Using Azure OpenAI GPT 4o to audit retail planograms and product assortments. The AI auditor inspects each shelf of a cooler photo, determining if it complies with the planogram rules. The system identifies the presence of required SKUs and reports any discrepancies or competitor products.
- Automated Auditing: Uses AI to analyze images of cooler shelves.
- Compliance Checking: Evaluates whether the products on each shelf meet specified planogram rules.
- Detailed Reporting: Generates an audit table that summarizes compliance for each shelf.
- Azure OpenAI with GPT 4o enabled
- Open the .env.sample file and replace the placeholders with your Azure OpenAI and Search credentials, save the file an name it
.env.
The payload sent to the AI model consists of structured messages, including:
- System Message: Describes the auditor's role and the product SKUs being assessed.
- User Task: Specifies the auditing task, including the rules for each shelf and the expected output format.
- Planogram Rules: Detailed requirements for each shelf, specifying the number of units for each SKU.
- Image Data: The image of the cooler shelf is included as base64-encoded data.