The computer vision model segments, recognizes and interprets images of store shelves and their products, placements and signage.

‍

The interpreted images, ideal planogram information like price info, product metadata, image metadata, and full product names are processed using a GPT. Additionally, we also input stock management system info and photo metadata.

‍

The GPT gives outputs:

• Stock Missing

• Partial stockouts

• Planogram Verification

• Placement Analysis

• Price Verification

• Discount Analysis

‍

• A significant reduction in staffhours, saving up to 1000 staff steps a day through automation of repetitivetasks such as stock and label monitoring

• An unbiased point of verification –improving efficiency, ensuring repeatability, alerting store representatives asneeded

• Reduced stockouts leading to morereliably available products, increasing sales and average basket value

• Streamlined store tasks andplanogram integrity