AI Agents for the Frontline: 7-Eleven's GenAI Maintenance Assistant

Maintenance technicians at 7-Eleven rely on preventive maintenance, catalogue, and parts documents for equipment like HVACs and ovens—often using only their phones. These documents span thousands of files in formats like PDFs and spreadsheets stored in shared drive, making it hard to find the right information quickly. High attrition and missed training sessions compounds knowledge gaps. To address this, 7-Eleven developed a GenAI agentic system using Mosaic AI, Agent Bricks and Databricks Hybrid Search that enables technicians to 1) Retrieve relevant maintenance documents with concise answers 2) Capture equipment images and receive troubleshooting solutions 3) Access parts information by equipment brand 4) Get solutions from the internet. This chatbot bridges knowledge gaps with instant access to institutional knowledge. We will walk through how we built this system, our evaluation methodologies and the 3-step prompt framework including guardrails for safe internet searches.