Current engagementFirst-stage PoC live across 25 stores · scale-up in progress
Applied AI and production engineering
Evidence-backed revenue risk review across a multi-site retail estate
A production Retail Control Center that correlates transaction and camera activity, prepares evidence and carries each alert through review and resolution.
Deployment scale25stores in the live first-stage PoC
The problem
Manual review made it difficult to connect transaction anomalies with the right camera evidence and a traceable investigation outcome.
What we delivered
Built a non-blocking event pipeline that reconstructs completed transactions and evaluates configurable revenue-risk conditions.
Correlated transaction timelines with camera activity and prepared the relevant evidence window for review.
Delivered alert ownership, investigation, classification, resolution and audit workflows in one control center.
What is proven
✓The first-stage PoC is live across 25 stores.
✓Scale-up beyond the initial estate is currently in progress.
✓It remains outside the billing path, so loss of oversight does not stop a store from serving a customer.
✓Reviewers can claim, investigate and resolve alerts with operator and timestamp attribution.
Technical methods
Authenticated transaction ingestionIdempotent event processingConfigurable risk evaluationVideo and transaction correlationOperator workflow and operational diagnostics
Scope of this evidence
This evidence covers the live 25-store first-stage PoC and its operator workflow. Scale-up and commercial-impact measurement continue against the customer’s agreed baseline; customer and site details remain confidential.
