If you’ve got one of our agents who’s taking a call from someone and they say ‘suicide’, that flags up to their team leader, who can then walk over and make sure that the actual agent handling the call is okay, or it gives the person prompts and it automatically transcribes all the calls.
“These are things that can facilitate a better customer outcome. Thinking with a Consumer Duty hat here, a better customer outcome is not just about replacing bodies, it’s actually about allowing us to give a better-rounded service through the lifecycle, not just up to completion.”
AI could impact entry-level jobs
Tolfree said there was a concern with AI being able to do administrative tasks regarding how this would impact entry-level jobs.
He explained: “I worry a bit about entry-level jobs more than whether it is going to hugely disrupt our businesses. If you think about big broker firms, fact-finds used to be done by a human being running through a bunch of questions and then passing over to an adviser to give the advice.
“All of those questions can be taken through different automated mediums now, so how does someone get into a broker firm if the first entry point is you need to be a CeMAP-qualified adviser? That’s probably where it’s heading.
“I think we need to keep an eye on that as the wider financial services community of: How do people get in if we’re automating away all of these easy jobs?… Where’s the next generation of people coming through?”
