One thing I've learned after 15 years in test automation — a flaky test is never "just flaky". It's always hiding a real problem. Race conditions, hardcoded waits, shared state. Fix the root cause, not the symptom.
Just set up a Playwright + Cucumber + TypeScript framework with parallel execution, HTML reports, and Azure DevOps CI/CD integration. Happy to help any team struggling with flaky tests or framework setup. Feel free to reach out!