Agentic Testing: Where Agents Fit in the E2E Testing Stack
Here's a 3-sentence summary of the engineering blog post: Slack engineers ran over 200 agentic end-to-end (E2E) tests to understand how agent-driven approaches fit into their testing stack, finding that agents can be highly reliable and suitable for complex workflows despite high execution costs. The experiment compared three execution models: agent + Playwright MCP, agent + Playwright CLI, and generated Playwright tests, with the MCP model showing the most consistent near-zero failure rates and the generated tests running the fastest but with lower reliability on complex flows. The findings suggest that agentic testing can be an effective addition to the testing stack, particularly for complex scenarios, but with considerations around reliability, execution time, and cost.