found a QA team still writing test cases by hand in 2025
AI writes them now. runs them. spots the flaky ones.
the job didn't disappear. it moved up a level.
that part nobody's ready for.
test automation tools are "having a moment"
we've been automating tests at QAwerk for 10 years
the moment was always happening
what changed: the tools got smarter. the hype got louder.
noise isn't the same as signal.
found a qa team running 3 agentic loops in parallel last month
none of them knew what the others were testing
AI found bugs. also introduced 2 new ones
tools got smarter. process didn't.
we shipped 40+ automated test cases for a client last month
all playwright
the shift from selenium isn't hype. it's just less pain
browser automation finally works the way you expect it to
found a market report claiming AI test automation will 'revolutionize QA by 2032'
2032.
at QAwerk we've been doing it since 2015.
the future is already unevenly distributed.
a tool that writes playwright tests for every PR automatically
kinda wild how fast this became table stakes
we spent years convincing clients to automate at all
now the debate is which AI does it for you
progress is weird.
found a qa team that automated 80% of regression last year.
now leadership wants to cut the team in half.
"AI does it now"
no. AI runs the scripts. humans still decide what to test.
confusing the tool with the judgment.
found a forecast: AI test automation market hitting $36B by 2032
the number is real. the hype around it isn't.
most teams still can't write a stable test without it flaking on CI.
market size β problem solved.
6/6 So yes, AI is changing QA. But the teams treating that as "less testing needed" are the ones who'll have the most interesting post-mortems.
Ten years running QAwerk. The tools change. The need for someone who actually checks the work never does.
1/6 Everyone's calling agentic AI a "paradigm shift" for QA. Maybe. But I've seen enough hype cycles to know the gap between the press release and the production environment.
5/6 We started writing about EvalOps because testing AI outputs is genuinely different from testing deterministic code. There's no "expected result" when the system reasons its way to an answer.
2/2 The automated and manual testing work we did with them was delivered to a high standard, and weβve become trusted partners."
-- Benjamin Stuart, Product Manager at Kazidomi
More client stories: https://t.co/vKn1c7Gqrv
1/2 "QAwerk team has been of enormous help in improving the quality of our releases at Kazidomi. They are communicative and highly professional, with a proactive attitude.
1/9 Every few months a new article drops declaring AI has finally "solved" software testing. 2024, 2025, now 2026.
At QAwerk, we've spent the last year actually running these tools on real client projects. Not benchmarks. Real codebases, real deadlines, real bugs.
8/9 Ten years running QAwerk, we've watched every wave of "this will automate QA." Codeless testing, record-and-playback, model-based testing. Each one moved the floor up. None of them replaced judgment.