A readiness risk was closed by a stakeholder confirming the manager roles were live. The first end-to-end run produced a Highest-priority defect: the manager could not open the Approvals app.
"Testers just pester people." Why questions are the cheapest test available, and what a QA audit found when it looked at teams that had stopped asking them.
A QA methodology written out in full in two files, and every customer journey documented twice — hand-written Gherkin beside notes from 313 executed cases. Both copies had drifted.
Same-tier discounts looked like interchangeable price options on one bundle. In the billing system each was a separate product, so none of them could be collapsed into a representative case.
A status called "Ready for SIT Deployment" sat in the Done status category — so every roll-up counted the work as finished with SIT, UAT and production still ahead of it.
Rather than re-run a manual journey to find out whether a fix had deployed, a script read the class source and classified it as old, new or mixed logic — then audited the test class too.
A story suppressed a contract document and an order confirmation for add-on discounts. The hard part was proving the suppression did not take the billing discount down with it.
Eight green tests on a discount-approval resolver asserted every failure path and none of the intended behaviour. A @TestVisible bypass had also routed the production query out of the suite.
UAT started on time with an environment, users and a plan — and nobody knew where to send the first defect, or who verified fixes on which side of the contract.
A tester closed a requirement, a developer changed the same area, and nobody could say whether what happened next was a retest or a regression. The two words are not interchangeable.