Instacart shares their comprehensive approach to integrating AI into software development workflows, detailing Project Tomato's findings on effective AI use cases like automated refactoring, design-to-code conversion, and debugging assistance. The company built Fizz, a group ordering app, in record time using AI-assisted development. Key insights include treating AI as a teammate rather than replacement, starting with low-risk tasks, optimizing development environments for AI effectiveness, and implementing structured learning programs. The post outlines both successful applications (ripe tomatoes) and current limitations (unripe tomatoes) of AI in engineering workflows.
Table of contents
AI-Driven Development at Instacart: Scaling Impact and Increasing VelocityFrom Grassroots Adoption to Scaled EnablementCase Study: Project Tomato — Accelerating Workflows with AI✅ Ripe Tomatoes❌ Unripe TomatoesCase Study: Fizz — Building Fast and SmartBest Practices for AI-First DevelopmentStructuring Success: Workshops, Playbooks, and Mindset ShiftsFinal ThoughtsSort: