You don't need new cameras
The fastest way to kill a safety project
Tell the customer they have to replace all their cameras first.
It happens constantly. Someone scopes a genuinely good system, and step one is a full hardware refresh across every site. Now it's a capital project. Now it needs more approvals, a budget cycle, a committee. Eighteen months later the person who championed it has moved on, and the whole thing quietly dies in a procurement queue. Nobody got safer. A lot of people got tired.
Add the intelligence, keep the infrastructure
Dhi was built specifically to avoid that. It runs on the CCTV you already have, and it works with the video management system you already run — Genetec, Milestone, Avigilon. No new wiring. No forklift swap. Nothing to rip out and re-cable.
The cameras stay where they are. The intelligence is the part you add.
This isn't a compromise version of "real" AI. It's the recognition that most sites are already sitting on perfectly good hardware that has simply never been asked to pay attention. The bottleneck was never the cameras. It was that no one can watch forty feeds at once and react in time.
Start with one camera
The other thing rip-and-replace gets wrong is the size of the first step.
You don't have to commit a fleet to find out whether this works. Pick one camera — the feed that actually keeps you up at night. A platform edge, a loading bay, a restricted door. Put Dhi on that single camera and watch what it catches over a week.
If it earns its place, you expand. If it doesn't, you've spent one camera's worth of effort finding that out instead of a year and a budget. "Start small" isn't the timid option here. For safety-critical work, it's the honest one: prove it where it matters, then grow.
Integration is a feature, not an afterthought
Riding on existing infrastructure isn't just easier to buy. It's easier to live with. Your team keeps the tools they already know. Your existing VMS stays the system of record. Dhi adds a layer of detection on top and sends events into the workflows you already have.
The best safety upgrade is the one that's actually deployed, not the perfect architecture still stuck in approvals. Keep your cameras. Keep your VMS. Add the part that was missing.