There’s more to artificial intelligence than the algorithms that make Siri, Facebook, and Amazon work. It holds the potential to instantly scan and compare mountains of data. That’s why the Texas Department of Insurance (TDI) is testing its use to review insurance policies.
“Imagine being able to instantly compare every policy we’ve reviewed over several years to see what language we’ve approved or denied,” said Insurance Commissioner Kent Sullivan. “This could be a game-changer in terms of helping us protect consumers and ensure regulatory consistency.”
The artificial intelligence pilot is one of the latest projects in Sullivan’s efforts to improve the agency by building on best practices, modernizing technology, and focusing on clear communication.
To get started, TDI is entering five years of data on auto policies into the new system.
“We’re doing a trial now that lets us compare policies we’ve seen in the past,” said Nancy Clark, TDI deputy commissioner of Administrative Operations. “The first phase of the pilot – where we test the concept – is on an aggressive timeline. We’re already planning the next phase, where we start using the system to review new policy filings.”
Today: Mountain of work, antiquated tools
When companies want to sell a new insurance policy in Texas, they must file the policy language with TDI. Companies filed almost 30,000 property and casualty policies and related forms with the agency last year and almost 15,000 health policies.
One personal auto policy can be more than 40 pages long. While standard policies once dominated the market, insurance companies now develop products targeted to individual consumer needs. This creates countless policy variations. Ride-sharing and other industry changes also add complexity to each review.
TDI employees have had reference charts and other tools to help with reviews, but they also had to rely heavily on experience, carefully reading every policy and manually comparing provisions in similar policies as needed, said Talia Bright, who works in the office that reviews property and casualty policies.
“We look to see if the language conflicts with a Texas law,” Bright said. “When language is unfamiliar, I check to see if it’s like something we’ve approved before. That’s challenging because I have to dig through multiple databases.”
The agency has lacked modern technology to facilitate the most efficient and highest quality reviews. That’s changing.
Future: Faster, more comprehensive reviews
Because the artificial intelligence review is nearly instantaneous, the project promises dramatic improvement.
“You put in a policy, the system reads it, and it immediately sees if there’s anything similar in the database,” Clark said. “It can also look for required consumer protections.”
Clark emphasized that the decision-making process remains with the regulator. “But this tool gives us faster and more comprehensive information to make good decisions,” she said.
Sullivan has been a champion of upgrading and modernizing technology at TDI. Other projects include adding Tableau software licenses to improve data analysis and putting document sharing technology and processes in place.
“It’s been said that government runs on an Excel spreadsheet,” Sullivan said. “It’s not far from the truth, but it’s time to change that. If we’re going to regulate a high-tech insurance industry, we must have cutting-edge tools to keep up.”