AI helps, but mortgages need humans

IN Partnership with

70% of brokers are either currently integrating AI into their business or planning to do so in the future

WHETHER YOU love it or loathe it, there’s simply no escaping AI. From our professional lives to our personal hobbies, our daily routines to our night-time scrolling, there’s no area, sector or market that AI hasn’t touched in some way. And the mortgage world is no exception. Research from HSBC found that seven in 10 brokers are either currently integrating AI into their business or planning to do so in the future, with only a small number now wholly resistant to the new technology. But what role will AI play in the mortgage sector’s overall future? Chris Lynch, IT director at Source Insurance, explains that while AI can be a helping hand for some aspects of the mortgage process, it’s by no means a replacement for human expertise. “It’s fair to say that generative AI tools are becoming part of the everyday search function for customers; Google includes its own AI-generated content at the top of search results, and the use of AI chatbots as a first search by consumers is increasing,” he says. However, when it comes to generative AI tools such as ChatGPT,

Source Insurance is an award-winning intermediary with over 30 years of expertise in the UK mortgage market. Specialising in home and landlord insurance, we offer advisers powerful, intuitive quoting platforms that deliver fast and fair results. Our flexible approach puts intelligent insurance at the heart of everything we do – giving advisers the freedom to work their way. Trusted by thousands across the UK, Source combines expert guidance, dependable protection and smart technology to simplify insurance without compromise.

AI in the financial sector

of firms are already using AI

“The term used online for misinformation disseminated by generative AI is ‘slop’. This creates many problems, including the risk of widening financial literacy gaps as consumers could be bombarded with incorrect information”

Chris Lynch, Source Insurance

the fundamental question that you must always ask ought to be, “Is this answer right?” It may not be. While AI is useful for basic research, it can’t be entirely trusted without further fact-checking. Lynch points out that errors from AI outputs are labelled as “hallucinations” – with ChatGPT-5 reporting a current “hallucination rate” of 9.6%. “There are a variety of causes for this,” he adds, “from models being fed incorrect information to the fundamental way generative AI ‘predicts’ the answer to a query. If a consumer took the output from an AI-generated solution at face value, it could lead to financial and legal difficulties. Advisers are a vital line of defence against this.” Lynch explains that in many cases generative AI may try to match its output to unrelated information it already knows about you, meaning that the other risk for consumers is that, in an attempt to personalise, the system will ‘overfit’ advice to the profile it has built up of you, using inappropriate information based on previous conversations. “This is where the adviser adds true value,” says Lynch. “They can get below the surface level to truly understand the detailed requirements of each customer, specific to the situation at hand. Their experience and knowledge allow them to differentiate between accurate and up-to-date information versus hearsay when providing advice.”

Lynch’s sentiment seems to chime with the general feeling in the sector. According to research from Darlington Building Society, 68% of homebuyers would rather go ‘in branch’ for their mortgage advice, with just 2% preferring to chat to an AI assistant. Still, that doesn’t mean customers are not using AI to gather facts first. It’s clear that to both meet customer expectations and thrive in the current market, brokers need to position themselves as essential partners for AI-savvy buyersespecially as many may already feel ‘informed’ before speaking to a professional in the first place. “The human touch is everything here,” says Lynch. “Personally, I would always recommend finding out what advice the customer has been given by AI, or any other online source, and asking probing questions to validate what they think they

know and what they don’t. Don’t be afraid to do some myth busting of your own.” As Lynch explains, it’s all about asking the questions that the customer hasn’t thought to ask themselves, building rapport and joining the dots in a way that AI can’t – all of which is key to providing a superior service. “It’s also worth remembering that most people actually do want to deal with human beings,” he says. “In a recent study by Kinsta, 94.8% of respondents preferred interacting with a human over AI, and 71% had encountered situations where AI struggled with complex issues. The latest AI models, and particularly ChatGPT-5, have also been widely criticised for being ‘robotic’ and lacking the depth and empathy simulated by previous models. Across the business world, there have also been several high-profile reports of companies replacing people with AI, only to find that they had to rehire when the systems failed to deliver to expectations.”

An adviser delivering nuanced, tailored, considered advice with empathy, experience and insight remains something that cannot be simulated or replaced by AI. Because, as Lynch makes clear, AI can’t be left to run unchecked. The rise of AI could, in fact, unintentionally widen financial literacy gaps through the spreading of misinformation. Lynch says, “The term used online for misinformation disseminated by generative AI is ‘slop’. This creates many problems, including the risk of widening financial literacy gaps as consumers could be bombarded with incorrect information.”

There are, however, two key ways that advisers could help to combat this. Firstly, Lynch recommends contributing to the accurate information available online and raising their profile as a trusted source of knowledge – whether that be through a social media presence or by contributing to industry articles and podcasts. “The good old-fashioned email newsletter is even enjoying a renaissance at the moment, with the rise of platforms like Substack – the home for many of the web’s most popular newsletters,” Lynch says. “Whilst blog and web content can be easily scraped, consumed and regurgitated by large language models, subscriber-only content remains private. Setting up a Substack newsletter is simple, and you can offer free content as well as subscriber-only and paid content. This approach also has the added advantage of letting you know who your readers are.” Secondly, Lynch reminds brokers that once a customer has found them, they actually have an advantage over AI. “Advisers don’t have to do everything online, meaning they can tailor their communication. Whether that be offering short consultation phone calls or even drop-in times at their office, advisers are empowered to connect with customers in

the way that best suits them. We all recognise this is especially important when it comes to tailoring your style to vulnerable customers and helping them differentiate financial facts from fiction.”

Partnering with AI-savvy buyers

A partnership between adviser and AI

Published September 22, 2025

“AI, in one of its many forms, may be able to help, freeing up time for you to do the things that only you can do – especially providing that all-important human connection with your clients”

Chris Lynch, Source Insurance

75%

are planning to use AI over the next three years

10%

of insurance firms are currently using AI – the highest rate of all sectors

95%

Source: Bank of England

Source: Cloudvirga

of younger homebuyers say that if AI played a major role in the application process, they would consider switching lenders

Younger homebuyers PREFER balance of AI and human support

60%

of these homebuyers rely on loan officers to navigate mortgage applications

58%

of first-time homebuyers want a more enhanced digital experience

63%

say they want a fully digital process

77%

Navigating ‘AI slop’

It’s about attaining a sort of marriage between human expertise and AI – a blend of personal touch and innovative tech. Lynch believes that advisers should be looking at deploying AI tools internally to improve the more monotonous tasks, such as workflow automation or personalised recommendations. “Even the most ardent AI sceptics accept that AI is already ‘baked into’ many pieces of software and will continue to be part of the business landscape for the foreseeable future. If you’re a Microsoft Office user, you’ve likely come into contact with Copilot. Copilot can help you write emails, summarise a document and research new topics. If you want to level up your marketing, you’ll find lots of options in new products from Adobe and Canva that help non-designers improve their output using generative AI. “Google’s Notebook LM product is an exceptional tool for working with your own documentation and files to create your own, more personalised AI experience. There are new products released almost every day that can help you manage your to-do list, calendar and emails. These tools can be great time savers and ways to build the skills to succeed.”

However, before you start jumping onto the AI train, Lynch recommends that brokers ensure that they:

include human checking in the process – the AI term for this is ‘human in the loop.’ Remember that you are responsible for the output, no matter how it was generated. remember their data – take time to consider what data you are sharing with third parties and whether this is in line with your own policies and the wider data protection framework of GDPR or DPA.

“Most importantly, remember that generative AI cannot ‘think’

or solve problems,” adds Lynch. “We’ve all been led to believe in AI’s ‘thinking’ and ‘reasoning’, but thanks to continued research and one very famous research paper from Apple, we now know better. In this context, it is best to view AI as an assistive technology. There are plenty of things that we all do every day that we’ve done many times before and are time-consuming but not complex. This is where AI, in one of its many forms, may be able to help, freeing up time for you to do the things that only you can do – especially providing that all-important human connection with your clients.”

Embracing AI as an assistive technology, not a replacement for human intuition, knowledge, thinking and experience, has been the cornerstone of Source Insurance’s approach to artificial intelligence. Lynch and his team of experts have been busy behind the scenes looking at how AI can augment, but never replace, the human touch in the mortgage process. “Source already takes a technology-led approach to providing intelligent insurance solutions,” Lynch reveals, “whether it be harnessing data enrichment in our UX platforms or enhancing our policy administration systems through automated processes. What’s coming next for us is utilising AI in our innovative work allocation solution so that we can further increase efficiency when responding to queries and processing admin. “The key for us, as will be the case for advisers, is to use AI as an assistive technology that helps our people so their focus is always on delivering that all-important human touch.”

Source Insurance’s technology-led approach

Sitemap

Newsletter

About us

Authors

Conditions of Use

Terms and Conditions

RSS

E-mag

Jobs

Events

White papers

Bridging

Buy-to-let

Commercial

Equity release

Residential

Second charge

Interest rates

Growth

Guides

Market trends

Opinion

Technology

Resources

News

MORTGAGE TYPES

Best in Mortgage

Mortgage Industry

RSS

Sitemap

Terms & Conditions

Conditions of Use

Authors

About us

Newsletter

Resources

MORTGAGE TYPES

BEST IN MORTGAGE

MORTGAGE INDUSTRY

News

RSS

Sitemap

Terms & Conditions

Conditions of Use

Authors

About us

Newsletter

News

MORTGAGE INDUSTRY

BEST IN MORTGAGE

MORTGAGE TYPES

Resources