Explanation:
Box 1: Yes
Yes - Microsoft 365 Copilot uses contextual information from your organization to suggest prompts.
Microsoft 365 Copilot leverages contextual information from your organization to suggest relevant prompts, a process aimed at improving productivity by reducing the need to search for information. This feature, often enhanced by Context IQ, scans your recent activity--including emails, calendar meetings, chats, and documents-to provide personalized, tailored suggestions.
Box 2: No
No - Microsoft 365 Copilot can use contextual information from your organization to train the large language model (LLM).
How Microsoft 365 Copilot actually uses organizational data:
*-> No Training on Customer Data: Customer data, including prompts, responses, and data accessed through Microsoft Graph (emails, chats, documents, etc.), is not used to train the underlying foundation LLMs that power Copilot.
Grounding (Contextual Use): Copilot uses organizational context to "ground" responses. It retrieves relevant data in real-time via Microsoft Graph to ensure answers are accurate and tailored to your organization.
Box 3: Yes
Yes - Microsoft 365 Copilot uses contextual information from your organization to suggest augment a prompt, before sending the prompt to the large language model (LLM).
It uses a process called grounding to improve the relevance and accuracy of its responses by connecting to your organization's data via the Microsoft Graph.
Instead of just relying on the general knowledge of a Large Language Model (LLM), Copilot acts as an intermediary, augmenting your prompt with real-time organizational context before the LLM processes it.