Many people talk about automated tests and end-to-end testing. But once you try it, you realize it’s not that easy: those tests are time-consuming to create and maintain, QA engineers may lack coding expertise, and every major product change requires starting over!

AI has already transformed testing—LLMs excel at generating unit and integration tests from code, dramatically reducing boilerplate work for developers. However, end-to-end testing remained untouched until early 2025, when Model Context Protocol (MCP) allowed AI to control browsers directly. This breakthrough is particularly powerful for WordPress, facilitating sophisticated UI-based testing automation.

Drawing from months of experimentation at group.one (WP Rocket, RankMath, WP.one, Imagify), I’ll demonstrate how we leverage Playwright MCP to turn manual E2E testing into an AI-driven process.