Microsoft Copilot is an AI-powered chatbot developed by Microsoft and designed to boost productivity, unlock creativity, and help users understand information better through a simple chat experience.

Copilot leverages the power of large language models (LLMs) and works alongside popular Microsoft 365 apps such as Word, Excel, PowerPoint, Outlook, Teams, and more. It’s like having your own personal companion to assist you in your everyday tasks.

There are different versions of Copilot available, including a free version and a Pro version that offers premium AI features. For businesses, there’s Copilot for Microsoft 365. Each version is designed to meet different user needs and enhance productivity in different ways.

Using your business data, Copilot synthesizes information from documents, emails, and messages to give you summaries, answer your business questions, and generate draft content from your files, messages, and more.

Are there limits for how much I can use AI?

As with other AI tools, there are limits to how many turns (prompts and responses) you can do per chat conversation and per day. If you reach those limits, you will be notified.

• Microsoft 365 E5
• Microsoft 365 E3
• Office 365 E3
• Office 365 E5
• Microsoft 365 Business Standard
• Microsoft 365 Business Premium

Generative AI is AI that can generate new content - like text, images, etc. - on its own. Copilot is a generative tool: you can prompt it with something, and it will generate a response based on its underlying AI models and the web.

Copilot aims to base all its responses on reliable sources, but AI can make mistakes, and third-party content on the internet may not always be accurate or reliable. Copilot will sometimes misrepresent the information that it finds. This is why Copilot is transparent and shows the sources of information behind its answers. Always check the sources before making decisions or taking actions based on Copilot's responses. Copilot does not replace human judgment. 

Copilot presents only the data that each individual can access using the same underlying controls for data access used in other Microsoft 365 services. The permissions model within your Microsoft 365 tenant will help ensure that data will not leak between users and groups.

Why is commercial data protection important when using generative AI at work?

Company data and intellectual property are crucial to our competitive advantage. The problem with using generative AI without commercial data protection is that prompts or content you generate can be exposed to the public, as part of another answer. That's where generative AI with commercial data protection comes in: what you put in and get out is not used to train the underlying AI model so company data stays safe. 

How are we guaranteed that our data isn't stored by Microsoft?

Microsoft is committed to ensuring the privacy and security of your data. When you use Microsoft Copilot, your data is processed and used to provide the service and improve Microsoft products. However, it’s important to note that Microsoft does not store your personal conversations or use them to train AI models. The data is also designed to be forgotten after the conversation ends. Microsoft has strict policies and procedures in place to prevent unauthorized access and use of data. For more detailed information, you can refer to the Microsoft Privacy Statement. It’s always good practice to review the privacy policies of any service you use to understand how your data is handled.

Copilot prompts are instructions or questions you use to tell Copilot what you want. Prompts can include four parts: 

• Goal (what you want: "I want a list of 3-5 bullet points to prepare..."
• Context (why you need it and who is involved): "...for an upcoming meeting with Sam..."
• Expectations (how Copilot should respond); "Respond with a tone that's friendly and authoritative..."
• Source (what information Copilot should use): "Use email and chats over the last two weeks."

You can put a little or a lot into a prompt, but all that's required is a clear goal. If you want to be more specific, add other parts.

Writing good prompts is key to getting better results with Copilot.

• Include details like context, expectations, and source information in your prompts.
• Provide positive instructions. Tell Copilot what “to do” rather than what “not to do.”
• Iterate and regenerate, especially if you’re not happy with the initial results. You can revise your prompt to try again or try regenerating.
• Always review and verify the responses from Copilot, cross-referencing with trusted sources when needed.

Copilot is a generative AI service grounded in data from the public web in the Bing search index only. It doesn't have access to organizational resources or content within the Microsoft 365 Graph, such as documents in OneDrive, emails, or other data, even when commercial data protection is turned on.

Copilot for Microsoft 365 adds three things:

1. Access to data within the Microsoft 365 Graph in your tenant.
2. Prompts and responses are processed entirely within your Microsoft 365 service boundary, along with other Microsoft 365-specific security, compliance, and privacy features.
3. Access to this generative AI capability from Microsoft 365 applications like Teams, Outlook, and Word.

Copilot can help employees quickly generate content, analyze or compare data, summarize documents, learn new skills, write code, and much more. Eligible users also receive commercial data protection when they signed in with their work account (Entra ID).

For example, Copilot can help them:

• Understand the implications of a decision: 'What are the pros and cons of offline marketing strategies?' or 'How can I measure the success of my offline marketing campaign?'
• Learn new skills: 'What are the top five things I should know when managing a large project?' and 'What is agile project management and how does it differ from waterfall?'
• Analyze data: 'If we're forecasting 7% EPS growth this coming quarter, how does our internal forecast compare with EPS growth in the top US public CPG companies?'
• Summarize a work PDFs open in Edge: 'Recap the findings of this internal cybersecurity report and the top three concerns' or 'What are best practices for addressing these kinds of security issues?'
• Write better code faster: 'Write a regular expression in Python that matches email addresses' or 'How can I use the unittest module to write test units for my Python code?'
• Plan a business trip: 'Where should I stay in Manhattan that has a good running path close by?'
• Generate social media content: 'Use this messaging framework to generate five social media posts describing its value to healthcare workers.'