Rising trends: AI and automation in the mortgages sector
by Jerry Mulle, UK Managing Director, Ohpen and Abdul Khan, General Manager, Cognizant
One of the key economic trends of the last forty years has been the growth of technology and automation in the workplace. The mortgage market is no exception. As the market has matured and evolved, technology has become the key to injecting efficiency and preserving margins. Nor have those drivers of change diminished. Supply side factors, in the shape of increasing regulatory obligations, fiscal interventions such as stamp duty cuts or government support schemes for buyers, and monetary interventions in the shape of rising or falling interest rates, all impact operational models. They also change risk appetites, and often demand a swift response to protect margins.
Most recently, Consumer Duty and the imminent arrival of Basel 3.1, will mean lenders will once again reflect on what markets they want to be in and what their operating models need to deliver. And then there is the UK Government’s Autumn Budget in October to consider.
Whether you are a high-street brand or a specialist in a market like Buy-to-Let, there is a lot to consider as legislation continues to shape the UK mortgage market. Being fleet of foot in adapting to market and industry change has never been more important. Today’s mortgage market has no place for legacy platforms and legacy thinking. So, what can we learn about the value of automation in this sector?
The primary benefits of AI in the mortgages sector
Dialling it back to basics, Artificial Intelligence, or more appropriately machine-learning, is already enabling lenders to gain more ground in delivering a more efficient mortgage process. While external data feeds and automation have been a feature of the market for some years, AI offers a new opportunity to offer predictive workflows, speed and scale that can really support the people in a business to make better decisions and allow staff to spend time on issues that really require human experience and expertise.
In that respect, elements like automated verification can be hugely enhanced by processes that swiftly verify and analyse applicant data from multiple sources. Client supplied data, enhanced with external data feeds can save firms and their distributors many hours in packaging cases and understanding exactly how appropriate an applicant’s circumstances are for the loan they desire.
At this point, if there are disconnects between what a borrower wants and what is possible to offer, AI software can suggest which products may be suitable for a candidate. This is a real advantage at affordability stage where a binary decline may see business simply ‘up and go elsewhere’.
Customised prompts of this kind benefit brokers, borrowers and lenders alike as time is taken out of the process without losing nuance along the way.
Tackling archaic mortgage application processes
While the business of applying for a mortgage is improving in many areas, it remains time consuming and prone to review on account of the inconsistencies that come with incomplete or inaccurate forms – not to mention any inconsistencies in the application of underwriting rules.
The mortgage lending process in most cases still involves numerous systems, databases, and workflow tools, that in turn can lead to the creation of substantial numbers of documents volumes for processing applications. Without centralised data, underwriters can miss out on crucial information that will likely impact a borrower’s risk profile. Furthermore, underwriters must manually access each platform to retrieve documents, causing significant delays. AI can streamline these processes and allow underwriters to concentrate on the cases that really should demand their attention.
AI models use historic data to analyse trends and identify likely patterns in, for example, the payment profile of an applicant. There are areas in the product modelling process that are already using AI to analyse vast data lakes of borrower behaviour to create new products for specific market segments. Borrower data offers lenders the keys to better origination decision making.
Understanding those patterns can unlock insights on key dynamics in the market from borrower switching to payment difficulties. But it goes further, advanced algorithms and machine learning techniques enable AI to detect and prevent fraudulent activities more effectively than traditional often slower methods, so risk is identified and dealt with more swiftly. For mortgage lenders, machine learning can be an effective way to proactively identify problems before they turn into significant losses.
Final thoughts: AI as a strategic lever
After all, over and above the focussed implementations within the process, AI unlocks the very agility all lenders need to survive in the current age. Efficiencies are always welcome but the strategic implementation of AI is where the technology has the pawer to be truly transformational. When margins are under pressure, and new markets open, being able to respond quickly and understand the impacts on your operations sales and funding strategies is essential.
AI is already present in the mortgage industry and is doing much to streamline current operating models, but its role is growing. Over and above the automation of lower risk tasks, lenders know they can make better, more consistent decisions by judiciously applying its power in the right places. From product modelling to sales and underwriting, AI offers quick wins but also more profound change that can really put your business ahead of the pack.
Image: Freepik