Testing strategies for AI-integrated tools and system stability

Hi everyone, I’m currently setting up some automated workflows using Genny and am looking into how to properly handle non-functional requirements when integrating voice AI into CI/CD pipelines.

I’ve been researching best practices to ensure system stability under load and came across this guide on non functional testing. https://testomat.io/blog/the-basics-of-non-functional-testing/

Has anyone here implemented stability or performance testing for their AI-integrated tools? Would be curious to hear how you handle latency and reliability checks when using Genny in automated environments.