Home >> Daily Dose >> Differentiating AI Applications in the Mortgage Space
Print This Post Print This Post

Differentiating AI Applications in the Mortgage Space

This piece originally appeared in the November 2023 edition of MortgagePoint magazine, online now.

Souren Sarkar, CMB, is President and Co-Founder of Nexval, a Miami, Florida-based technology company specializing in mortgage automation processes and IT infrastructure upgrades. Sarkar has more than 25 years of experience as a technology leader in the mortgage and banking arena and is an expert at improving the performance and scalability of service-driven businesses using workflow automation.

MortgagePoint had an opportunity to chat with Souren about the emergence of artificial intelligence (AI) in the mortgage space and how it is easing processes for both customer and lender alike.

Q: What is generative AI, and what makes it different than other types of AI?
Generative AI stands out from other types of AI due to its unique capabilities and focus. Generative AI, or generative models, uses a wide range of AI techniques, such as machine learning algorithms, deep learning, generative adversarial networks, and large language models to mimic the human creative process.

Unlike other types of AI that use and recognize existing data to generate output, generative AI can create new, original data, like images, texts, and music. It does not just ingest human inputs to return a “Yes” or “No” answer or, at best, to churn out an insight or recommendation. It takes human input and then actively creates an artifact that resembles the input provided, yet it is different and wholly original. It can also assimilate data from existing content (images, video, audio, or text) to create digital assets that feel completely unique.

From creating a fine art masterpiece and composing original music to writing text, generative AI can not only generate new outputs, but can also improve them.

Q: How could generative AI be applied in the mortgage industry to help reduce human effort and errors?
Generative AI can provide plenty of use in the mortgage industry in cases that save time and money, and reduce errors as well. For example, it can craft servicing messages that are nuanced, polite, yet effective and comprehensive, enabling servicers to connect with customers on a personal level without human assistance.

It can help create new, accurate borrower profiles rather than just finding and matching patterns that it has met in the past, like traditional AI does. It can help analyze a borrower’s financial profile to generate a more accurate assess the borrower’s creditworthiness. It can even recommend the amount of loan and interest rates based on the borrower’s financial history.

Q: How can generative AI be used to improve customer experience and engagement?
Generative AI is a game-changer when it comes to enhancing the customer experience and engaging borrowers. For example, many borrowers face lengthy delays even support interactive virtual assistants to help borrowers throughout the mortgage application process. Generative AI-powered chatbots can also use natural language conversations to answer common queries and provide financial advice.

Q: What are the challenges involved in training and operating generative AI models?
The old saying, “garbage in, garbage out” applies here. Generative AI models produce results based on the data you use to train them. If you input faulty, outdated, or incomplete data, the output will also be incorrect. Additionally, generative AI can be challenging to implement legacy systems and complex IT infrastructures, which may not be compatible with generative AI solutions. On the other hand, businesses that already have adopted automation and have an updated infrastructure in place have the flexibility and agility needed to train, implement, and operate generative AI models.

Q: How can mortgage companies prepare for the costs associated with generative AI?
Adopting generative AI requires careful planning and consideration. Mortgage companies should start by conducting a comprehensive cost-benefit analysis to measure the potential return-on-investment (ROI) and identify areas where implementing generative AI will have a positive impact. The next step should be to develop a budget that covers the implementation costs. Because these are not simple tasks, mortgage associated with infrastructure, software, and expertise.

Q: What are the potential risks associated with generative AI adoption, such as lending bias and security? How can these risks be mitigated?
Businesses must have a clear and actionable framework for using generative AI and a plan to be prepared for AI-related risks and mitigate them. Data security, wrong analyses, and lending bias are some of the common concerns that surround generative AI adoption in the mortgage industry.

Let us take the risk of lending bias as an example. Again, generative AI models are as good as the data they are trained on. If you feed generative AI models inaccurate and biased data, they will amplify that bias and generate a discriminatory result.

As far as security—which is paramount in our industry—robust data protection measures like encryption, access controls, and secure storage facilities must be in place to safeguard sensitive information. Regular security audits of generative AI models can also help identify vulnerabilities and stay current with the best data security practices. Creating policies and guidelines for the ethical usage of AI models can also mitigate security risks.

Employees using AI systems must be educated about these risks, so they can use the system wisely and protect sensitive data.

Q: How can generative AI and augmented reality be combined to provide predictive insights into things like real estate construction, valuation, and property preservation? What impact will this have on the housing industry?
Generative AI and augmented reality (AR) are emerging technologies that can be transformative for the housing industry. When combined, they can fuel a kind of interactive design approach in which data constantly inspires new models and virtual projections that are visually presented.

These immersive experiences have enormous potential in the future to provide predictive insights.

For example, generative AI and AR can be combined to generate realistic 3D models of buildings and enable users to visualize and explore virtual construction sites for a wide range of purposes.

They can be used to see how a building’s position and design would affect exposure to the sun and local weather patterns, or to appraise properties by analyzing various factors such as location, features, market trends, and historical data to generate predictive valuation models. They can also create immersive virtual property tours, where buyers can visualize potential modifications without being on site. They can help find and correct existing issues in a building’s structure, which could also help in property preservation needs, or even predict potential risks and hazards associated with real estate projects by analyzing construction and safety data.

Q: How can mortgage companies build a culture of innovation that supports the adoption of generative AI and other emerging technologies?
Building a culture of innovation that supports the adoption of disruptive technologies is like unlocking a hidden treasure. It begins with the unwavering commitment of a company’s leadership, whose vision sets the stage for a thrilling journey ahead. Companies must also develop an innovation strategy that aligns with the organization’s goals and vision. It is also wise to build cloud capabilities so that the company is ready at the time new technologies are implemented. Finally, it should also include training programs for employees to build their knowledge and skills in emerging technologies.

Encouraging continuous learning, taking calculated risks, and learning from failures are all part of building a culture of innovation. Such an effort is rarely accomplished alone, however. When it comes to generative AI and other new technologies, there is so much to learn and so many use cases yet to be discovered. For many lenders and servicers, tapping the expertise and resources of a partner that already has a culture of innovation can shine a bright light on the path ahead and guide them on their journey.

About Author: David Wharton

David Wharton, Editor-in-Chief at the Five Star Institute, is a graduate of the University of Texas at Arlington, where he received his B.A. in English and minored in Journalism. Wharton has nearly 20 years' experience in journalism and previously worked at Thomson Reuters, a multinational mass media and information firm, as Associate Content Editor, focusing on producing media content related to tax and accounting principles and government rules and regulations for accounting professionals. Wharton has an extensive and diversified portfolio of freelance material, with published contributions in both online and print media publications. He can be reached at [email protected].

Check Also

Survey: Homeownership Remains Elusive for Baby Boomer Renters

A recent look into housing affordability by NeighborWorks America has found that three in five long-term baby boomer renters feel homeownership remains unattainable.