How to run user acceptance testing for software
Prepare realistic work scenarios and agree which issues block release before users begin testing.

In short
- Ask people with different roles to complete full tasks.
- Observe where people stop or go back to an old workaround.
- Decide which defects block launch and which can have a documented temporary route.
- Next steps
On this page
Cover a realistic working day
Ask people with different roles to complete full tasks. Include a normal job, an incomplete record, a rejection and a cancellation. Use recognisable examples without exposing real private data.
Check results beyond the screen
Observe where people stop or go back to an old workaround. Check saved statuses, audit events and data sent to connected systems. A correct-looking interface can still produce an incomplete export.
Define who can sign off
Decide which defects block launch and which can have a documented temporary route. Rerun critical scenarios after fixes. Name the person who confirms readiness and handles new reports after launch.
Next steps
Creagrid / actie
Practical checklist
Describe the problem and intended outcome.
Involve the people who use the process every day.
Test the riskiest assumption with a small prototype.
Define tests, ownership and maintenance up front.

