FBN.2025.09
AI-Driven Customer Insights
Industrial trainee, Intel® Unnati
Computer Vision, Python, Machine Learning, Business Intelligence, 2025
Live
video analysed as it streams
CV+BI
vision models meeting business intelligence
3
decisions informed: placement, promotion, restocking
The problem
Retail teams fly blind inside their own stores: which aisles draw people in, where help is needed, when a shelf is about to run empty.
The approach
A working prototype reads live video with computer vision, following movement and dwell to reveal busy zones and preferences, then layers analytics on top to recommend placement, promotions and restocking, with live alerts for staff.
The outcome
Raw camera feeds turned into placement, promotion and restocking decisions, with alerts that arrive before a problem does.
What it does
3 capabilitiesBehaviour, observed
Computer vision follows movement through the store to show how people really shop.
Heat, mapped
Busy zones and preferences surface on their own from the feed.
Alerts in time
Live notifications for assistance and inventory, before the problem lands.
The strata
How it standsCapture
Real-time video ingestion
Vision
Models tracking movement, dwell and preference signals
Analytics
A BI layer turning signals into placement and restocking insight
Alerting
Live notifications for assistance and inventory events
Next exhibit — X
Bitcoin Price Prediction