Delivering AI at scale in public transit

Transit agencies today face increasing pressure to deliver reliable and safe bus service in increasingly congested urban environments. One of the most persistent and preventable sources of disruption comes from vehicles stopping or parking in bus-only lanes, creating avoidable delays and heightening safety risks for operators and riders. AI-enabled Automated Camera Enforcement (ACE) has emerged as a practical tool to address this issue by capturing clear, time-stamped images of illegally stopped vehicles and their license plates.

These images are reviewed and processed through established enforcement workflows, enabling agencies to issue citations consistently and defensibly. But the effectiveness of ACE programs is not determined by technology alone. The real challenge lies in deploying AI systems at scale, within live operations, in a way that delivers measurable value while minimizing risk.

As demand grows, many vendors have entered the market to install cameras and configure ACE software, but very few have the expertise to help transit agencies operationalize AI as part of critical public infrastructure. At Hatch, our work in AI-enabled transit programs focuses not on the novelty of technology, but on the disciplined delivery models required to ensure AI performs reliably, defensibly, and at scale.

Integrating AI, not just installing it

In practice, the success of ACE programs depends less on hardware and more on governance, integration, and execution. AI systems deployed in public transit environments must function consistently across large fleets, be transparent and auditable, and integrate seamlessly into daily operations that cannot pause for experimentation.

When AI is introduced without strong oversight, agencies risk performance variability, rework, and reputational exposure. AI should be treated as critical infrastructure. This approach is grounded in experience delivering complex, technology-enabled transit programs where systems must perform under real-world conditions from day one.

Three considerations agencies must get right

Based on our work supporting large-scale ACE deployments, including the largest bus fleet in North America in New York City, three considerations consistently determine whether AI-enabled enforcement programs succeed:

1. Governance and defensibility: AI systems used in public infrastructure must operate within clearly defined and approved configurations. They embed governance directly into delivery by validating software versions, system configurations, and operating parameters, and by ensuring that updates or changes are introduced in a controlled and documented manner. This approach supports consistency across fleets while providing a well-documented account of system behavior that can be used to demonstrate adherence to regulatory requirements.
2. Integration with live operations: ACE systems are deployed across vehicles that are part of an active transit fleet, requiring close coordination with fleet operations and maintenance management. They work with on-site teams to align installations and commissioning activities with scheduled vehicle downtime, ensuring work is completed before buses return to service—without disrupting daily operations—while maintaining clear visibility into schedule, quality, and risk. This level of planning gives agency leadership real-time insight into delivery progress and constraints, enabling informed decisions as programs scale while preserving service reliability.
3. Program control and documentation discipline: Successful ACE programs depend on disciplined documentation and quality control. They ensure that configuration data, inspection results, and acceptance records are complete, consistent, and traceable across the program. This structured oversight enables informed decision making at key milestones and gives agencies confidence that systems are delivered in accordance with approved requirements while supporting audit readiness and long-term program continuity.

Business impact

When implemented effectively, AI-enabled ACE programs can reduce reliance on manual enforcement, improve consistency, and resolve issues faster through real-time visibility and oversight. In many cities, bus lane enforcement has been associated with improved bus speeds and service reliability, directly supporting performance goals. Importantly, ACE programs can generate revenue that agencies reinvest into other critical priorities—supporting safety initiatives, service improvements, and state-of-good-repair needs.

The value of AI in transit is not just technological; it is operational, financial, and strategic.

Proof at scale

At Hatch, we are supporting this shift through our role with New York City Transit on its ACE program. To date, more than 1,000 buses have been retrofitted, commissioned, and placed into service, with additional deployments planned. What distinguishes this effort is not simply the number of installations, but the disciplined way AI is governed, validated, and operationalized within an active transit fleet.

This capability is reinforced by Hatch’s broader experience overseeing several thousand technology retrofits across transit systems, including digital passenger information platforms and onboard communications systems. That combined delivery experience allows us to apply repeatable processes, informed judgment, and structured oversight to AI deployments in complex operating environments.

Looking ahead

Across the global transit sector, agencies are increasingly evaluating AI-enabled applications for safety monitoring, fleet condition assessment, predictive maintenance, passenger information systems, and operational decision support. As AI becomes more embedded in transit operations, the differentiator will be effective deployment.

Agencies that establish strong governance, operational integration, and oversight structures today will be better positioned to unlock long-term value while managing risk as these technologies scale. Hatch helps clients make that shift—turning AI from a promising technology into a dependable, operational part of transit service.