What acceptance tests are needed after robot software updates?

Why

Begin by recording the old and new software or firmware versions, release notes and a change-impact assessment. Identify workflows that depend on modified motion planning, tip tracking, liquid classes, labware definitions or integration APIs. In a nonproduction environment, run representative and worst-case test scripts with traceable inputs and expected outputs.

Verify volume performance and sample maps with physical measurements, not only simulation. Challenge emergency stop, interrupted runs, network loss, bad barcodes and restart behaviour where those failure modes matter. Confirm the exported sample or result schema and timestamps are unchanged or deliberately migrated.

Document deviations, retest decisions, approvals and the rollback procedure. A cosmetic interface change may require less testing than a firmware motion change, but neither should be declared safe solely because the software launches.