Drive.ai, a California-based computer software company, will launch an autonomous car ride-hailing service in Texas.

The pilot program begins in July, amid increased scrutiny over the safety of driverless cars. But Sameep Tandon, the company’s co-founder and CEO, said the fleet of cars has been tested and they are safe. He said the vehicles’ bright orange color will help them avoid accidents.

“By making our cars orange, we’re really able to set the expectation [that] this is a self-driving car,” Tandon told CNBC’s Phil LeBeau on “Power Lunch” Monday.

“When you see a school bus, you have a slightly different behavior when you drive around it,” he said.

Tandon, who holds a doctorate in computer science from Stanford University and was a research assistant there for deep learning on autonomous driving, said other safety precautions include multiple sensors, cameras, radar and lidar on all vehicles. The automobiles have also been tested by way of simulator systems and in geo-fenced locations, or places that use software with GPS to test the service.

The pilot program will roll out in Frisco, Texas, a city outside of Dallas, with Nissan NV cargo vans, but Tandon said the service will expand over time.

The service will change the way people “micro-transit,” that is, those five- to seven-minute trips, Tandon said.

“It’s a little bit too hot and you don’t want to take a car ride,” he said. “Or, you’re a little bit feeling guilty that you don’t want to take your car there.”

“We think we’re solving a real transportation problem here,” Tandon said.

Still, public concerns over autonomous vehicles have heightened since an Uber vehicle in self-driving mode struck and killed a pedestrian in Arizona in March.

On Monday, Uber released a statement saying the accident was likely caused by “false positives,” or computer software that programs autonomous vehicles to ignore random objects on the road, such as floating plastic bags. The software, the company said, is an issue for all autonomous vehicles, not just Uber.

Mass adoption of driverless cars may not come anytime soon, Tandon said. But his company, he said, is moving one step closer.

“The cool part about this entire AI approach is that as this system gets more data and experience, it’s going to continue and learn and get better,” Tandon said.