How long does a table stay occupied, and how many times does it turn?
Is the counter busier than the tables, and at what times?
It's raining tomorrow: how much should we prepare?
What is each cover worth?
Distinct customers and dwell time per table and per dining room.
Dwell time and footfall at the counter compared with the tables.
Sales by weather condition, product mix by temperature and a 7-day forecast.
From the till: covers, average ticket and best-selling products.
Waiting times and queues in pickup and drive-thru areas.
Breakfast, lunch, aperitivo: when people come in, and when they don't.
Kontrol Flow cross-references the forecast with your venue's history.
Illustrative example.
Kontrol Flow was born in the Ammu Cannoli stores in Milan: every metric was tested in the field before reaching your venue.

Go deeper into the metrics that matter most in your industry.
Movement heatmaps and dwell time by zone: average, median, P95 and engaged customers.
Average, median and P95 wait from waiting area to counter, queue, walk-outs and alerts.
How it works
Kontrol Flow reads the video streams from the IP cameras already installed in your store.
An NVIDIA edge device processes the images locally. Video never leaves the store.
Real-time dashboards, an AI briefing every morning and alerts when something is out of the ordinary.
Privacy
Numbers leave the store, not images.
Demographic data only in aggregate, optional, and enabled only after an assessment with your DPO.
Employees are automatically excluded from every number.
Images stay on site; aggregated data is hosted in Europe.
Yes: you start with a single venue and the cameras already installed.
Yes, integration with your POS system (e.g. Quenta) or via API. During the pilot we can also start from an export of your POS data.
For each venue we cross-reference the 7-day forecast with your footfall and sales history.
We work with the IP cameras you already have installed. We check compatibility together on the first call.
No. Images are processed on site; only anonymous, aggregated data leaves the store.