expect(answer).toContain(“refund”) – PASSED ✅
Actual result: “Sorry, we never give a refund.”
Test automation is great at checking exact results. But more and more products integrate generative AI, and it never answers the same way twice. Your assertion passes when it should fail, and you can’t read thousands of answers by hand. Our answer: AI evals, automated checks that score AI answers. We built them for our own product. In this hands-on workshop, you will learn to add them to Playwright tests and leave ready to test AI at scale.
Senior Software Engineer, Infobip
Staff Quality Engineer, Infobip
Prerequisites – Your own laptop with VS Code (or another IDE), internet access, and permission to install software. We will send setup instructions in advance.
