Benjamin Blyth, director of Houz Mortgages, has founded CriterAI, a mortgage criteria sourcing system that allows brokers to search products using natural language.
Confronting broker frustrations
Blyth said advisers already primarily communicated through WhatsApp and wondered if it would be useful to consolidate these messages into one place with the ability to receive answers.
He said the decision to build CriterAI came from “the frustrations of the day-to-day of being a broker”.
Blyth has worked as an adviser for around 10 years across mainstream and specialist lending and founded his current firm in 2022.
Click here to view our Sponsored Content Hub
He recently built his own CRM system and helped other brokers do the same on a consulting basis. Blyth said he had a natural interest in technology and began considering how AI could support an adviser’s daily activities.
Working on CriterAI helps Blyth to “flex his muscles” in technology and still do his day job. He said it was early days and there was a lot the company could bring to brokers.
His CRM took him three years to build, born out of his dissatisfaction with the options on the market, and over that time, it has been improved and developed.
He said: “Having developed that, it was a natural progression to fill this pain point on criteria. Actually, I started building it to sit in my own CRM, then decided to build it for a wider audience.
“The frustrations that led to it were how lengthy it can be to do a bit of criteria research, especially if you’re using more than one system to verify it. Equally, you can search for something on one criteria system, and it won’t be there on another because one system uses one lender terminology and the other uses another.”
Blyth added: “I had an experience a few years ago where I searched for Airbnb because a client of mine wanted to let out an annexe. I spelt it as per the brand and it didn’t come up. But when I typed ‘A, I, R’, it autofilled the rest for me and showed a completely different way of spelling it, which was A, I, R, space, B, ampersand, B.
“That’s just not the brand. No one’s going to search for that, so it won’t get found. At the time, I notified that particular provider and they changed it.”
Conversely, CriterAI can interpret what a user means.
“If you spell it wrong because you’re not familiar with the brand, it will know that’s what you’re looking for,” Blyth added, saying this was the beauty of natural language AI technology.
How CriterAI works
CriterAI is connected with lenders in the same way as traditional sourcing systems, allowing firms to provide accurate, up-to-date product information.
For broker users, the interface is similar to ChatGPT, where type queries can be typed in using natural language, “in the same way you might text your BDM or drop them an email”.
CriterAI launched with residential only, and buy to let (BTL) will be launched soon, with bridging pegged to launch in March.
“[Bridging is] an undoubtedly under-served market, so I think that will be quite an advantage,” he added.
CriterAI addresses the challenge of speed and has an average response time of 12 seconds and results are returned without having to enter specific keywords. The platform is also ad-free and works on mobile.
A broker-built technology ecosystem
Blyth said he had plans to build more technology systems for advisers, saying his main focus this year was CriterAI, so he was hiring someone to replace him at Houz.
He said he wanted more people to use the platform so they could help develop it and he could research how to build on it, which would need the funding of paying users of CriterAI.
“I have dreams of building an AI-powered affordability calculator towards the end of the year… with features such as being able to drag and drop documents safely and have it read them for you,” Blyth said.
He is also building a suite within CriterAI that will enable brokers to use AI as an assistant.
He wants CriterAI to be a trusted brand for residential and BTL mortgages before he brings the affordability aspect in. Additionally, Blyth has more planned for other gaps he’s identified in technology, “but that will rely on much more money than this does, in terms of investment, and therefore this needs to be successful”.
What are tech firms getting wrong?
When asked where existing tech firms were missing the mark and whether brokers should take the lead, Blyth said: “I don’t think it’s necessarily time for brokers to take the reins, but as brokers on the ground, we are quite nimble. I have an interest in technology and AI, so naturally, my inquisition took me into doing what I’ve done.
“I think if you look at the big fintech companies, they’ve got systems they’ve been building for years and years, so it may be difficult for them to pivot and ditch everything they’ve built and trust something new. Or, to take a new technology and have it work seamlessly with what they’ve been building for years.”
He said this could get easier with different AI plugins, but these firms also had roadmaps that had been agreed, sometimes with investors, “and it’s very difficult for them to swerve away from that”.
Blyth said AI had come at such a rapid pace that only a few years ago, something like an AI-backed criteria search was not on a firm’s roadmap and now they probably wish it was or had been, or if it was, there would be other features prioritised ahead of it.
He said CriterAI was built in a way that allowed changes to come to market quickly, adding: “This was only a concept in June, and I launched in December. Within six months, I’ve gone from an idea in my head to a tangible product with paying subscribers who have ditched other criteria search systems in favour of ours.”
Blyth said the intention was to make something AI-powered and bring it into today’s world “because some of the traditional platforms have existed for a while and haven’t really changed”.
Essentially, Blyth wants to help brokers work faster and be more efficient.
