Comments on: QA Metrics That Still Mean Something in the AI Era https://www.softwaretestingmagazine.com/knowledge/qa-metrics-that-still-mean-something-in-the-ai-era/ Mon, 21 Sep 2026 20:12:55 +0000 hourly 1 By: Enrique Sanchez https://www.softwaretestingmagazine.com/knowledge/qa-metrics-that-still-mean-something-in-the-ai-era/#comment-50482 Mon, 10 Aug 2026 07:51:25 +0000 https://www.softwaretestingmagazine.com/?p=11774#comment-50482 The important words about software testing metrics in this article is that metrics should measure whether the software testing activity protects the product, not just how much software testing activity is happening.

]]> By: Karina Jenkins https://www.softwaretestingmagazine.com/knowledge/qa-metrics-that-still-mean-something-in-the-ai-era/#comment-50235 Mon, 03 Aug 2026 12:17:35 +0000 https://www.softwaretestingmagazine.com/?p=11774#comment-50235 The good question of this article: “if AI writes all these tests, why do customers still find bugs?” There is so much hype around AI for generating both code and tests, that I have the impression that by removing the human control on software organizations are just going drive more quickly towards the big hard wall of failure.

]]> By: Ihor Boiarynov https://www.softwaretestingmagazine.com/knowledge/qa-metrics-that-still-mean-something-in-the-ai-era/#comment-50292 Thu, 30 Jul 2026 17:32:02 +0000 https://www.softwaretestingmagazine.com/?p=11774#comment-50292 This hits the nail on the head, Yevheniia.

In the trenches, we are definitely feeling the weight of the “AI green dashboard illusion.” Leadership sees 1,000 new generated tests and a 98% pass rate and assumes quality is bulletproof, but half those tests are just checking that the application didn’t crash, rather than asserting actual business logic. You end up burning CI/CD minutes on fluff while edge-case defects still slip right through to production.

Your callout on internal new-defect yield and mutation testing is where the real value lies.

Shifting the conversation away from volume and back toward risk-weighted journey coverage and test freshness gives us a pragmatic way to keep testing accountable to the end user, not the machine.
Love the article. Thank you!

]]> By: Аlla https://www.softwaretestingmagazine.com/knowledge/qa-metrics-that-still-mean-something-in-the-ai-era/#comment-50167 Thu, 23 Jul 2026 22:18:10 +0000 https://www.softwaretestingmagazine.com/?p=11774#comment-50167 This matches exactly what worked for us, tying defect escape rate to release delay cost finally got leadership buy-in. Would love to hear if you found a metric that worked even faster than that.

]]>