Use case
AI Retail
See the shelf, predict the demand, help the shopper.
Camera analytics, forecasting and in-store assistants on your dedicated node, fed by the cameras and tills you already have over an encrypted tunnel.
AI Retail · Feature 01
CCTV heatmap
Existing ceiling cameras become a live map of where shoppers walk, stop and reach. See dwell time by fixture, dead zones by hour and how a new display changes traffic. Every frame stays on hardware reserved for you.
- Works with the cameras you already have
- Dwell, footfall and reach per fixture
- Compare layouts before and after a change
AI Retail · Feature 02
Shelf and queue vision
Detect out-of-stocks, planogram drift and price-tag mismatches as they happen, and open a till when the queue grows past your threshold. Alerts reach staff handhelds in seconds.
- Out-of-stock alerts in under a minute
- Queue length triggers for staffing
- Planogram compliance scored per bay
AI Retail · Feature 03
Demand forecasting
Store-level forecasts that combine sales history, weather, local events and the heatmap, retrained every night over your private fabric. Orders and staffing follow the forecast automatically.
- SKU-level forecasts per store
- Retrained nightly on local data
- Feeds replenishment and rosters
AI Retail · Feature 04
Multilingual kiosk assistant
A voice and chat assistant that knows your catalogue, promotions and store layout. It answers where things are, what is on offer and whether it is in stock, in the shopper's language.
- Grounded in your live stock and promotions
- Voice and touch, 30+ languages
- Runs on your dedicated node, never a shared cloud
Use cases
Where AI Retail goes to work.
Deployments customers run today. Each one starts on a single node and grows with the fleet.
Footfall and layout optimisation
Saturday's heatmap shows shoppers pooling at the bakery and skipping the new drinks endcap entirely. The store manager moves the endcap to the aisle mouth on Sunday night. By Wednesday the agent reports dwell time on it has tripled and drink sales are up 18%, with a before-and-after view she can forward to head office.
Fitting-room and till queues
It is the lunchtime rush. The camera above the tills counts five people waiting and the fitting rooms are full. The agent pings the floor lead's handheld with one message: open till three and send someone to the changing rooms. The queue is gone in four minutes, and the roster for next Saturday already reflects it.
Shrink prevention
A shopper slips two razor packs into a coat pocket in aisle 4. The agent flags the unusual shelf activity, attaches an eight-second clip and alerts the guard by the exit, who politely offers to help at the self-checkout. The video never leaves the store's dedicated node and is deleted after the retention period the retailer chose.
Stock forecasting
Every night the pharmacy's agent reforecasts each branch from sales, weather and the local school calendar. It sees hay-fever season starting a week early in the north and raises antihistamine orders there before the shelves empty, while trimming cough syrup in the south where demand has passed.
Also included
Everything in the box
Shelf and queue vision
Detect out-of-stocks, planogram drift and queue length from existing cameras in real time.
Demand forecasting
Store-level forecasts that combine sales history, weather and local events, retrained nightly on the node.
POS and ERP integration
Connectors for common point-of-sale and inventory systems keep recommendations in sync with stock.
Multilingual kiosk assistant
A voice and chat assistant that knows your catalogue, promotions and store layout.
Runs on
Supported nodes
AI Retail runs as agents on these nodes. Models are served from a Node Cluster or an API you choose. Pick a node and deploy it with AI Retail included.
Dedicated
Node One
Up to 50 Agents at once
one-time purchase25000000
Dedicated · 2U
Node Pro
Up to 250 Agents at once
one-time purchase45000000
Dedicated · 8U
Node Cluster
Hosts models up to 405B
one-time purchase750000000