User acceptance testing (UAT)
A release check where real users or the client confirm the software supports the job and requirements they actually asked for.
See it
What it is
User acceptance testing asks the people who requested or will use the software to confirm that it supports the real work they need to do. Unlike a technical test that asks whether the code matches its implementation contract, UAT asks whether the delivered workflow meets the agreed business requirements well enough to release.
Run it in a production-like environment with realistic roles and data before a major launch, client handoff, or process change. Scenarios should come from acceptance criteria and real tasks, such as approving an invoice from submission through audit history. Teams often write them in Cucumber's Given-When-Then form, but the important artifact is a recorded result and an authorized sign-off.
Gotcha: UAT is not a late bug hunt or a substitute for automated testing. If requirements are vague, users can only approve their latest interpretation. Name the decision-makers, freeze the acceptance criteria for the test round, and record exceptions instead of treating a hallway 'looks good' as approval.
Ask AI for it
Create a user acceptance testing pack for <feature> from these signed requirements. Turn each business workflow into a Cucumber-style Given-When-Then scenario with role, prerequisites, realistic test data, steps, and an observable expected result. Add a Jira Xray test set with an owner and Pass, Fail, or Blocked status for every scenario, plus entry criteria, exit criteria, defect severity rules, an exceptions log, and a named business approver sign-off section.