Everything Announced at Microsoft's AI Future of Work Event

Speaker 1: Welcome and thank you for joining us. We are here today to talk about something that's very fundamental to the human experience, the way we work. Speaker 2: Introducing Microsoft 365 co-pilot, your co-pilot for work. Co-pilot combines the power of large language models with your data in the Microsoft Graph and the Microsoft 365 apps. To turn your words into the most powerful productivity tool on the planet. [00:00:30] Co-pilot is integrated into Microsoft 365 in two ways. First, it works alongside you embedded in the apps. Millions of people use every day Word, Excel, PowerPoint, outlook, teams, and more. Second, we're introducing an entirely new experience business chat that works across all of your data and apps, your documents, emails, calendar chats, meetings, and contacts to do things that you've never been able to do before. [00:01:00] And co-pilot is built on Microsoft's comprehensive approach to security, compliance, privacy, and responsible ai so you know it's enterprise ready. Using just your own words. You can put co-pilot to work, gathering information in real time from across your data apps to unleash creativity, unlock productivity, and up-level skills. Now you can be more creative in word, more analytical and Excel, more [00:01:30] expressive in PowerPoint and more productive in Outlook and even more collaborative in teams. Speaker 3: You ask co-pilot to visualize what contributed to the decline in sales growth. This period co-pilot adds a little color to make the problem jump off the page. Now you want to dig deeper and ask a follow-up question. Co-pilot not only answers your question, it creates a simple model of the scenario [00:02:00] and it even asks you if you want to learn more about what it did with a step-by-step breakdown. Finally, you can ask it to create a graph of your projected model. Copilot turned a sea of data into clear insights and actions. Let me show you how co-pilot can help you tackle your email. We all struggle with the pull of the inbox. Sometimes it keeps us from focusing on our [00:02:30] most important work, but at the same time, you do have important emails to respond to. Co-pilot separates the signal from the noise and gives you hours of time back. Co-pilot can help you triage your inbox, highlighting the most important emails to prioritize and co-pilot also helps you on the go in Outlook mobile. It can summarize long email threads. You can draft a reply [00:03:00] using data from an Excel file, and copilot will generate a reply for you. Speaker 3: You can edit it, you can make it more concise. You can even change the writing style or add more context. Now all that's left to do is review the proposed response and hit Send. Co-Pilot is a productivity game changer. Speaker 2: Let me take a moment to explain what you just saw. [00:03:30] To build co-pilot, we didn't just connect Chat G P T to Microsoft 365. Microsoft 365 co-pilot is powered by what we call the co-pilot system, a sophisticated processing and orchestration engine. It harnesses the power of three foundational technologies, the Microsoft 365 apps, that's word, Excel, PowerPoint, outlook, teams, and more. The Microsoft Graph, that's all your content and context, your emails, files, meetings, [00:04:00] chats and calendar, and a large language model or L L M A creative engine capable of parsing and producing human readable text all accessible through natural language. It starts with a prompt from you in an app, co-pilot, pre-process the prompt through an approach called Grounding Put simply grounding improves the quality of the prompt so you get answers that are relevant and actionable. One of the most important parts of grounding is making a call to the Microsoft Graph [00:04:30] to retrieve your business content and context. Speaker 2: Co-pilot combines this user data from the graph with other inputs to improve the prompt. It then sends that modified prompt to the L L M. You may be familiar with what an L l M can do based on your experience with chat. G P T or Bing Chat co-pilot takes the response from the L L M and post processes it. This post-processing includes additional grounding calls to the graph, responsible AI checks, [00:05:00] security compliance and privacy reviews and command generation. Finally, co-pilot sends a response to the user and commands back to the apps co-pilot iteratively processes and orchestrates these sophisticated services to produce a result that feels like magic. Speaker 4: It looks like you have a dentist appointment that conflicts with your weekly business review. What if you could do both? Speaker 4: With co-pilot, you can follow a meeting and catch up to see what you missed, [00:05:30] even if you can't attend. You'll get a notification in Teams when the recap is ready and you can see the content that was shared, detailed notes and any action items. You see there was some discussion about your customer. You can ask co-pilot specific questions to get more context and it provides a detailed response. You can ask questions like why a certain decision was made or what other options were considered, and it citations with a snapshot of the transcript. It's [00:06:00] a huge time saver, so that was pretty cool. But the real magic of co-pilot happens during a live meeting. Here. You're in a meeting with your team. You can ask co-pilot to summarize what's happened so far. You can see who said what, what points are made. Co-pilot is really capturing the spirit of the discussion. As the meeting progresses, you can check on where people stand. You can even ask co-pilot what questions are unresolved. You may not have heard of Loop. Loop is a multi-player [00:06:30] collaboration canvas with co-pilot as an AI peer to your team along with everyone else. Let's try it out here. The team is collaborating in real time on product offers and pricing, any notes and action items to keep everyone on track and up to date. Speaker 4: You can ask co-pilot to pull data from Excel on the top, selling products by category and within seconds it's pro processing the information. Another member of your team jumps [00:07:00] in and adds a column and anyone can prompt co-pilot to build on each other's ideas. Everything is captured in prompt history and Speaker 5: We not only built grounding into the system, we also built verification into the experience. We added citations and introduced forms of friction to encourage back and forth collaboration and included feedback mechanisms so people could let us know when we get it wrong. Our goal is to give people agency, tell people get started. Co-pilot suggests good [00:07:30] prompts. If you don't like what it does, there's the try again button. You always have the option to use, adjust, discard, or undo. You saw that in the demos today. Speaker 2: We're deeply committed to listening, learning and adapting with speed to help create a brighter future of work for everyone with ai. Thank you.