Researcher and analyst agents in Copilot

- [Instructor] Copilot in Excel is your AI assistant inside the workbook. We've asked natural language questions like, "Summarize sales by region," or, "Find trends in the data." But now we're entering what's called the agentic era, where AI doesn't just assist but it starts to act. So inside Excel we can actually call on a few specialized agents. These are prebuilt task-oriented versions of Copilot designed to work with or around our data depending on the use case. In this example, our workbook is Analyst Researcher and we have global CO2 emissions. What I'm going to do here is open Copilot. We're going to go to chat here, and if we go to this navigation pane here, there are going to be a few prebuilt, like I said, agents for us. We're going to focus on the first two, Researcher and Analyst. So if I open Researcher, what we'll see here is we can use this to get assistance to understand our data higher-level reasoning. It actually doesn't really read this data per se. It's going to go out to its wider LLM knowledge. So if you want some contextual support here, for example, "List major global policies that influence emissions since 1990," I'm going to go ahead and ask for that. And this is going to be a fairly in-depth research report. So this will take a while. Often, you are going to be getting some clarifying questions here, if there is something specific that you want: "Are you only interested in specific countries?" "Are you interested in different policies?" I just want something high-level, so I'll go ahead and just use the first option. And Researcher will head off here. Like I said, this will take some time to put this all together and it actually will kind of show you step-by-step what it's doing. If you expand this, you'll get a little bit more detail about what it's actually doing here. So be patient. Okay, after a while of thinking and thinking, we do get a report here and it's pretty thorough. We get resources, we get summary tables, in my case, and again, with this being generative AI, this might look a little bit different. And let me add some space for us here. You may or may not get a summary table like this, but we do have resources, links, and so forth. And this just helps us, again, kind of add context and understand what's going on with our data here. Now let's head over to the analyze agent. So I'll go back to our navigation panel and we'll head down to Analyst. So this is going to work a little different. This one does interact with your Excel data. We are going to prompt it, and this kind of works like the advanced analysis with Python feature, especially the Think Deeper feature. So I'm going to ask something like this. And you do see there are some suggested cards here. "Find the top five countries with the highest CO2 emissions in the most recent year." So Analyst is going to go through our data and similar to Researcher, it will go, step-by-step, list its reasoning steps and give us a result. So we do have the result here. We could continue on with the visualization. We do have the various steps. And a lot of the times, if you notice, see when I click on here, we'll get a little bit more detail. In this case, we are using Python step-by-step to complete these results. So that's pretty interesting. Let's ask for one of these. Yeah, sure, "Compare emissions with previous years." We'll see where this goes. So we're getting this nice research report and in a lot of ways, this is a much more user-friendly and seamless experience. And it does have that agentic feel to it where it's working a little bit more autonomously, it's adding a little bit more reasoning to its steps here. So we have an interesting chart here. This is Python code. I believe you can copy it, you can download it. So some of these things you can kind of hold onto. And we do have our insights here. So if you like this stuff, I would suggest copying it and moving it elsewhere. Otherwise, there's not a great way to save this history, unfortunately. But these two agents really show us some of the future applications of a agentic AI here. Researcher gives us the background, Analyst gives us the data-driven answers, and together they're showing how AI and Excel is evolving from simply answering basic questions to orchestrating reasoning, connecting knowledge, and executing tasks.