It sounds like the product is intended to help you avoid performance-related regressions, and argues that the risk of performance regressions will increase with AI writing more code. That seems plausible, but it's unclear how or why this product is superior to existing performance testing as part of your CI/CD pipeline.