AI-Driven In-Store Operations Intelligence

The client needed greater visibility into in-store staff activities to ensure consistent service quality and effective operations. Limited insight into customer interactions and daily tasks made it difficult to measure engagement, optimize workflows, and maintain standards across stores.

Challenge

• No systematic way to monitor staff activities and customer interactions
• Standards varied across stores without objective measurement
• Managers relied on periodic walkthroughs rather than continuous data
• Solution needed to work with existing CCTV infrastructure

Solution

• Real-time face detection, pose estimation, and staff activity classification
• Edge AI deployment for fast, on-device video stream processing
• Real-time people detection and tracking integrated with existing CCTV systems
• Dashboard-ready traffic metrics for operational decision-making

Architecture

Outcome

• 97-99% staff activity detection accuracy across all monitored stores
• 30-40% incident response time reduction through real-time alerting
• 20-30% store operational compliance improvement via data-driven management
• System integrated seamlessly with existing CCTV infrastructure

Tech Stack

• Backend: Java
• Frontend: React
• AI/ML: Python, YOLO, TensorRT
• Infrastructure: AWS, Edge GPU