One store
With the existing CCTV, after an approved site survey.
The pilot exists to validate the product in a real store and convert into an annual contract. It is paid, with credit on conversion: commitment is part of qualification.
With the existing CCTV, after an approved site survey.
Fixed price, with credit on conversion to an annual contract.
Fortnightly checkpoint with the sponsor and a monthly report.
Analytics and Digital Twin according to the product phase at the time.
| Criterion | Target |
|---|---|
| Counting accuracy | Error of up to ±5% against manual counts in agreed samples |
| Processing coverage | At least 90% of the agreed hours processed on time |
| Actionable insights | Three or more, documented and acknowledged by the sponsor |
| Tested decision | At least one operational or layout change decided with pilot data |
| Adoption | Weekly use by the manager or analyst |
At least three one-hour windows, on different days and times, counted manually by a person from the customer and checked against video. The comparison is recorded in the report. Divergence above the target triggers a correction plan and a new measurement.
Processing is the same as in production: there is no test mode with less privacy. At the end, the data is deleted or formally migrated, as decided by the controller.
The targets above are engineering goals to be validated, not contractual guarantees.