AI in the Escrow Industry: A Rising Tide

BY

Bobbi Pronin

.

December 19, 2023

A group of colleagues working together

Artificial Intelligence (AI) has been around for a while, but the pros and cons of its use are increasingly on the lips and minds of students, teachers, business executives, and people in every walk of life – the more so since a start-up called OpenAI released its user-friendly chatbot, dubbed ChatGPT, just over a year ago.

ChatGPT saves the user time and effort by creating finished material from a prompt. So, it’s not surprising that poll numbers gathered during a recent webinar hosted by the American Land Title Association (ALTA) show that nearly half of title specialists viewing the webinar are either already using AI in their business (25 percent) or want to know more about it.

Acknowledging the almost eerie aspect of using large language models like ChatGPT for the first time, Robert Zwink, Chief Technical Officer for Razi Title Exchange, compared it to the first use of grammar-correcting software, robotic floor sweepers, or even self-driving cars – all technical advances that have now been around long enough to have lost the wonder factor.

Despite their relative newness, however, large language models, like ChatGPT, he said, and technologies like Robotic Processing Automation (RPA) are built to understand and generate logical responses to user input.

Razi, for example, is using its proprietary large language model for such tasks as:

  • Creating a viable mission statement and/or other company communications from suggestions loaded into it
  • Creating friendly, ‘personalized’ customer letters out of information contained in a company’s mission statement or other formal documentation
  • Creating an improved version of a company logo from sample attempts
  • Extracting meaningful, easily understandable language from a land deed loaded into it

Tasks like these are typically done with 98 percent accuracy, Zwink said. But the other two percent might be misspelling a name on a title insurance policy, for example, which is why careful human reviewing and editing of all AI-created material is vital.

That’s why first steps for title companies considering the use of such models, Zwink said, include developing a responsible AI policy; one that emphasizes caution, review and edit standards, procedures for spotting AI errors, and vigorous efforts centered on fraud prevention, as cyber-criminals can potentially use the same technology to steal data and money.

But implementing AI for repetitive tasks, such as a need for typing and re-typing data in the title industry, is a proven way to increase efficiency, said Aaron Davis, CEO of Florida Agency Network, noting that he has been able to decrease his staff by 25 percent since going all-in on AI technologies.

“Humans will always be the drivers,” said Davis. “They are critical to our success. But the use of RPA overlaid with AI results in major efficiencies for the humans who use it.”

Two webinars on these topics may be viewed by going to ALTA.org and selecting ALTA Insights Archive.

---

This material is meant for general illustration and/or informational purposes only. Although the information has been gathered from sources believed to be reliable, no representation is made as to its accuracy. This material is not intended to be construed as legal, tax or investment advice. You are encouraged to consult your legal, tax or investment professional for specific advice.

About Bobbi Pronin

Bobbi Pronin is an award-winning writer based in Orange County, Calif. A former news editor with more than 30 years of experience in journalism and corporate communications, she has specialized in real estate topics for over a decade.

Bobbi is not an employee of Anywhere Integrated Services or affiliated with its title companies.