Visitors
Footfall by hour, day and branch, with entries and exits counted separately.
Retail analytics for Saudi operators
Turn your existing cameras into smarter decisions — without sacrificing privacy.
PrivacyVision reads operational behaviour, not identities. It measures how many people arrive, where they wait and how long queues run, so your managers act on numbers instead of guesswork — while every face is blurred on the device before anything is sent.
Faces blurred on device · No video stored · Data stays in KSA
Privacy is the architecture, not the brochure
Processing at the edge
Frames are analysed on an appliance in the branch. The video never leaves the store network.
Faces blurred on the device
Anonymisation is applied before any data is transmitted — not afterwards in the cloud.
No raw video stored
PrivacyVision keeps counts and timings. There is no video archive to request, leak or subpoena.
Data stays in the Kingdom
Analytics metadata is stored on Saudi-hosted infrastructure under your tenant.
Built for PDPL
Data minimisation, no biometric identification and no identity tracking, by design.
What PrivacyVision measures
Seven operational signals, all derived from anonymous counts. None of them identifies a person.
Occupancy
How full the branch is right now against its declared safe capacity.
Queues
Queue length per checkout and service area, lane by lane.
Waiting time
Average and maximum wait against the target you set.
Heatmaps
Aggregated activity density across the floor — never a path or a track.
Crowding
Density that crosses your threshold, with the action to take.
Branch performance
One branch against another, this period against the last.
From camera to decision
Five steps. Only the last two ever touch the cloud, and by then there is nothing personal left to protect.
Your existing CCTV
Standard IP cameras. No new hardware on the ceiling.
Edge AI appliance
An in-branch device analyses the stream locally.
Privacy processing
Faces are blurred and frames are discarded on the device.
Anonymous metadata
Only counts, timings and densities leave the branch.
PrivacyVision dashboard
Managers see the numbers and the action to take.
We analyse behaviour. We do not identify people.
There is no face recognition in PrivacyVision, no identity matching and no customer profile. The system counts anonymous presences and measures how long they wait. A returning shopper is a new count every time — by design, because the alternative would require knowing who they are.
Read the privacy architectureWhat PrivacyVision measures
- No facial recognition and no biometric templates
- No identity tracking across visits or across branches
- No customer profiles and no personal records
- Counts, not people: every figure is an aggregate
Branch intelligence for the people who run them
Compare branches on the measures that change a decision: footfall, peak period, queue performance and occupancy trend.
- Branch against branch, on one axis
- Peak periods, by day and by hour
- Queue performance against your wait target
- Occupancy trend and utilisation
- This period against the last, like for like
Alerts your team can act on
Each alert names the place, the measurement and the next step. No sirens, no noise.
Crowding
Queue wait
Occupancy threshold
Camera offline
Edge device degraded
Analytics interruption
Ask PrivacyVision
An analytics assistant for management questions. It answers from your own data, inside your own permissions — never from a general-purpose search of your database.
The assistant can only reach the branches your account can already open, and it runs a fixed set of reviewed analytics functions — it is never given free access to the database.
Built for the person on the shop floor
- Which branch had the longest queues yesterday?
- What is the busiest hour this week?
- Which branches need operational attention?
- Compare the Riyadh branches this month.
- What changed compared with last week?
See it on your own floor plan
We will walk through a live branch, the privacy architecture and what the first month of data looks like.