How the Texas transportation department has embraced AI

It’s been a busy summer in Texas, as the state hosted 16 World Cup matches across Dallas and Houston, as well as numerous other major events that could create major traffic headaches.

But the Texas Department of Transportation, one of many agencies responsible for such events, looked to put itself on a strong footing ahead of the influx of millions of people to the Lone Star State, embracing technology — especially artificial intelligence — to do so.

In the past few years, TxDOT has turned to AI to help it identify issues in its infrastructure, and to collect and analyze traffic crash data. Meanwhile, employees have leveraged the technology to be more efficient in their jobs and to help alleviate some of their more menial tasks. It’s all a far cry from when technology was not at the forefront of the agency’s operations, but comes just as the state experiences massive population expansion and needs infrastructure that can keep up.

“It's a great time to be in transportation right now,” Anh Selissen, TxDOT’s chief information officer, said during a Future Nation webcast hosted by Route Fifty and Google Public Sector this week. “There are so many things that we do that build a road; some of that is manual. Think about the power of being able to leverage AI as a tool to be able to expedite the building of a road. We are seeing monumental growth in Texas; population growth and expansion of the roads is critical to the success overall of Texas. I'm excited about where we can take transportation, especially leveraging innovation like we are right now.”

Texas has thousands of video cameras across the state, and Selissen said TxDOT is already using AI to examine the quality of the infrastructure they are pointed at, rather than have to carry out manual evaluations, which can be time consuming. Using AI for preventative maintenance is something other transportation agencies see the value in, and Texas is no different.

“We can get real-time information on the quality of those video cameras, so that at all times our infrastructure is in a good place,” Selissen said.

AI also has a role to play in traffic management, Selissen said. That includes collecting and analyzing information from first responders to identify crashes on the state’s roads and remove that disruption more quickly so that traffic can flow again. The technology could also be helpful in helping TxDOT’s traffic safety division forecast traffic and better synchronize traffic lights to avoid congestion at intersections.

Using data from third parties can help make that forecasting seamless, Selissen said, and it is part of a process of understanding how AI can be best used to make the travelling public’s lives easier.

“Looking at how we can remove congestion on the roads by getting third-party responder data feeds, that's low risk,” she said. “It's really taking a look at viable use cases that can be applied that actually can minimize congestion and increase safety on the roads with the low risk of leveraging AI, because we're still trying to figure it out.”

One less headline-grabbing but effective way TxDOT is leveraging AI is by helping its employees with some of their more menial everyday tasks. Invoice reconciliation has been a big time saver and productivity gain, as rather than spend time manually processing invoices and payments to vendors, staff can use AI to reconcile purchase orders and charges in far less time.

“We pay a lot of invoices, with contracts and to vendors, because we do a lot of road construction, buying infrastructure, buying hardware, and in the past we've had people who've had to reconcile what the PO has to do from a vendor perspective with what the vendor charges us, and sometimes they take months, depending on the size of the invoice,” Selissen said. “We've been able to leverage AI to reduce that from months and weeks to possibly just minutes and hours, and so we've seen a direct correlation of savings.”

Governance, then, has been a crucial aspect of the agency’s embrace of AI, as leaders have been keen to ensure employees use the technology safely and properly. TxDOT first released its AI Strategic Plan in 2024, and has updated it in the time since to reflect how the technology has evolved along with its potential use cases.

Selissen said the agency has used lessons learned from prior technology adoptions to govern its adoption of AI, and that has stood it in good stead.

“We leverage the existing governance process that we have around large enterprise technology implementation,” she said. “When you're implementing a large technology, you are looking at what's the risk, and how do we mitigate any risks with implementing a technology? What's the data you're consuming inside it? Are there privacy and confidentiality risks to that data? Who is trying to use it from a user perspective? Who's your audience? Then what are you trying to achieve? By leveraging existing governance, we applied AI on top of that.”