Pilot a robot
Test a robot against a real operational use case, with measures agreed up front and an independent read at the end.
How a pilot runs
- 01Scope the use case
One task, one place, one shift pattern. We write down what the robot has to do and what would count as working.
- 02Shortlist the machines
Two or three candidates drawn from evidence in the Cyborg Index — what has actually been demonstrated, not what is claimed.
- 03Agree the measures
Throughput, intervention rate, uptime, safety conditions and the cost of running the pilot. Set before the robot arrives.
- 04Run and observe
A fixed window on site with the vendor. We record what happened rather than what was presented.
- 05Independent read
A short written verdict: what the machine did, where it needed people, and whether a wider deployment is justified.
What we bring
- Evidence on every candidate machine
- A pilot design with measurable outcomes
- Vendor coordination and question sets
- An independent written verdict
What you bring
- A real task and a site to run it on
- An operational owner for the window
- Access to baseline performance data
- Safety and site sign-off
We do not sell robots, hold inventory or take commission on hardware. Pilot work is paid for by the organisation running it, which is what keeps the read independent.