Case Study: Laurel achieves production-grade validation for AI-written code with Signadot

Laurel cuts change failure rate by 82% with Signadot

Laurel, an AI-native company building timekeeping and intelligence software for professional services, faced a challenge in validating a high volume of AI-generated code changes. They needed a production-grade validation solution that could operate without code or customer data leaving their own cloud due to strict compliance obligations for their law and accounting firm clients. Their previous approach using Vercel preview environments could not accurately replicate their full production stack.

Signadot addressed this by deploying per-PR sandboxes directly on Laurel's own Amazon EKS cluster. This solution, which integrated with their GitOps workflow, allowed for realistic validation of changes without data leaving their AWS account. As a result, Laurel saw a significant 82% reduction in change failure rate, dropping from 1.1% to 0.2%, while the share of agent-authored engineering pull requests increased to 76%. Signadot enabled Laurel to safely scale its throughput of AI-written code.

Laurel