Question

A project manager opens a lengthy proposal document in Word and asks Copilot to shorten the introduction. Later, she switches to Outlook and asks Copilot to draft a reply to a client email about the same project. She notices Copilot produces very different results in each app. What best explains why Copilot behaves differently between these two apps?

Accepted Answer

Copilot draws on the context of the app currently in use—the open Word document in one case and the email thread in the other—so each response is grounded in different source content.

Answer

Copilot in Outlook always searches the public web while Copilot in Word never does, causing the difference in output.

Answer

Each Microsoft 365 app runs a completely different underlying AI model, so responses cannot be consistent between apps.

Answer

Copilot in Word ignores the document entirely and relies only on the user's typed prompt, while Outlook reads all mailbox data.