UAT is not QA, and treating it that way costs you

I have sat in too many UAT cycles that were really a second round of QA — the same test cases, re-run by tired business users who click through and sign off because the sprint is ending.

Real UAT starts from the operator's day, not the requirement list. What does the ops executive do at 9 AM? What does she do when the customer on the phone is angry and the record is half-saved? Those flows rarely map cleanly onto a single user story, which is exactly why QA never catches them.

I write UAT scripts as narratives with a person's name in them. "Priya receives a disbursal request that was rejected yesterday and re-submitted with a new document." Business users test that willingly, because it is their actual Tuesday.

The other half is evidence. Manual screenshotting is where UAT quietly dies — people stop capturing, and then a defect becomes an argument instead of a ticket. That frustration is literally why I built a browser extension that captures and labels evidence per test case; the tooling exists to protect the testing habit, not the other way round.

Sign-off should mean "we can run our business on this". Anything less is just QA with a longer email chain.