Messagepoint’s Patrick Kehoe: Leveraging Generative AI for Borrower Communications That Resonate

Patrick Kehoe drives product strategy in collaboration with the product development team at Messagepoint, a provider of customer communications management software. Kehoe brings to the company more than 25 years of experience delivering business solutions for document processing, customer communications and content management.

Since the release of ChatGPT in late 2022, advancements in generative AI technology have taken center stage in conversations across nearly every industry. Among the most widely discussed aspects of generative AI–a category of artificial intelligence that can produce various types of content including text-producing tools such as ChatGPT and Bard–is its potential to cater to the demand for personalized communications and  streamline the workflows required to create them.

A recent report from research and consulting firm Gartner shows that enterprise interest in generative AI has continued to climb because of publicity surrounding ChatGPT, with 38% of respondents citing improved customer experience as a primary motivation behind their decision to implement these technologies. And while many organizations have understandable concerns about data security and the potential for biased or inaccurate responses, 68% of respondents reported a belief that the benefits of generative AI far outweigh the risks. 

For the mortgage industry, one of the most attractive features of generative AI is its potential to revolutionize communication between providers and borrowers. The official end of national emergency declarations surrounding the COVID-19 pandemic have left many borrowers unable to pay their mortgages, putting agencies in the tricky position of balancing their loss mitigation efforts with the need to maintain compliance with the CARES Act.

According to a recent JD Power mortgage service satisfaction study, the percentage of financially unhealthy mortgage customers rose from 48% to 54% during the past year—which, combined with climbing interest rates and declining consumer trust in financial institutions–requires mortgage providers to be strategic when it comes to their loss mitigation programs and the ways they communicate them to borrowers. Luckily, balancing this climate with the need for streamlined workflows is where generative AI excels, providing an efficient way to build relationships between borrowers and providers.

Finding the Right Words With Generative AI

As the saying goes, “Words matter.” The words we use can make our borrower communications easy or impossible to understand, helpful or threatening, actionable or dismissed. Too often, legal team editing of borrower communications steers them in the wrong direction. Overly complex or blunt language can have a significant impact on a borrower’s response. Striking the right sentiment in your communications is crucial, not only to establish trust, but also to prompt a desired response from consumers. Studies show that communications likely to elicit a negative response, such as those about foreclosure and loss mitigation, release stress hormones in the reader, often causing borrowers to shut down rather than take the necessary steps toward a resolution. On the other hand, communications that attempt to soften the impact of negative messages with overly positive language can give consumers the wrong impression about the true urgency of the situation, leaving them confused about how to respond.

One application of generative AI is the creation or rewriting of text-based content that conveys a particular sentiment to address this issue. Generative AI models analyze the rules of grammar, syntax and composition based on the vast swaths of data fed to them to help predict an optimal sequence of words to achieve a certain objective. The result of extensive training is a finely tuned language prediction capability that enables generative AI to formulate convincingly lifelike, situationally appropriate responses to user-generated queries. Additionally, the analytical function of generative AI makes it possible to assess and adjust the sentiment or other parameters of content, such as the reading level of borrower communications or rewriting content according to plain language principles to reduce the impact of stress caused by overly complicated language.

This can be particularly helpful with low English proficiency (LEP) borrowers and the 31.5% of U.S. adults whose reading comprehension falls between a third- and fourth-grade level. Generative AI enables teams to accomplish this while dramatically reducing the amount of time spent aligning communications to the desired sentiment and readability level, allowing human employees to focus on higher order concerns.

Establishing Trust Through Consistent Multichannel and Multi-Language Messaging

Customers increasingly want to access their communications using a variety of platforms, including non-traditional channels such as mobile apps or WhatsApp. They tend to spend more on brands that earn their trust through consistent, recognizable messaging across all points of access. However, this doesn’t mean messages should be communicated in exactly the same way from channel to channel since communications that aren’t optimized for a particular platform can be an added source of inconvenience or confusion. For example, while a PDF may be easily read on a computer screen, a borrower who accesses this communication via their phone will be required to pinch, scroll and zoom to see all the information being communicated.

Adapting communications in this way, particularly to be read on the smaller screen of a smartphone, often begins with a significant reduction in length, a task that can be deceptively time-consuming when you consider the intentional phrasing behind messages that deliver a consistent voice, concept and tone in fewer words. That’s where generative AI comes in: as an efficient tool for reformulating communications in a way that preserves meaning, reduces word count and adjusts formatting so messages are optimized for individual channels. 

Similarly, when borrowers are best served by receiving communications in their first language, AI can help to save the tedious process of human-led translation. Today’s translation capabilities are remarkably accurate and can lift the heavy burden off translation teams.

In this current economic climate, it’s important for the mortgage industry to consider ways to ensure communications resonate with customers, even communications that are likely to be perceived as negative but that still require a specific, prompt response from the borrower. Of equal importance is the need for agencies to streamline the process of tailoring communications to individual channels, as organizations in every industry are tasked with doing more with fewer employees. When fully integrated with an existing customer communications management platform, generative AI enables mortgage providers to accomplish each of these goals simultaneously by providing an efficient means of optimizing sentiment, readability and channel compatibility—all while focusing on the human element of every transaction.

(Views expressed in this article do not necessarily reflect policies of the Mortgage Bankers Association, nor do they connote an MBA endorsement of a specific company, product or service. MBA NewsLink welcomes your submissions. Inquiries can be sent to Editor Michael Tucker or Editorial Manager Anneliese Mahoney.)