Nearly 87.1% of respondents to Fleet Advantage’s Use of AI in Fleets survey report using GenAI large language models (LLM) for back-office tasks, driver feedback, and accessing and extracting insights from internal fleet documentation such as maintenance manuals, SOPs, and compliance guides. This category did not appear in last year’s 2025 edition, making its near-immediate dominance one of the most significant single-year shifts the study has recorded.
Predictive analytics (38.7%) and machine learning (35.5%) trail by a wide margin. Computer vision and robotic process automation both registered 0%, despite clear applicability to damage assessment and automated invoicing.
"The data tells a story we see playing out across the industry every day,” says Mac Hudson, senior off-lease manager at Fleet Advantage. “Private fleets are embracing AI faster than anyone anticipated, particularly GenAI, but enthusiasm alone does not create results. The organizations that will lead this next phase are the ones investing now in data quality, telematics integration, and structured measurement frameworks. The biggest opportunity lies in applying AI to core financial and operational disciplines like TCO modeling and safety management, where adoption still lags significantly. Without this foundation, AI remains a back-office convenience rather than a true operational advantage. The window to build that advantage is open right now, and the gap between early movers and the rest of the industry will widen quickly."
Key takeaways:
· AI adoption in driver safety is also gaining traction, with 61.3% of organizations reporting the use of AI tools to monitor and manage driver behavior and coaching programs. However, 6.5% of respondents indicated they do not have any formal driver safety monitoring program in place, highlighting a small but notable gap in foundational safety practices. While relatively low, this figure underscores that even as AI adoption grows, a subset of organizations with transportation fleets still lack baseline safety frameworks.
· Data integration issues jumped from 38.1% to 71%, and inaccurate data concerns rose from 23.8% to 64.5%. Lack of expertise climbed from 19% to 45.2%. The pattern is consistent: as AI deployment scales, foundational data infrastructure problems are becoming impossible to ignore for organizations – something many in the industry are struggling with today.
· The share of respondents not utilizing agentic AI at all nearly doubled, from 19% in 2025 to 38.7% in 2026, while active usage declined slightly from 19% to 16.1%. Interest in agentic AI for procurement fell from 57.1% to 9.7%, and for asset lifecycle management from 57.1% to 6.5%. Fleet modernization planning is the one area holding steady, declining modestly from 57.1% to 51.6%. These comparative results indicate the 2025 survey captured aspirational intent; whereas the 2026 data reflect actual deployment behavior, and the gap is significant.
· Only 9.7% of respondents have a formal AI ROI tracking framework. The majority (51.6%) track some metrics without structure, and 19.4% assess ROI anecdotally. Investment plans reflect caution: 41.9% plan to modestly increase AI spending, 35.5% are unsure, and 22.6% plan to hold current levels. Zero respondents plan significant increases or reductions.
· The survey also revealed a critical disconnect in Total Cost of Ownership (TCO) modeling for Class 8 assets. Nearly 32% of respondents still perform TCO modeling manually, while 29% do not perform TCO modeling at all. Across remaining responses, adoption of AI-driven TCO modeling averages just 12.1%, signaling that the vast majority of organizations with transportation fleets are not yet leveraging advanced analytics in one of the most impactful areas of asset management. This represents a major opportunity for organizations to unlock cost savings and improve lifecycle decision-making through more sophisticated, AI-enabled modeling approaches.
· A majority of respondents (51.6%) collect telematics and ELD data but have not integrated it with AI tools, with only 9.7% feeding it into AI models for real-time insights. And 64.5% report no use or evaluation of AI for lease-end processes including damage scoring, remarketing, and excess mileage assessment, the largest single area of non-usage in the entire survey.
