Each year, the Transportation Research Board Annual Meeting provides a clear signal of where the transportation industry is headed.
At TRB 2026, one thing was unmistakable. Artificial intelligence is no longer viewed as a future capability. It is quickly becoming an expectation.
What felt different this year was the tone of the conversations. There was less discussion about what AI might enable someday and more focus on how agencies are trying to use AI tools today. Many of these conversations were centered on improving day to day workflows, helping staff move faster, prioritize more effectively, and spend less time on manual or repetitive tasks. In many cases, when pilots struggled, it was not due to the AI itself but because the underlying data was inconsistent, incomplete, or difficult to maintain over time.
The Most Common Question: Are We Ready for AI?
Across technical sessions, hallway conversations, and meetings with agencies and researchers, a consistent question emerged. How do we practically use AI in transportation.
The answers rarely focused on algorithms. Instead, discussions repeatedly returned to core fundamentals. Agencies are thinking deeply about how to ensure reliable and consistent data streams, how roadway geometry data is created and maintained as networks evolve, how operational datasets fit into broader data ecosystems, and how signal timing data connects with performance, planning, and operations.
There was broad agreement that AI can unlock real value. There was also growing recognition that without trustworthy, well structured, and continuously maintained data, AI tends to accelerate existing uncertainty rather than reduce it.
The Foundation: Data Quality and Consistency
Many conversations at TRB circled back to a familiar set of challenges:
- Data quality
- Consistency across sources
- Reliable, continuous data streams
- Clear ownership and long-term stewardship
Speed data, geometry, signal timing, and other operational datasets surfaced repeatedly. While access to data has improved significantly over the past decade, the harder challenge now lies in maintaining consistency as systems, vendors, and priorities change. Standards are emerging, but the real work is happening in how those standards are applied, trusted, and sustained over time.
This focus on consistency is not abstract. It directly affects whether analytics can be scaled, compared, and operationalized across large networks.
Standards, Ownership, and Long-Term Maintenance
Another recurring theme at TRB 2026 was data governance. As new datasets and exchanges mature, the industry is increasingly focused on questions that go beyond technology.
- Who owns the data over the long term?
- Who is responsible for keeping it current?
- How agencies balance innovation with day-to-day operational reality?
These discussions reflect an industry that is maturing. Data access alone is no longer enough. Long term stewardship and accountability are now seen as prerequisites for successfully scaling analytics and AI.
Why This Change Matters
What stood out at TRB is how well positioned the industry is compared to even a few years ago. Today, we have extensive historical performance data, improving consistency across operational datasets, and growing alignment between agencies, researchers, and industry.
This shift is particularly visible in signal operations. A recurring theme was the need for consistent performance measurement across entire networks, not just at individual intersections. One advantage of probe-based signal performance measures is the ability to apply the same methodology everywhere. This allows agencies to compare performance across hundreds or thousands of signals using a common framework.
That consistency changes how signal data is used. Instead of evaluating intersections in isolation, agencies can identify patterns, rank priorities, and understand tradeoffs across corridors, districts, and regions. This network wide view supports more efficient workflows and better use of limited resources.
From Reporting to Insight: Unlocking AI’s Potential
For years, transportation data has largely been used to explain what already happened. That is beginning to change.
With long term historical data and increasingly consistent real time feeds, agencies are moving beyond static reporting toward insight driven analysis. This evolution helps streamline workflows, reduce manual analysis, and allow staff to focus on decisions rather than data assembly.
This is the point where data becomes a strategic asset rather than a byproduct of operations.
What’s Next: Forecasting and “What If” Scenarios
Another theme at TRB was forecasting. The next evolution isn’t just understanding what has happened, it’s anticipating what will happen and how we can plan accordingly.
By combining:
- Rich historical datasets
- Consistent real-time data streams
- AI and machine learning models
We can begin enabling engineers and planners to explore if–then scenarios:
- If a work zone is placed here, how does congestion change?
- If signal timing is adjusted, what happens to travel time reliability?
- If demand grows by X%, where do bottlenecks emerge?
These capabilities further improve workflows by allowing agencies to test ideas virtually before investing time and resources in the field.
Final Takeaway
TRB 2026 made one thing clear. Progress in transportation AI depends as much on discipline and collaboration as it does on technology. Agencies, researchers, and industry are increasingly aligned around the idea that better outcomes start with better data and better workflows.
As AI becomes the expectation, the real differentiator will not be who deploys the most sophisticated models. It will be who invests in consistent data, clear standards, and practical analytics that help teams work more efficiently and make confident, defensible decisions. The future of transportation is not just about smarter tools. It is about turning high quality data into insight and insight into action.