City DOTs do not need another sensor-first project
ITS program managers are under pressure to cut congestion, improve pedestrian and cyclist safety, and prove results without turning every corridor into a capital replacement program. The Seattle DOT’s University District Transportation Technology & Safety Improvements project shows a more practical path: use existing transportation assets, layer intelligence onto the corridor, and measure whether travel time and reliability improve.
That matters because many cities already have camera coverage at key intersections. The missing piece is not always another pole, cabinet, or procurement cycle. It is camera-agnostic monitoring that turns video feeds into operational awareness. The defensible position for traffic operations is simple: start with the cameras you already own, validate the outcome, then scale.
The University District project validates an outcome-first approach
Seattle’s University District is a demanding test bed: buses, bikes, pedestrians, freight, curb activity, and signalized intersections compete for limited right-of-way. In its project report, Seattle DOT describes transportation technology and safety improvements aimed at corridor performance, travel time, and reliability.
The lesson for other city DOTs is not that every corridor needs the same configuration. It is that measurable improvement should come before hardware expansion. When a major city demonstrates that transportation technology can cut travel times and improve reliability, it gives ITS leaders a stronger argument for phased deployment. A retrofit model lets teams test one intersection, one camera group, or one corridor segment before committing to broader rollout.
Computer Vision AI turns existing CCTV into operational awareness
nureal.ai’s Computer Vision AI is designed for this exact constraint: works with the cameras you already own. Pre-trained traffic and behavior models can detect objects, count and track pedestrians, track vehicles, recognize corridor patterns, and surface events that affect operations. The platform is camera- and infrastructure-agnostic, so teams do not need to replace CCTV just to evaluate a congestion or safety use case.
For a University District-style corridor, that could mean monitoring pedestrian volumes near crossings, vehicle queues at signalized intersections, or patterns of behavior that indicate recurring conflict points. This is monitoring, not identification. The purpose is to help operators understand corridor conditions in real time and compare those conditions against measurable transportation goals.
Operational workflows matter more than raw video
Computer Vision AI creates the behavioral signal, but city teams still need a way to act on it. That is where nureal’s three capability pillars work together: Computer Vision AI; Agentic AI; Generative AI. For this use case, Computer Vision AI is the foundation. Agentic AI can translate detected traffic events into operational alerts, automated response steps, or workflow prompts for signal timing and incident response teams.
A practical example is a corridor where operators see recurring travel time variability during peak periods. Instead of reviewing video manually after the fact, monitoring can flag queue growth, pedestrian surges, or incidents as they happen. Generative AI can optionally summarize event data into shift handover notes, giving the next operator a clearer picture of what changed and when.
Retrofit first, then scale what proves itself
The mildly contrarian view is that many city transportation teams are overbuying new sensors before they have exhausted the value of their installed camera fleets. New hardware has a role, but it should not be the default answer when existing cameras already cover intersections, crossings, and high-friction corridors.
A lower-friction path is to start with one camera. Activate one model. See it work in your environment. If the model produces useful operational awareness, expand to a few more intersections and tie the signals to a measurable KPI such as travel time reliability, pedestrian delay, or response time to detected events. nureal is ready to run day one, with edge inference available where on-device processing supports the operating model.
For procurement-aware teams, the point is not to avoid process. It is to reduce avoidable replacement scope and validate outcomes before scaling. Available on Sourcewell Contract #041525-NURL.
Talk to an expert.