QA trailblazers are rewriting the rules in the age of AI

Julius Mong (Perforce) discusses the challenges of traditional software testing, specifically the issue of flakiness where testers spend 70% of their time fixing automation scripts rather than testing. They explain how current automation relies heavily on manual locator identification and code generation via LLMs, which breaks when UI designs change. He introduces the concept of agentic AI, which focuses on user intent rather than code or locators, allowing for automation that can handle pop-ups, visual variations, and multiple languages without constant script updates. The presentation concludes by suggesting that AI will not replace human testers but will instead require them to evolve into QA orchestrators who use AI to scale their capabilities.