XBOX & Microsoft · 2026.03

Copilot Localization Review Agent

Localization errors were surfacing after publish. A custom M365 Copilot agent moved that review upstream, checking localized content against the source at import, before anything went live.

reactive QAbecamecaught at source

The problem

Localization errors (incorrect context, the wrong language shipped to a market, off-brand terminology) were being caught late in the production cycle, sometimes after content had already published to multiple surfaces. Each catch meant updates across several markets and a manual review sweep.

The approach

The agent I set up in M365 Copilot reviewed localized content against the source at the point of import, before it entered production. It applied rules for grammar, style, sales terms, and brand terminology. It flagged issues with enough context for producers to fix them directly or share specific examples with the localization team, even in languages they couldn't read. In our case, the review helped us identify errors and patterns we hadn't been aware of.

The result

Manual spot-checking gave way to an automated review step. Errors that used to spread across several markets got fixed before they left the producer's hands.

Before and after.
MeasureBeforeAfter
Review processManual spot checksAutomated review with producer follow-up
When issues were caughtLate in production, sometimes after publishFlagged at import, before publishing

Open to work

Got a manual process worth automating?