Case Study: Semgrep achieves 20% fewer false positive security alerts with Anthropic Claude in Amazon Bedrock

How Semgrep delivers AI-powered code security with Claude in Amazon Bedrock

Semgrep, a cybersecurity company, faced the significant industry challenge of overwhelming developers with high volumes of false positive security alerts that wasted valuable engineering time. To address this, they leveraged Anthropic's Claude platform, specifically Claude in Amazon Bedrock, to bring AI-powered context understanding to their security tools.

By implementing Claude, Anthropic's solution empowered Semgrep to intelligently analyze code and filter out noise. The results were substantial, including a 20% reduction in false positive alerts, a 92% user agree rate on its safe-to-ignore labeling, and 16% higher accuracy in identifying false positives compared to their previous model. This allowed Semgrep to process thousands of security findings daily with greater precision and efficiency for their customers.

Semgrep

Bence Nagy

Staff Engineer