Ivan, co-founder of Amikoo (AMO), discusses how his AI-powered QA agent helps testing teams keep pace with accelerating code velocity driven by AI coding tools. AMO uses 12 specialized agents and 43 purpose-built tools to generate Playwright tests with page object model structure, suggest regression test sets based on pull requests, detect and bulk-fix flaky tests, and analyze test coverage gaps. The system is designed to be token-efficient — claiming 8-10x improvement over direct Claude Opus usage — by using narrow, purpose-built agents and custom MCP servers. Ivan argues that AI increases the need for testing rather than reducing it, and that testers should evolve toward test architecture and UX judgment roles rather than manual scripting. He recommends keeping humans in the loop for P0/P1 critical scenarios and treating AI agents like junior interns that need guidance and context.