Hourly density and queue time from existing cameras; checkout staffing plan is 12% more efficient.
A supermarket chain planned checkout staff from past sales data. Ilmek Akış derives anonymous people counts and queue time from existing IP cameras; the heat map shows density along shelves and at the entrance.
Checkout staff were scheduled dynamically by hourly density; average queue time fell 31% and staff productivity rose 12%. No facial recognition was used and images were anonymised.
| Queue time | -%31 |
|---|---|
| Staff productivity | +%12 |
Surface defects on cast parts are caught at 99.2% accuracy without stopping the line; customer returns halved.
Intake forms and referral documents are read automatically; registration time fell from six minutes to 90 seconds.
92% of 4,000 daily documents flow directly into the system; the remaining 8% are completed on a verification screen.
