Selected work

Applied AI · Retail intelligence

Computer vision loyalty

Using computer vision and event-driven cloud services to make loyalty recognition work without a physical card.

Outcome

Cardless customer recognition connected in-store interest signals with relevant follow-up offers.

Context

A retail loyalty experience normally begins at checkout. This concept explored how customer recognition and in-store behavior could help the business respond earlier and more personally.

Challenge

Translate computer-vision signals into useful business events and customer workflows, rather than stopping at face detection as a technical demo.

Approach

Make the system work.

  1. 01

    Develop face-recognition capabilities for cardless loyalty identification.

  2. 02

    Model customer presence and shelf interest as events that downstream services could act on.

  3. 03

    Connect Azure Functions, API Management, Cosmos DB, Azure SQL, queues, and blob storage.

  4. 04

    Design mobile and web touchpoints around the event-driven service architecture.

Impact

What changed.

  • Extended loyalty recognition beyond the physical card and checkout moment.
  • Connected in-store behavior with timely, relevant follow-up opportunities.
  • Demonstrated how applied AI, cloud architecture, and business workflow create value together.

Disciplines

  • Computer vision
  • Event-driven architecture
  • Retail AI
  • Azure
Next case · 01Agentic SDLC governance