With this series of weekly emails, we’ve arrived at the “Show me the Money” moment, where ideas meet results. We know public dollars come with public accountability; that’s precisely why this matters.
Let’s take a quick step back before I show how one initial investment can unlock a cycle of reinvestment, where Operational Expense savings fuel the scaling of 20-40% faster transit trips, not just in one corridor or one city, but let’s say, across the entire State of California (or any State). And how we are planning on delivering this approach over the next year.
Seeking direct value for citizens and businesses: Waco (TX) SMART Grant
If technology doesn’t deliver measurable benefits for people and businesses, then why invest in it?
Funded by the U.S. DOT SMART Grant, the Waco MPO is deploying a multi-mode traffic-flow optimization and AI-Edge sensing solution to improve daily life directly, from boosting downtown vitality and local sales to cutting red-light delays for transit, improving student safety and travel across the freeway, and enhancing event access around the basketball arena.
Each outcome will be measured through data, surveys, and community feedback, turning innovation into proven results that matter to citizens.
Technology proven, operational, and delivering results
From Northern Europe to Marysville and Dublin (OH), the outcomes are consistent. Dublin City Council sought to prioritize pedestrians and allocate green time based on real-time demand – not just to optimize the main line. The result: a 28% reduction in pedestrian wait times and 4% less delay for all approaches, even on corridors already finely tuned.
In Marysville, the focus was reducing truck delays to minimize emissions and noise. Traffic Engineering reported a 15% reduction in truck delay, leading to 17% lower greenhouse gas emissions.
And safety? Dublin reported a 75% (!!) reduction in incidents – clear evidence that performance-based management benefits all modes.
The evolution of a city’s ability to serve through transportation
Transportation has followed a path remarkably similar to computing.
It began with manual control – officers and fixed timers responding by sight, much like early standalone computers. Then came coordinated signals, synchronizing corridors for smoother flow, the first local networks. Adaptive systems followed, using sensors to react to real-time conditions, the Internet’s early packet routing.
That’s where most U.S. cities stand today. Not for lack of effort, but because the tools have only recently matured.
The next phase is policy-driven optimization, where networks prioritize by community goals, as seen in Marysville and Dublin. Waco represents the next leap: a Controlled Operational Model, where infrastructure is managed like the Internet itself: adaptive, predictive, and continuously optimized.
In Stage 1, Waco will control infrastructure to deliver measurable impact for its four selected use cases. In Stage 2, the city will define local issues and opportunities by stakeholder, craft new use cases, and deliver the technology to drive results county-wide.
Paul, get on with it – show me the money! San Francisco MTA demonstrated this potential clearly. In a proof-of-concept along a ten-intersection corridor, red-light delay for light-rail vehicles was cut by 73%, requiring 18.6% fewer service hours while increasing speed and reliability. Transit planners estimated that 2-3 vehicles could be removed from service, each representing over $1.4 million per year in Operational Expenses. This success led the Metropolitan Transportation Commission (MTC) to fund expansion to 101 intersections (fund transfer pending).
Scaling the model
What if this same model were applied to other routes and agencies? The U.S. DOT reported $33.67 billion in 2024 Operational Expenses for Bus and Light Rail. We’re not suggesting we can remove all 30+% opportunity, but we for sure should aim for it to find the optimum between fiscal responsibility and service performance, especially since this could mitigate operator shortages and enable the retirement of older diesel vehicles that are costly and difficult to maintain.
The Transit Opportunity Index (TOI) Assessment
We developed the TOI Assessment, a data-driven framework that cleans, organizes, and analyzes large transit datasets to understand real-world vehicle behavior. It identifies where vehicles stop for passengers, where they fail to travel at posted speeds, and how long trips would take if they could flow at the posted speed. The outcome: a quantified measure of service-hour savings potential.
Together with the SamTrans Policy, Planning, and Transportation team, we first applied this to El Camino Real, then scaled it to all 74 routes spanning 20+ jurisdictions on the Bay Area Peninsula.
Findings were striking: In some cities, SamTrans could reduce service hours by 33% when intersections were predictable and delays minimized. In others, opportunities exceeded 50%.
Using the same method, SFMTA’s Route 49 (Van Ness), even after $170 million in capital investment, still revealed a 34.2% opportunity, confirmed independently by SFMTA.
In a recent TOI for a Los Angeles-area city, comparing four weeks before and after an AI-based TSP implementation, results showed no material improvement as they are conceptually still in the adaptive, not controlled, stage. To deliver impact, we need infrastructure to move vehicles efficiently through intersections and coordinate pedestrian phases safely and predictably.
The lesson is clear
Technology alone doesn’t deliver results; governance and measurement do.
Cities need structures that align around defined outcomes, support implementation cycles of under six months, adjust schedules, and improve running headway performance, delivering tangible benefits to riders and reinvesting operational savings into the next phase of improvement.
It’s encouraging to see this message resonate with city leaders, State Senators, transit agencies, MPOs, and advocacy groups. Staff have been assigned, Concepts of Operations written, and demonstration projects are now being formed to showcase scalability.