Blog
Notes on software QA
What we see across the industry: how QA gets bought, how teams get staffed, what counts as evidence, and where AI earns its place.
These are perspective pieces, not news. They come out of running conventional testing for product teams, and they say what we would say on a first call.
All articles
What AI changes in testing, and what it does not
AI is genuinely useful on the parts of QA that are volume work — first-draft cases, triage, spotting the gap in coverage. It is not useful for the one decision the whole engagement exists to support.
Read the articleJudgement is what you are buying
Execution is the visible part. What a buyer actually pays for is someone who has seen this failure before, in a different industry, and knows where to look.
The shift to AI-assisted testing
Separating the two AI shifts, choosing a tool by the problem it solves, keeping data and access under control, and measuring the whole workflow rather than the generation step.
Evidence every cycle, not a report at the end
A pass rate is not evidence. What a release decision actually needs, and why it has to arrive while the work is still happening.
Want this applied to your product?
A scoping conversation, and a straight answer about whether we are the right fit.