Every warehouse has the same poster on the wall: "STAY CLEAR OF FORKLIFTS." And every warehouse has near-misses anyway, because posters don't watch aisles. Here's the full build of a system that does — one depth LiDAR, one model, and an agent short enough to read in a sitting.

1. Mount and enrol

The Spatial Robot's SPAD dToF covers 8 m of choke point — the intersection where near-misses actually happen — from a single pole mount, unfazed by skylight glare that would wash out an indirect ToF sensor. One cable, then visionlibra fleet enroll cam-047, and it appears in the dashboard.

2. Load the model

The forklift-safety model classifies forklifts and people from pure depth — no RGB, so it works in the dark and sidesteps the privacy questions an ordinary camera raises in a workplace. It tracks both classes and emits a person_near_forklift event when their trajectories converge. It runs on-device, in real time.

3. Write the agent

Subscribe to the event and wire three actions: post to the safety Slack channel, log a structured record to your database, and generate an incident report with the depth clip attached. That's the whole program — it's the exact code on the developer quickstart.

4. Watch it change the building

Events land in Slack in under a second. The dashboard heatmap shows convergence hotspots by aisle, which — in every deployment so far — moved a pallet staging area within the first week. That's the difference between AI and an AI agent: the first one sees, the second one changes the floor plan.

Why depth, not RGB, for this job

A depth sensor gives you distance in metres directly, so "a person is within 2 m of a moving forklift" is a measurement, not an inference from pixel size. It works in the dark, tolerates glare, and produces imagery a works council will actually approve, because you cannot identify a face in a depth frame. For safety and presence tasks that combination is hard to beat — the reasoning behind it is in our piece on why physical AI starts with spatial intelligence.

The Spatial Robot ships this pipeline out of the box, and the Spatial Hub runs the model on-site for a whole fleet — see the store.