DoorDash Deploys AI to Automate Quality Checks for 19,000 Agents

Traditional quality assurance methods, which relied on limited manual sampling, struggled to keep pace with DoorDash’s rapid global expansion. The new system replaces binary scoring with deep conversational intelligence, allowing the company to identify product friction points and sentiment shifts as they happen. According to Xenia Strunnikova, Head of Customer Experience, Fraud, Trust & Safety S&O, the goal was to avoid automation for its own sake and instead capture the behavioral drivers behind customer feedback.

This transition allows human teams to move away from repetitive manual scoring toward high-value analysis and agent coaching. Joaquin Dufeu, Director of Strategy & Operations for Customer Experience & Integrity, noted that issues which previously required weeks to surface are now detected almost immediately. By integrating these signals, DoorDash aims to bridge the gap between reactive support and proactive operational response. The collaboration with Observe.AI and AWS focuses on augmenting human judgment rather than replacing it, ensuring that accountability and empathy remain central to the platform’s decision-making process.