Technology trends may ebb and flow, but separating hype from achievable return on investment (ROI) is a constant, regardless of application, functional area and industry. Nowhere is this more apparent than in the mortgage lending space, where artificial intelligence (AI) powered by natural language processing (NLP) is solving critical business challenges and delivering ROI to early adopters. (In part 2 of this series, we’ll take a look at the application of blockchain to mortgage lending.)
Cognitive assistants already permeate our lives – scheduling appointments, issuing reminders, playing songs, controlling house temperatures, activating security alarms. What if these advancements were applied to mortgage lending? Let’s consider two possibilities:
- A home buyer looking to prequalify: Typically, buyers either walk into a bank, work with a personal advisor to fill out a prequalification form and get a prequalification letter or, more likely today, browse a site like Zillow or Redfin and use the prequalification calculators advertised. Each has its strengths and weaknesses. While faster and more convenient, the Zillow/Redfin route is more of a ballpark estimate; the personal banker route is more accurate but more time-consuming and requires more personal interaction.
- A home buyer needs information on various product types: Similarly, the conventional approaches are to speak with a loan officer either in the branch or over the phone, or rely on Google searches. Here again, the pros and cons have to do with speed and convenience over personalization and accuracy.
A Smarter Way to Facilitate Mortgage Lending
Now imagine if a lending institution made its services available through a cognitive assistant like Alexa or Google Home. These scenarios would play out quite differently:
For the lender, the steps to review financial and personal information would include:
- Bank balances
- Mortgage statements
- Credit report
- Verification of assets
- Verification of income
- Verification of employment
- VA eligibility
- Millennial/senior eligibility
- Location/zonal discounts
Investing in prequalification services alone could help not only drive growth in prospects/leads but also reduce the cost of loan origination and education. Loan officers may hesitate to spend time on complex loans, given the lower ROI for the time spent, but a cognitive assistant could walk prospects and customers through countless scenarios at the home buyer’s pace.
For example, our Intelligent Mortgage Advisor platform offers prebuilt flows (see below) that support prospective customers with product information, information on the loan lifecycle, document requirements and other services.
Similar use cases can be developed for loan servicing – imagine the multitude of questions contact centers manage that could easily be handled by a cognitive assistant. Has my payment posted? How much is my escrow? Why has my escrow amount changed this month? What is my payoff quote?
Or what if we added predictive analytics to turn the cognitive assistant into more of an advisor. For example, if you received a bonus, the assistant could provide advice on how to invest it: “If you invest it into your mortgage, it would save you $X in interest over the course of the loan. Would you like me to make that transfer?” (Read how the insurance industry is applying chatbots.)
Overcoming Security Concerns
The key concern we hear from clients and prospects regarding these types of chatbot applications pivots around security and, specifically, authentication. Consider these options for authentication when using a voice-based AI assistant like Alexa:
- Have Alexa send an authorization code either via SMS or to the Alexa app on your phone.
- Use a numeric pin that the user has to say out loud.
The first option of using an SMS authorization code option disrupts the entire hands-free conversation experience, while the second option of a numeric pin could easily be overheard. One increasingly popular alternative is biometric authentication, a realm of security that’s rapidly growing in popularity (and reliability).
Voice biometrics leverages a sophisticated statistical algorithm to digitize the distinctive characteristics of a person’s speech to produce a unique stored model or voice print. While impersonators can copy the pitch and intonation, other physiological features such as vocal cord location and larynx structure are unique, which make voiceprints as distinctive as fingerprints. Earlier platforms relied on active voice authentication (customers reading one or a series of set phrases), but during the last few years, passive voice authentication has emerged, where set phrases are no longer needed.
Realizing ROI from AI and NLP
One of our clients, a UK-based major bank, sought to reinforce how it authenticated transactions in its self-service mortgage modules to reduce the back-office time and effort spent on verification. We implemented an on-premise voice solution using Nuance Voice, integrated seamlessly with the bank’s existing Avaya and Cisco IVR infrastructure, which passively analyzes users’ tone, pitch and rhythm of voice. This solution enabled superior customer experience by eliminating the need for a password, PIN or passphrase.
Just as important, the solution drastically decreased fraudulent transactions and reduced verification time by about 80% (~1 minute to ~10 seconds), which has not only enhanced customer experience but also delivered clear ROI.
For financial institutions, it’s about penetrating new markets and potentially creating a new channel of communication to extend hyper-personal customer service. Any bank with the vision and courage to enter these territories could rapidly build a strong brand position as a game-changer. With demanding consumers, ever-thinning margins and the need for differentiation, the combination of AI and NLP may well be the nostrum that gains favor.
Stay tuned for Part 2, where we’ll explore how blockchain is enhancing lending globally.
Subscribe to our newsletter and get expert insights straight to your inbox.×
SUBSCRIBE TO OUR NEWSLETTER✖
Thanks for your interest in Digitally Cognizant.
To complete the subscription process, please click the link in the email we just sent you.