Get more relevant AI Proofing reviews with workspace and project memory

AI Proofing can now learn reusable preferences from recent, similar human proofing work, helping your first AI review produce more relevant findings without extensive setup.

Workspace memory captures shared terminology and recurring conventions, while project memory keeps client, brand, tone, and spelling preferences specific to that project. Explicit instructions and configured guidelines take priority over inferred preferences.

Learning starts when you enable AI Proofing for your workspace or start a one-off AI review. A one-off review uses similar proofs you can access without enabling workspace-wide automatic reviews.

Who it helps
Teams with existing proofing history can carry recurring review preferences into AI reviews with less manual configuration. Teams working across clients or brands can keep project-specific conventions separate.