Key metrics
70–75%
Of inbound call volume handled with zero human touch each week
~80 hrs
Of talk time freed per week across the team
200,000 calls
Fully automated
400,000 min
Of automated talk time
Products
“I’ve seen many vendors come and go, but few evolve into true strategic partners. Happy Robot team is one of them, from taking an early risk on them to watching their rapid growth, they have consistently delivered innovative, reliable solutions while maintaining exceptional responsiveness, integrity, and a rare ability to listen, adapt, and proactively guide us toward better outcomes, making them not just a technology provider but a trusted extension of our team.”
Frankie Clement
Sr VP of IT, U.S. Xpress
About U.S. Xpress
USXpress is one of the largest trucking companies in the United States, providing freight transportation solutions nationwide. Founded in 1985 and headquartered in Chattanooga, Tennessee, the company operates a vast fleet of trucks and trailers, serving a diverse range of industries with dedicated, over-the-road, and intermodal transportation services.
U.S. Xpress's carrier sales team manages high volumes of inbound calls from carriers looking to move freight. Behind the scenes, a small product team translates business needs into operational systems, and increasingly, into AI.
Opportunity
Every day, carriers make thousands of calls looking to move loads. And every day, the team had to answer, because there was no way to know which calls were worth taking without taking them.
An estimated 60–70% were non-value-added, unqualified carriers, callers who couldn't move the specific load they were asking about. The team had to work through every bad caller to find the good one. There was no shortcut, no filter, no systematic way to separate the two without actual human conversation. "It was certainly accepted as just a part of the job," Hayden explains. "But it was something we wished there was a solution for." A previous vendor could flag the most obviously fraudulent callers, roughly 5%, using a phone number API. The remaining 65%+ couldn't be screened by data alone. You had to engage them verbally. And so the team did, tens of thousands of times, on calls that added nothing.
As the AI boom accelerated at the end of 2024, U.S. Xpress’ product team pulled their call volume data, mapped it across every business unit, and started asking: which of these known problems can AI now solve? Carrier sales were at the top of the list.
Solution
U.S. Xpress deployed CASS: Carrier Authentication & Sales Support: HappyRobot's AI voice agent for inbound carrier sales.
CASS picks up the call, vets the carrier, and determines whether a human needs to be involved at all. The first version, handling just the basics, starting the call and screening the carrier, effectively doubled team efficiency out of the gate. Today, CASS handles 70–75% of inbound call volume each week with zero human touch.
"It was double the efficiency out of the gate, even with that first version of CASS, just doing the basics, starting the call, vetting the carrier, it freed up about 80 hours a week of talk time."
That's 80 hours a week returned to a team that didn't get any bigger.
Building in the platform
U.S. Xpress’ product team built CASS directly inside the HappyRobot platform alongside Dhruv, an FDE for HappyRobot. The low-code, no-code approach meant that people with deep carrier sales knowledge, not engineers, could translate that knowledge into agent logic in natural language. Native TMS integrations meant data flowed in and out from day one.
Early on, the team experimented across many use cases at once. They quickly learned to rein it in. "We realized we had to get serious about a few use cases and really focus," Hayden recalls. "That's when we felt the value really start to accelerate." For the team building on the platform, it felt like gaining powers they'd always wanted but never had, going from nothing to a working MVP faster than anyone expected.
The name
U.S. Xpress named the agent CASS from the start, a deliberate call. With multiple stakeholders, business units, and frontline workers involved, a name meant everyone could stay on the same page. "Rather than saying 'the inbound carrier sales voice agent,' you just say CASS," Hayden explains. "It takes on a life of its own. It kind of helps it stick." CASS is an acronym, but on the floor, they are a teammate. There are even memes in the group chats about them.
Getting buy-in
Leadership came first, business unit by business unit. "Win leadership over," Hayden says. "If you can't, you should consider not doing it." The key framing: the agents belong to the business unit leaders, not the product team. "We make them the vision owners. These agents are their employees." Frontline adoption followed from there, because it came from their own leaders, not from a technology rollout.
Impact
Across all HappyRobot use cases, U.S. Xpress recently crossed 1 million total interactions.
The carrier sales team at U.S. Xpress, freed from unqualified call volume, has expanded its reach into higher-value work. And the ideas haven't slowed down; if anything, they're accelerating faster than the team can keep up. Tracking and tracing agents are next. Beyond logistics, other business units are already asking when agents will be coming to them.
"The ideas are growing faster than we can currently keep up with," Hayden says. "We keep having to reorganize just to increase our velocity."