Proof of concept
Test one robot task before you scale.
A PoC is a bounded exercise: one station, one SKU mix, clear pass/fail metrics. It is the fastest way to find out whether a task is ready for automation.
What it is
- + Single task definition and success criteria
- + Recorded runs on real objects with lighting variance
- + A short report with videos and failure tags
- + Recommendation for the next phase (stop, iterate, or expand)
What it isn't
- − Full factory rollout in four weeks
- − Guaranteed ROI figures
- − Custom hardware design
- − 24/7 unattended production
What we measure
Success rate
Completed cycles vs. attempts over a fixed test matrix.
Cycle time
Median and p95 time per pick-place or navigation leg.
Recovery behavior
How often the stack replans after slip, occlusion, or collision.
Typical timeline
01
Kickoff
Agree on objects, fixtures, and acceptance thresholds.
02
Integration
Mount the arm, calibrate cameras, and collect teleop demonstrations.
03
Readout
Share clips, metrics, and a plain-language go/no-go note.
Run in our lab or on your line
After the PoC
You either stop, rerun with new objects, or move to a wider pilot. No automatic lock-in.
Example task
Shelf restock in a store aisle
Client
Retail
Task
Front-face beverage cans on two shelf heights
Duration
Bounded pilot
Labels rotate, gaps vary, and the gripper must not crush slim cans. The policy uses vision plus light force limits.