A seller at a software company opens an opportunity record just before a renewal call and wants a quick recap of the last month of email exchanges and meeting notes without reading every thread manually. She uses Copilot in Dynamics 365 Sales to generate a summary directly on the opportunity form. Which statement correctly describes how this summary is produced?

The correct description is that Copilot analyzes emails, meetings, and notes already tracked to the opportunity in Dynamics 365 and produces a natural-language synopsis of key points and recent activity. Copilot for Sales generates opportunity and record summaries by reading the activities, emails, meetings, and notes already tracked against that specific record, then producing a synopsis a seller can scan before a call. This matches the scenario, where the seller wants a recap built from what has already been logged to the opportunity. The statement saying Copilot searches the seller's entire mailbox, including personal and untracked emails, to build the most complete summary possible is incorrect because Copilot does not reach into a seller's full mailbox or surface untracked personal correspondence; it works from records associated with the Dynamics 365 entity, respecting existing tracking and security boundaries. The statement saying Copilot pulls only the most recent single email and rewrites it in a more concise tone is incorrect because the summary draws on the full set of tracked interactions over the relevant period, not a single most recent email, which is why it can recap a month of activity rather than one message. The statement saying Copilot generates the summary by re-running the opportunity's AI score calculation and translating the numeric result into sentences is incorrect because the summary feature and the predictive opportunity score are separate capabilities; the summary is a language-generation output over activity content, not a narrated version of the numeric AI score.