Andrew January

🚨 Playwright 1.59 just dropped… and it’s time to stop pretending this is “just another tool upgrade.” It’s not. I’ve sat through countless interviews where the JD screams Playwright… but the conversation still lives in the Selenium dark ages - “How do you handle waits?” - “What’s your strategy for flaky locators?” - “XPath or CSS?” Meanwhile, the tool itself has quietly evolved into something far more powerful. Look at what just landed 👇 → page.pickLocator() No more DevTools guesswork. No more copy-pasting broken XPath. Hover → Highlight → Click. Playwright hands you the perfect locator. Game over for trial-and-error. → locator.normalize() Takes your messy locators and intelligently converts them to best practices (test IDs, ARIA roles, etc.). Cleaner. Smarter. Less brittle. → Built-in auto-waiting on steroids + the brand new Screencast API Your tests don’t just run. They explain what they did, show where they failed, and create visual “agentic video receipts” with action annotations. The old world was: Open DevTools → Inspect → Try selector → Fail → Add arbitrary wait → Debug race conditions → Repeat forever. The new world? Deterministic execution + intelligent locators + built-in observability. With shared browser sessions, CLI debugging superpowers, and features clearly designed for both humans and AI agents… Your automation is no longer just code that clicks buttons. It’s becoming a system that understands the UI, documents its actions, and sets the stage for AI to debug and fix it. Playwright isn’t “better Selenium.” It’s redefining what test automation even means in 2026. If we keep judging modern tools with legacy habits — we’re not just falling behind. We’re missing the entire direction the industry is sprinting toward. The future of testing isn’t writing more code. It’s writing less code… that understands infinitely more. Who else is feeling this shift? Are your teams still stuck in Selenium-era thinking, or have you fully embraced the Playwright mindset? Drop your thoughts below 👇 And if you’re hiring or interviewing for Playwright roles — let’s talk about what questions actually matter now. #Playwright #TestAutomation #QA #SoftwareTesting #AIinTesting #AutomationEngineering

1

🧮 Great testers don’t stop at results - they analyse and iterate. With AI evaluations, exporting structured results (like CSVs) enables deeper insights, scoring, and continuous improvement loops. See how to build this workflow in our Practical AI Testing Guide: 🔗 https://lnkd.in/enB9a-UX #AITesting #AIResults #QATools #AIQuality #SoftwareTesting #Testmo

1

I had the pleasure of managing Greg when he was part of my development team, and his professionalism and attention to quality consistently stood out. He has a calm, methodical approach to testing and a strong instinct for risk that always helped the team make better release decisions. This kind of independent QA review is exactly the sort of pragmatic, high‑value support many teams underestimate until they’ve worked with someone like Greg.

1

BDD and AI assisted coded is an awesome combo. Have your user stories include acceptance criteria written in Given/When/Then (aka Gherkin) whenever the change impacts users. You can use an agent, don't write it yourself This stops the uneasy feeling when the work is completed where you think to yourself "really? This does not feel right" (even with excellent unit test coverage) I force my agents to get a pass 100% across lint, unit test, integration, BDD (end to end across components). I means I know it works from a user perspective. Across systems.

9

1 Comment

One thing I’ve noticed recently in QA: Writing automated tests is no longer the hardest part. Maintaining them is. Modern applications change constantly, and many automation suites become fragile over time: • selectors break • flows change • environments become unstable • false positives increase That’s why I think the future of QA automation is not only about creating tests faster… It’s about creating resilient testing strategies. Working with Cypress has shown me how important it is to build: • stable selectors • reusable commands • clean test architecture • reliable API validations • smart retry strategies AI is also starting to influence QA workflows, but even with AI-assisted testing, the foundation still matters. Bad test design + AI = faster chaos. Good QA engineering is still about reliability, maintainability, and confidence in every release. Tools evolve. Quality principles don’t. #QA #Cypress #SoftwareTesting #AutomationTesting #QualityAssurance #WebTesting #Testing

1

One thing I’ve noticed quietly damaging quality on many projects isn’t tooling, process, or even time. It’s handoffs. - Requirements handed to dev. - Build handed to QA. - Results handed back to delivery. And in between? Context gets lost. I’ve seen issues slip through not because no one tested — but because: - QA didn’t know why a change was made - Dev didn’t know what assumptions QA was validating - Product didn’t realise which risks were still open The strongest teams I’ve worked with reduced handoffs by sharing context, not artefacts. Less “here’s the ticket”, more “here’s the intent”. Quality improves when ownership flows, not when work is passed around. 👉 Where do handoffs create the biggest blind spots in your delivery today? #QALeadership #SoftwareTesting #DeliveryTeams #QualityCulture #AgileReality

6

4 Comments

🔥 If you explain BDD like this in an interview… you’ll get shortlisted instantly. Most candidates talk about Cucumber, step definitions, tags, and runners. But a real SDET explains BDD the way it’s meant to be used: ⭐ **“BDD is not a testing framework. BDD is a collaboration framework that produces testable requirements.”** That one line changes the entire interview. Here’s how I explain BDD clearly: 1️⃣ BDD starts before automation It begins with the 3 Amigos (BA + Dev + QA) discussing user stories. Goal → achieve shared understanding. 2️⃣ Gherkin converts business language → executable scenarios Instead of writing test cases in QA language, we write: Given (context) When (action) Then (expected outcome)** This becomes living documentation. 3️⃣ Automation connects Gherkin to code Using tools like Cucumber / SpecFlow / Behave, we bind scenarios to step definitions using code (Java/Python/JS). 4️⃣ CI/CD runs these scenarios as acceptance tests This ensures: ✔ Business rules aren’t broken ✔ Regression stays clean ✔ Requirements stay up-to-date 5️⃣ The true value of BDD Not the tools. Not Gherkin. Not automation. 👉 The real value is eliminating ambiguity in requirements. 70% of defects come from unclear requirements — BDD solves that. #AutomationTesting #SoftwareTesting #SDET #TestingFramework #TestAutomation #AutomationFramework #QualityEngineering #QACommunity

23