Case Study: Datadog builds trust in AI code review with Greptile

Datadog surfaces ~5,700 issues per month with Greptile

Datadog, a monitoring and analytics platform, faced a challenge in scaling its code review process to keep pace with a significant increase in AI-generated code output from its internal Opus 4.5 system. Their engineering teams were concerned they lacked the time and resources to properly review this volume of code, potentially impacting quality and accountability. They evaluated several tools, including Greptile's AI code review service.

Datadog implemented Greptile after it outperformed other tools, including in-house and artisanal options, in an unbiased trial. The solution was adopted organically by engineers who began to trust its recommendations and even shared their "Greptile scores" with peers. This integration into the review culture helped ensure code quality and shared accountability for AI-assisted work. The implementation resulted in approximately 5,700 issues surfaced and 3,800 code reviews completed per month.

Datadog