Bolstering Retail Profits with Loss Prevention

AI can help reduce shrinkage at stores and help drastically improve store performance

Client Background

A top US-based grocery retail chain with over 300 stores across the country approached us for solutions for the revenue shrinkage they were facing due to bottom-of-the-basket (BOB) losses. With over a 100-year-old legacy, the client is one of the top 20 largest retailers in the United States. Our AI experts evaluated the extent of the problem and realised that shrinkage due to undetected items at the bottom of the basket was leading to millions of dollars in lost revenue. After a careful assessment of the underlying issues and requirements, Cogniphi designed a Vision Intelligence-based solution framework to help the client overcome the challenges they were facing across their grocery store outlets.


Challenges and Solutions

Loss Prevention:

Revenue shrinkage due to items that are left undetected at the bottom of the basket (BOB) is a major problem for most modern retailers across the world. The client was encountering a similar issue; throughout their grocery stores in the US, hundreds of customers checking out would invariably have an item/s at the bottom of the basket that would go unnoticed while billing took place. With scores of customers to be served in a time efficient manner, cashiers would virtually have no way of detecting these items and the BOB problem continued to translate into revenue losses for the client.

  • Using AI-powered Vision Intelligence, Cogniphi’s team of software developers and engineers devised a solution that leveraged the store’s already existing CCTV camera network. Video feeds streaming from cameras inside a billing counter were trained to detect objects that were left behind in a trolley and send out immediate signals to Cogniphi’s AIVI engine. This would be followed by a real-time alert being sent out to a cashier/billing agent at the POS counter who would cross check whether any item has been left out during final billing and thereby avoid a potential BOB-related loss.
  • This Vision Intelligence augmented technique of spotting un-emptied items at the bottom of a basket/trolley was implemented at both regular checkout lanes and self-checkout counters. At self-checkout counters, the real-time alerts were programmed to be sent out to the store supervisor instead of cashiers/billing agents.

Buddy Billing/ Sweet hearting

This is a common form of theft by employees at the cash register where they give away items to a “sweetheart” customer such as a friend or family member or coworker by scan avoidance.

  • Cogniphi’s team worked on an AI-based solution wherein camera feeds could communicate with the POS software and correlate data in real-time. Cogniphi’s AIVI system could verify whether or not the POS software has completed the scanning process for every sale item. In cases of a data mismatch, image-based evidentiary alerts would be sent out to the store manager for identification of repeat offenders.

Value Delivered by Cogniphi

  • 70% reduction in over-the-counter losses.
  • Real-time detection and alert system designed using stores’ existing CCTV networks.
  • Reinforced learning enhanced decision-making dashboard, reports and notifications for real time inference, alerts and action.
  • BOB was accounting for more than 65% of the retailers shrink. Post VisionAI implementation BOB instances has reduced by 80%
  • Increase daily per-lane profits: Continuous monitoring for unscanned items has helped boost profits per lane, per day by up to 10 percent
  • Improve throughput and front-end efficiency: Cashiers no longer have to lift heavy products from beneath carts. Transactions move smoothly for cashiers and customers alike
  • Identify cashier compliance problems: Whether it’s operational error or cashier-customer collusion, retailers can quickly uncover and address cashier compliance problems
  • Capitalize on flexible, cost-effective scalability: The solution’s industry-standard open architecture facilitates easy integration of next-generation retail technologies into store infrastructures