How A.I is Changing the Mortgage Lending Industry

Artificial intelligence (AI) is increasingly becoming a part of our everyday lives, assisting and influencing our work, decision making and experiences. Generative AI applications like ‘ChatGPT’ are dominating watercooler conversations around the globe, and machine learning is being utilised by industries such as transportation, healthcare, marketing, and yes, you guessed it, the mortgage industry.

In real estate, AI is streamlining how people purchase, sell, and rent property, by monitoring and translating relevant data, making the process faster and easier for users to understand. Programs such as “Zestimate” owned by US company Zillow, uses machine learning to estimate property values. Zillow states that its median error rate for on-market homes nationwide is 3.2% and 7.52% for off-market homes.

However, these AI-generated estimates aren’t foolproof, as the technology is unable to account for the unpredictability of the market. AI learns from historic data, patterns and trends, and while housing markets do follow trends, they’re also adversely impacted by unpredictable events (like pandemics). These AI-powered tools also don’t take homes’ specialty features or cosmetic damage into account.

Within the finance industry, lenders are actively investigating recent advancements in conversational AI technology, and some are already using virtual assistants, chatbots and speech recognition technology to engage with their customers. AI has also been embedded into mobile banking apps for some time, tracking user behaviour to provide relevant financial insights, predicting upcoming bills and forecasting cash flow. The big four banks have also integrated AI and data programs into their operations, enhancing their customer service offerings.

Perhaps the most contentious issue relating to AI and mortgage lending is the use of algorithmic underwriting, where AI assesses a loan application and decides on an individual’s borrowing capacity. While in theory the AI’s objective, data-based decision-making eliminates the possibility for discriminatory bias, in practice this isn’t necessarily the case. AI is trained on previous credit decision data, meaning that conscious and subconscious biases from lenders in the past might impact the AI-generated result.

For example, an AI application might identify a large number of rejected loan applications belonging to residents of a particular postcode. Whilst the postcode likely has nothing to do with why the applications were rejected, AI technology may recognise this pattern and take the postcode into account when making its decision.

AI has the potential to positively impact the finance and mortgage industry in a big way, streamlining processes, saving customers time, improving risk management and increasing business efficiency. However, the implementation of AI technology across the sector is still evolving and human skills and qualities such as creativity, empathy, intuition, and face to face customer service, are still imperative to the home loan process.

• SHARE