From cameras to evidence

Two projects of similar size signed contracts within a month of each other. Both wanted "site cameras". One ended up with a passive feed nobody opened after the first week. The other ended up with an evidence layer that survived a delay claim and a regulator visit. The hardware on the wall was nearly identical. The difference was the brief.

The first project bought cameras

Procurement asked for camera count, resolution, and storage. The successful tender did exactly that. The team got crisp footage and a viewer. There was no named workflow attached. There was no defined trigger that turned a clip into a decision. When the first programme dispute arrived nine months later, the team scrolled through hours of footage looking for a moment they half-remembered. Nothing was tagged.

The second project bought evidence

Procurement asked a different question. Which decisions on this project depend on a record that does not yet exist. The list named four: delay claims at the steel package, subcontractor attendance disputes at peak fit-out, an expected regulator visit on access, and the handover pack. The brief was written backwards from those decisions. The hardware was a means; the records were the end.

• Capture points were chosen by what they would prove, not by site geography.
• Retention windows were tied to the contract and to insurer requirements.
• A named owner was attached to every workflow before the cameras were powered on.
• A weekly evidence review went on the project director's diary from week one.

“The first project had hours of footage and no record. The second had fewer hours of footage and a defensible archive.”

The category boundary

Cameras are a means; RDI is a discipline. Buyers who can name the decisions they want to change have already crossed the boundary. Buyers who cannot will end up paying for the hardware twice — once for the install, and again in the cost of the disputes the record could have closed.