How To Access Autonomous Agents in Copilot

Speaker A: I have been talking to a ton of people about Microsoft Copilot lately and what I am finding is nine out of 10 of them are not truly happy with the results. So what I did over this past weekend is I sat down and looked at how I can unlock agentic use within Copilot, because that's what they were talking about at Microsoft Build last week. And in three steps in this video, I'm going to show you exactly how you could do that, unlock it and start to get really good value and an ROI out of Microsoft Copilot. For those of you who don't know me, my name is Ryan Staley. I work with companies for AI transformation and basically upskilling the entire team, scaling revenue without adding headcount. And I'm seeing amazing things starting to happen now. One of the things that brought this to my forefront is I was at Gartner CSO event, which is for Chief Sales Officers. There's about a thousand people there, maybe 1200. And a lot of the executives there were in the sales space, as you can imagine, Chief Revenue Officer, Chief Sales Officer, Chief Commercial Officer. There were some rev ops leaders and enablement leaders as well, but really, really big on sales leadership. One of the things that I had as a reoccurring theme in the conversations is a lot of those organizations are ranging from 500 million to 5 billion, 10 billion plus was that they had Microsoft Copilot, but they weren't getting a ton of value out of it. Besides the basics and what I mean by the basics are writing, email, summarizing transcripts, things along those lines. So like I said, I started going through and working and seeing how we can unlock this value. I'm going to show you this right now, today. So let me open up my instance of Copilot and I'm going to show you how you could access this and start taking things to the next level with what's possible. Okay, so as you see here, this is in the researcher frontier Agent, if you will. You're probably running like, Ryan, where the hell is that? I haven't seen that. And it's funny, I had to take some time and do this. So click on the get agents tab that I just did on the right hand side and this is what's going to come up. So you're going to see apps come up like this. Now what the interesting thing about it is there's a lot of different areas where you could dig into this. And this is kind of overwhelming. There's a lot of companies that built these now what I looked and figured out through my research is if you identify the built by Microsoft, this is how you access those frontier agents. So as you can see right here, this is the researcher and the other frontier agent is the analyst. Now think of the analyst as looking at more chunks of data and then reverse engineering outputs from that and it could do it in a lot of different formats. The other one, which is the researcher, that's more like deep research, which is something that OpenAI has had on effectively since February. And I was really confused as why copilot did not have these uh, capabilities embedded within there or the forefront. So what I did is I basically found those, researched it, and then what you do is you click on it and once you click on it, I already said get agent. So if you get the agent, it'll put that in your sidebar. Okay, now what you're going to see is this is really, really strong because I added this one and the analyst. So now when you go back to the COPILOT tab, what you're going to see is they're over here on the side. So I basically have two different agents I could leverage to do deep work that are very, very strong in terms of the outputs they deliver. Now what you're going to see here is this is the prompt. So this is step two. I'm basically saying as a Big four consultant, execute a democ driven data research flow which is define, measure, analyze, improve and control data on. And I gave the old example of AI upskilling for non technical employees. Right? It's a little meta, that's what we're talking about today and I basically asked it to do it like a concise executive summary. Now what's going to happen when you use this agent is basically it's going to ask you for clarifications or confirmations on exactly what you want. The reason why it's doing this is because the fact that it's going to work for sometimes 10 minutes or 15 minutes, comb a ton of different sources and it really wants to give you an amazing output. Now I gave it the answers and this is the executive summary. So basically I'm getting a big four consulting executive summary on this and complete report now just from one prompt and a little bit of feedback. So what it's starting to do is it's updating this and the interesting thing about this is it started to search across my data. So some of these ROI metrics are actually real ROI metrics that I had with clients and it's identifying exact areas Now I wonder if, and I didn't like go back and check it, but basically say don't use my data, only use external data. But what this did is it looked through my entire drive and all the areas that I was leveraging in terms of presentations, roi and it was pretty, I don't know, I was impressed with this aspect of uh, it because I haven't seen deep research done or agentic use case. So to give you an example, basically this looks across files I had and it talks about what the ROI in the program was, how many hours a month were saved, and then you know, what the average rating was that I was getting in surveys. So some really, really strong indicators there. Uh, it also identified scope and objectives. So like what the initiatives were, the sales track, the marketing track, and it talks about metrics and performance over the last three months. So it talks about just improvements in terms of deal size. What were some of the outputs that we saw, increases in average deal side and percentages. And so like I was pretty blown away with this. So this is one of the biggest areas that I've seen companies struggle with is really trying to understand and look across massive mounds of data that they have and get insights. So it's also looking at trends, right? Exponential pipeline growth, after the curve, efficiency gains translating into volume and widespread usage adoption. Right. So every rep has basically so far had AI as a good thing in their qbr. So it even talks about some of the qualitative components and then as you could see, it talks about the strengths of the initiatives, right? Areas for improvement and then it finishes it up. It even talks about benchmarking, right? So now uh, granted it's referencing an older source so I guess I could give it like hey, only work within 2025 or other areas. But it was really interesting because it was a combination of what I've done from the analysis side as well as what I did, you know, with actual clients. Right. So this is an example of how you can unlock that analyst. Works pretty much the same way if you click it. So as you can see it look at uploaded files, it'll get quick insights and at the same time it could create um, a table with the volume of planets. Like here's an example of how it work in this situation for one of the pre built prompts. So what you could see. And so obviously it breaks down the planet, the volume, the volume multiple by earth. Now, right? That's not necessarily the best business use case but, but you kind of get the concept of how it could work. Super Fast. So once again, these are both the frontier agents. Like I said, that's the simple path that you take. Step one is basically get the agent. So you go and click get agent. Step two is basically filter for Microsoft built ones. Then you're going to see these two frontier agents. Three is pin these to the side like what you're seeing here and then you could start to leverage them. Now the other thing like I said is make sure you probably prompt them very strongly. Uh, I've noticed if you sometimes use the normal copilot model to prompt them and create a very good or long and detailed deep research prompt, it generates really good outputs. So um, just to kind of summarize, that's really simple and easy way and how you could do this other use cases and this is a little bonus, right that I've seen is you could develop entire business strategy plans simply from deep research with one prompt. Okay. So how I've seen that take place is I've seen C level executives say hey, we want to move into a new vertical and we want to move up market. At the same time they'll create this report and they're like, hey, this report would have cost $50,000 from basically a consulting company if we would have done this in the past, right? And it's creating this level of detail. You could use it on competition, you could use it on pricing models, you could use it basically on anything that you want. You just have to give it that focus and area. Now I've even seen it as another bonus. This is bonus part two. I've even seen it done to reverse engineer frameworks of experts or uh, entire books or other areas and then use that as training data to create your own GPT. Okay, so this is something that I've been working with a lot of executives on. I'm not going to get too much into it now, but maybe if you like that I can go into that in the next video. But feel free to drop your comments below. Anyways, I just hope you appreciate this video. Like I said, this was one of the biggest unlocks that I saw after hearing so many people say how frustrated they were with copilot. So I wanted to provide this for you, uh, to create value so that you can understand that there's massive unlocks for this as long as you know where to look in the right place. Thanks for joining me today and we will see you all on the next video.