Sponsored by Festina Finance

Photocredit: Festina Finance.

An Intelligent Mortgage Advisor Is the Human Side of AI

Festina Finance applies artificial intelligence (AI) to streamline the mortgage application ­process, and the Copenhagen-based firm’s first partnership with a UK based building society is transforming the end-to-end customer journey.

In the UK today, mortgage sales are dominated by ­face-to-face mortgage advice, paper-based documentation and little in the way of innovation.

Over the last three years, mortgage lenders have started to investigate how a range of ­technologies could speed up completion times and improve the way customers get advice. A crop of firms now offer “robo-advice”, but in reality, these are just automated fact-finding forms for customers to fill in rather than genuine AI.

As a result of this, a common misconception has built up within the UK mortgage industry that AI is not yet advanced enough to handle the complexity of the mortgage advice process, in particular, the selection of a suitable product.

That’s all changed with Festina’s recent partnership with Hinckley & Rugby Building Society.

The mortgage lender is using Festina’s AI-based system to transform its mortgage process, improving the experience for customers and the quality of advice provided, as well as dramatically speeding up application-to-offer times.

This is an example, argues Jesper Lauritsen, CTO of Festina Finance, of the human side of AI – especially in financial services – and the role it can play in radically improving customer interactions and experiences.

“Our approach to AI has been to see it as a complementary tool when designing a business process and customer journey. The software collects information about the customer’s financial situation and asks relevant questions on their views and understanding of mortgages and the overall economic outlook. The software then adopts what the human advisor does and continuously optimises the data set based on the knowledge of the human counsellor,” Lauritsen says.

Currently, Hinckley & Rugby uses it as a tool for its in-house branch mortgage advisers, but the next step will be to support the customer through a self-service, fully-advised mortgage process, without the aid of a human adviser.

AI is not a matter of magic

Festina’s ambition is to use AI to transform some of the time‐consuming work into improving the customer experience and to maintain compliance.

However, Lauritsen points out that there is still some work to do in terms of educating the general public, including the wider financial services industry, about the benefits that AI can provide.

“The challenge is that AI is like magic to many people. They don’t see what lies behind the technology. So, from the beginning, we wanted to create a tool where the human advisors can decode and explain the recommendations made by the AI to the customer. This makes the solution relevant across the organisation and on all levels,” Lauritsen says.

Festina Mortgage Advisor is part of a bigger financial advisory system called Festina Advisor, which is used by banks to advise on loans, mortgages, investments and retirement planning and is deployed as a holistic financial adviser. Festina Advisor is involved in more than 1000 cases each day and learns from each session, especially when there is a discrepancy between the recommendation given by the advisor and the decision made by the customer.

The software registers the actions of the human advisor and, as the dataset grows, the algorithm becomes more accurate. The extensive dataset allows clients to use the intelligent mortgage advisor to train junior counsellors and ensure they stay compliant.
“How to deal with a vital decision that will have a significant impact on how your economic situation will progress or how your family situation will be in the future is a very private matter. Some prefer to talk to a financial expert they trust. Others prefer to handle the process themselves and investigate the details on their own. And some will turn against the advice given. Regardless of the customer’s needs, we want to understand how we can provide the most valid outlook through the use of AI,” Lauritsen says.

A tailored process

Hinckley & Rugby’s implementation of Festina Mortgage Advisor uses AI to support its in-house mortgage advisers. The software gives both the customer and the building society advisor a specialist AI‐helper with integrated process support.

“It has improved our efficiency and the flow and appropriateness of the scripts used within our process when we go through our advisory process with customers. The AI tool is a second validation of the advice that is very valuable, as we can point out specific steps we need to emphasise or optimise in the process. We monitor the advice given by human mortgage specialists and compare results to the systems AI outputs,” explains Dean Waddingham, Chief Customer Officer in Hinckley & Rugby.

According to Waddingham, Festina Mortgage Advisor helps the human advisors by taking them through a process that is much more tailored to the individual in front of them.

“With a very high percentage of accuracy, the Festina Mortgage Advisor comes up with a recommendation for a product that is consistent with what we would expect. This means that we get enhanced confidence in the quality,” Waddingham concludes.

Tags: ,
Training the Festina Mortgage Advisor with adaptive questioning

Festina Finance is a Danish Fintech company working in the field of holistic financial advice. Our AI-based system Advisor helps 40+ financial companies around the world provide financial advice.

The key to optimising Festina Mortgage Advisor is to obtain the customer’s personal preferences for mortgages via adaptive questioning. The customer will answer questions as part of the mortgage advice process, and each question is tailored to the answers already provided by the customer, thereby ensuring that each individual customer is only asked relevant questions. The AI engine applies its “intelligence” to each answer to build a model of the customer’s knowledge and preferences. The more answers provided by the customer, the more the precision of the model improves.