Use case · Brand & venue analytics
Brand interactions at the shelf. Seat occupancy and service time at the chair. Throughput through the door. mediaIntel reads the existing camera fleet and turns it into structured operational signal — anonymous-by-default, hour-by-hour, per-zone — so merchandising, store ops, and venue operators stop guessing and start tuning.
What the engine does on the floor
The same behaviour model that scores brand engagement at a display also clocks an unoccupied chair, a finished service, or a queue forming at the till. You pick the metrics that matter for the venue; we tune the model against the cameras you already have.
Display dwell, pickup-and-replace, planogram compliance, end-cap engagement, sample-station throughput. Per-SKU and per-zone — so the brand team can finally answer 'did the activation work' with footage, not foot traffic guesses.
Seat occupancy, service start and end events, idle minutes between clients, peak loading by hour, queue length at reception. The metrics a hairdresser, clinic, or restaurant operator currently logs by hand — collected automatically from the cameras already on the wall.
Face recognition is off. The models score behaviour, dwell, posture, and zone — not identity. Edge-blur removes identifiable detail before frames leave the camera, so the data is operational from the first frame, not after a privacy review.
How it fits the venue stack
Standard IP camera support, plus an edge runtime for sites where you can't backhaul video. Outputs are structured events your scheduling, POS, BI, or back-office tool can already consume — not a separate dashboard that nobody logs into.
Works on the existing camera fleet (RTSP / ONVIF). Edge runtime ships on NVR-class hardware for sites where the network can't backhaul live video.
Identifiable detail blurred at the camera before frames leave the site — so the events feed is operational by default, no per-jurisdiction privacy escalation required.
Push events to scheduling tools (Booksy, Mindbody, Square Appointments), to BI dashboards (Looker, Metabase, PowerBI), or to the existing back-office system via webhook or REST.
Our promise to floor and venue operators
We are building per-zone, per-shift behaviour models trained on real retail and service-venue footage, with the false-positive rate shipped alongside the model so operators can sign off with eyes open. We are pre-launch on this surface; these commitments are how we will structure pilots, not metrics from a fielded venue.
Anonymous-by-default
Behaviour models — identity not collected, not required
Same camera fleet
Standard IP and edge runtime — no rip-and-replace
Hour-by-hour
Granularity matched to how venues actually staff and merchandise
Per-zone, per-SKU
Display, end-cap, chair, station — tuned to the venue's layout
From conversations with brand managers, merchandisers, store ops leads, and multi-location service-venue operators.
No. The models score behaviour, dwell, posture, and zone — not identity. Face recognition is off by default and only enabled where a customer has a documented legal basis and an explicit per-tenant configuration. The default product surface is anonymous-by-default and identifies behaviours, not people.
Yes — standard IP camera support (RTSP, ONVIF). For sites where the network can't backhaul live video, the edge runtime ships on NVR-class hardware and runs detection on-site. Most pilots run on the existing camera fleet without a hardware swap.
Display dwell time, pickup-and-replace rate, end-cap engagement, sample-station throughput, planogram compliance (is the right SKU in the right slot), and zone-level traffic counts. Per-SKU and per-zone — so an activation campaign can be measured at the display, not the store.
Seat or station occupancy, service start and end events (inferred from posture and zone transitions), idle minutes between clients, peak-loading hours, queue length at reception. The structured signal flows into your scheduling and POS tools so the operator gets real numbers without manual logging.
Counting hands at the door tells you traffic. Measuring zone dwell and pickup tells you whether the merchandising actually worked. Measuring seat occupancy and service time tells you whether the operation is profitable. We add the behavioural layer on top of (or instead of) the existing counter — operators usually keep the counter and use us for the why, not the how-many.
Yes — structured events out via webhook or REST. We integrate directly with common scheduling platforms (Booksy, Mindbody, Square Appointments), and ship into BI tools (Looker, Metabase, PowerBI) or whichever back-office system the operator already runs. The value is in routing the signal into your existing workflow, not in standing up another dashboard.
Brand & venue analytics — single-site pilot
Tell us the floor or the venue you want to start with, and the two or three metrics that would actually change a decision. We'll scope a one-site pilot inside three weeks, with the privacy-impact assessment in writing.
