AI and Playwright MCP for web testing

In this talk, Jaswanth Manigundan shares a practical view of how AI and the Playwright MCP server can support day to day web testing. He draws on his experience as a tester, automation engineer, and engineering manager to show how these tools can be used as helpers rather than replacements for human thinking.

Jaswanth starts by grounding the session in familiar testing practices, such as exploratory testing and traditional automation with tools like Selenium and Playwright. From there, he introduces AI agents and the Model Context Protocol (MCP), explaining them in simple terms and cutting through the hype. He frames AI as a set of tools that can speed up work, reduce setup effort, and help testers learn faster, not as something that magically does testing for you.

The talk includes hands on demonstrations using Playwright MCP inside an editor. Jaswanth shows how AI can automate browser actions, explore a web application, suggest test ideas, generate a starting test plan, and help set up a Playwright project from scratch. He also walks through using AI to write and run automated tests, inspect failures, and review Playwright reports and traces.

Throughout the session, Jaswanth emphasises the importance of context, judgement, and ownership. He explains why testers should stay in control of test strategy and architecture, and how AI works best as a collaborative assistant that handles repetitive or setup tasks. He also discusses where these approaches fit alongside other testing types, tools, and frameworks.

By the end of the talk, you will understand what Playwright MCP is, how AI agents can support web testing workflows, and how to use these tools responsibly to learn faster, work more efficiently, and focus on the thinking that really matters in testing.