A new industry survey from EVAI, an AI-first company focused on fleet analytics and decision support, is shining a spotlight on how GPT-style interfaces are making their way into commercial vehicle fleets. Spoiler alert: fleets are curious, cautiously optimistic, but still pretty early in the game when it comes to widespread use.

EVAI’s survey, sent to more than 2,000 fleet professionals in May 2025, explored how generative AI tools like ChatGPT are being used (or not used yet) in real-world fleet operations. And while awareness is growing fast, most fleets haven’t gone all-in just yet.

How Familiar Are Fleets with GPT Interfaces?

• 40.8% of fleet respondents said they’re moderately familiar with GPT-style tools

• 10.7% said they’re very familiar, and 9.7% claimed expert knowledge

• But only 10.7% are using these tools extensively, while 40.8% are just dabbling

• Nearly 39% aren’t using GPT at all but are actively exploring the idea

• And 30.1% said they definitely plan to start within the next 12 months

What’s Getting in the Way?

There are some very real concerns slowing things down:

• Integration complexity (20.4%)

• Lack of awareness and understanding (19.4%)

• Data security worries (19.4%)

• Accuracy and AI “hallucinations” (also 19.4%)

ROI uncertainty (9.7%) and comfort with existing systems (11.6%) also made the list but weren’t the main blockers.

What’s Getting Fleet Pros Excited?

Some of the most promising GPT use cases include:

• Total Cost of Ownership (TCO) analysis (21.4%)

• Trip and route optimization (20.4%)

• Advanced data analysis (17.5%)

• Vehicle health monitoring (15.5%)

In fact, 51.5% said GPT-style tools perform better than traditional dashboards when it comes to making sense of fleet data.

Real-World Results: Efficiency and Trust

For fleets already using GPT:

• 30.1% reported efficiency gains of 8-10%

• 9.7% saw boosts of 15% or more

And among those on the fence:

• 29.1% expect a 10–12% bump in efficiency once they implement

But trust is still a work in progress. Only 10.7% feel very comfortable using GPT tools for ops analysis, and another 39.8% are somewhat comfortable. When it comes to trip and route optimization:

• 10.7% trust GPT completely

• 39.8% trust it—but with some reservations

Human Oversight Isn’t Going Anywhere

Delegating tasks to AI? Fleets are open, but still cautious:

• 30.1% are OK with some AI help, as long as humans are watching

• Another 30.1% support moderate delegation with occasional reviews

• Only 10.7% are cool with high-level delegation

• And just 9.7% would go full-autonomy

Preferred oversight methods include:

• Periodic audits and checks (31.1%)

• Exception-based oversight (29.1%)

• Full-time human monitoring (20.4%)

Where GPT Could Add Major Value

Fleet managers see major upside in a few key areas:

• Automated safety alerts and compliance monitoring (59.2% rated this very/extremely valuable)

• Real-time geofencing and location tracking (21.4% said very important, 29.1% moderately)

But they’re still worried about:

• Data accuracy (31.1%)

• Integration with existing tech (29.1%)

• Privacy and security (20.4%)

The Bottom Line

“The data reveals a clear desire to embrace generative AI in the fleet sector but also underscores that we’re still in the early phases,” said Ian Gardner, Founder and CEO of EVAI. “Fleet managers are asking smart questions about integration, trust, and ROI—and that’s where EVAI is helping bridge the gap, providing real-time, actionable insights while keeping human oversight in focus.”

While GPT tools won’t replace dashboards and driver input anytime soon, this survey shows a future that’s charging toward smarter, more predictive operations, with the right tools and training leading the way.