The AWS Machine Learning Blog reports that monday.com is operating agentic AI, termed "AI Teammates," in production environments on Amazon Bedrock. This deployment is notable for its scale, as few engineering organizations run agentic AI in production at this level.
Key Points
- monday.com's "AI Teammates" are agentic AI systems running on Amazon Bedrock.
- Nine out of ten Builders at monday.com use AI coding tools monthly.
- This represents an increase from approximately half of Builders using these tools six months prior.
- Per-engineer pull request (PR) throughput has increased by over 50%.
- All figures presented are derived from monday.com's internal production data.
Context
According to the AWS Machine Learning Blog, monday.com has implemented an architecture that supports these production AI agents within a codebase that is a decade old. The blog post details the retrofits made to enable this functionality and describes a confidence-scored merge process that aims to advance toward full autonomy.
Why It Matters
This report highlights a practical application of agentic AI in a production engineering setting, demonstrating how such systems can contribute to developer productivity and throughput within an established codebase.
What To Do
- Note the reported increase in AI coding tool usage and its impact on PR throughput.
- Consider the implications of running agentic AI at scale in a production environment.
- Observe how monday.com's internal production data informs these findings.