AWS ML BlogWednesday · June 3, 2026FREE

How Baz improved its AI Agent Code Review accuracy using Amazon Bedrock AgentCore

Baz, a company focused on developer tools, leveraged Amazon Bedrock and the newly introduced Amazon Bedrock AgentCore to build an AI agent for automated code review, specifically targeting specification compliance. The Spec Review agent uses large language models to compare code changes against project specifications, identifying deviations and potential bugs. By integrating with existing CI/CD pipelines, the agent provides real-time feedback to developers. The implementation resulted in higher review accuracy and reduced time spent on manual reviews. Baz reported catching specification mismatches earlier, which lowered the cost of fixes. The solution is built on AWS infrastructure, utilizing Bedrock's model selection and AgentCore's orchestration capabilities. This case study highlights how agentic AI can be applied to code quality processes, with implications for reducing technical debt and improving software reliability.