EP 385: Microsoft Copilot - Autonomous AI agents released into the wild

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

AI employees are here, well, in a month. So this is no longer one of those things, theoretical things that we're talking about here like, oh, we're going to have AI autonomous agents that'll be able to help us with our everyday tasks. I think for the past two years we've maybe been a little bit early when talking about autonomous AI agents. But now Microsoft just had an announcement in bringing in its co pilot studio autonomous AI agents to all of us next month, not next decade, not next year, next month. AI employees that can handle those mundane day to day tasks that you can train with natural language. Yeah, I'm excited to talk about that in today's show. What's going on, y'? All? My name is Jordan Wilson, the host of Everyday AI. But before we get started, let me first give a little shout out to our partners at Microsoft Worklab. So the Worklab podcast from Microsoft is made for leaders who want to understand the future of work. It offers expert insights on everything from how to approach digital transformation to how AI can help unlock, unlock more value from your company's data. That's W O R K L A B no spaces, available wherever you get your podcasts. Speaking of podcasts and live streams and daily newsletters, that's what everyday AI is. Thank you for tuning in, y'. All. Like I said, my name is Jordan, I'm the host and this show is for you. This is your daily live stream podcast and free daily newsletter helping everyday people learn and leverage generative AI to grow their companies and to grow their careers. So I mean, one of the best ways to do that, I mean autonomous AI agents, if that doesn't get you a awake very early this Tuesday morning right here, I'm, I'm live on the road, taking the Everyday AI Show a little bit on the road today. If that doesn't get you excited, I don't know what will. But if you haven't already, please go to your everyday AI.com Sign up for that free daily newsletter. All right, before we jump into it, let's start as we do every single day by going over the AI news. And hey, for our live stream audience, got a little secret live stream poll. So go ahead and vote now. All right, in AI news, IBM has launched Granite 3.0, expanding its generative AI business to $2 billion. So IBM is making headlines with the launch of its third generation of Granite large language models marking a significant extension in its enterprise AI business, which is now valued at over $2 billion. So the new Granite 3.0 models include general purpose options with 2 billion and 8 billion parameters. So some smaller models as well as specialized mixture of expert models. If you've ever seen that moe designed for a range of enterprise use cases such as customer service, IT automation and cybersecurity. So the models are now available on IBM's Watson X service and platforms like Amazon Bedrock, Amazon SageMaker and Hugging Face, providing widespread access to these new AI models. A key feature of Granite 3.0 is its release under the Apache 2.0 open source license. Yes, these are open source models, so this allows enterprises to build on this new technology and have a more collaborative ecosystem for AI solutions. IBM claims that the Granite models outperform competitors like Google and Anthropic in various tasks, highlighting their state of the art performance and commitment to safety with advanced Guardian models to prevent harmful content all right, our next piece of AI news, some ChatGPT updates. So advanced voice mode has officially rolled out to all those other countries that didn't yet have access. So that includes users in the eu, Iceland, Norway and Switzerland. And this also does provide limited access for for free users as well. Speaking of new access to new features inside of Chat GBT Search GPT is starting to roll out to more users. Didn't even see an announcement from OpenAI on this one or anywhere on the web actually late last night. You know I check all of our, you know we have paid accounts and free accounts to every single platform and I noticed that actually our free account here at Everyday AI had access to search GPT but not our paid account. So if you do have a couple of accounts, go and make go and check. See you might have access to the new search GPT which is open AI's kind of competitor against perplexity or AI overviews from Google. So you can just use the backslash search command. All right, let's get into it y'. All. I'm excited for everyone tuning in. Uh, so Brian and Jackie and Marie and Christopher and everyone else, thank you for tuning in. I'd love to hear thoughts from you all and our live stream audience or sorry, our podcast audience. So as always, make sure to check your show notes but reach out. I'm I really want to know what you guys think about AI agents, but let's just get straight into it and talk about what's new here and what was announced. So if you missed this yesterday, well, and unless you were in London, maybe you didn't see this, right? So this happened very early in the US time, but Microsoft had its AI tour event in London. So this news happened very early here on Monday in the US And Microsoft CEO Sadia Nadella announced Copilot Studios new feature, which are autonomous agents. All right, so seems like maybe a small thing because we've heard of these co pilot agents before. So maybe you saw this and you're like, oh, okay, this has already been announced. So now they're just giving us a date, Right. Because they said that this is going to roll out to users starting in November. So maybe you heard this announcement and you saw it and you're like, okay, well, we've been here before. I knew that there were, you know, agents coming to Copilot Studio. But this is new small detail which is not small at all. Autonomous agents. So Microsoft had previously talked about essentially programmable agents inside of Microsoft Copilot Studio, but not autonomous agents. Right? Yeah, you still program these autonomous agents, but these are ones that are essentially AI employees. Right. If you saw any of the headlines in the past 20, you know, 20 or so hours. Hours, everyone's talking about AI employees. AI employees, right? So that's what's kind of happening here with Microsoft. And I think you also have to look at the bigger picture, which is kind of the race, the race to the first fully autonomous AI agent available to users everywhere. So now all of a sudden we're looking at Microsoft, whereas before we were looking at Salesforce. So Salesforce at its Dreamforce conference just over a month ago, announced its big offering in autonomous agents called Agent Force. All right, so Microsoft here swooping in and it looks like they may actually beat Salesforce to the punch as they announced that this would be available starting in November. Right. November is in like eight days. All right, So I don't know if this is going to be, you know, the first week of November or the very last week. I'm sure we'll, we'll see more reporting in the coming days and weeks on when Microsoft may actually announce this. However, it does look, look like Microsoft could be the first household name company that brings autonomous AI agents to millions or hundreds of millions of people. So what are autonomous agents and why do they matter? Well, in short, autonomous AI agents are agents that are powered by AI. They're powered by large language models. And for the most part, and you know, it, it works a little differently depending on what service provider that we're talking about. But you give these large Language model agents access to your data, then using normally either low code or no code. What that means is you just talk to these agents in normal language like I'm talking to you right now. And that's actually how you build them. Right? And you can kind of set up some guardrails, some fallbacks and dependencies on if something goes wrong. And then these autonomous agents will complete tasks without human intervention with your data. And in some instances, they'll be sending emails for you completing tasks. And y', all, it's, it's actually wild me saying this out loud because if, if you would have asked me when I started everyday AI, you know, almost two years ago, hey, would we have autonomous AI agents in 20, 20, 24? I would say, well, maybe, but I wouldn't think that they would be this mainstream. Right? I, I would, I was probably thinking they would come from, you know, a chat GPT or an anthropic Claude. But not big name companies, right, that have way more, I guess, influence on, Well, I guess OpenAI has a ton of influence and tons of users right now. But I would have expected it would have been essentially one of these AI companies, not traditional, you know, kind of tech companies, Microsoft and Salesforce. So pretty exciting news if you're a fan of the technology. But there's also downsides and cons, which we're going to get to those as well. All right, so like I said, let's go over the basics here. Let's talk about a little bit what they are, what was announced at the AI Tour event in London. So like I said, this was announced less than 24 hours ago by Microsoft CEO Satya Nadella during Microsoft's AI tour in London. So these run or can run once they are available, presumably in, you know, a couple of days to a couple of weeks, they can run 24, 7 and they are started by a trigger. Okay, that part's important and that's one of the differentiators between kind of traditional AI agents and you know, AI agents that you could even build in Copilot Studio and autonomous agents, they're connected to your real time data. That's huge. Right, because what's the most important thing or one of the most important things for all employees, right? Oh, you got to make sure you're looking at the most updated sop. You got to make sure that you're bringing in all the orders from today and not yesterday. Right. Whatever, whatever it may be. But you know, business happens in real time. So, you know, even when we talk about, you know, Kind of non agents, Right, but we talk about things like GPTs or projects from Claude or Gems from Google. Right. These kind of more consumer versions. I, I won't say they're agents necessarily, but kind of. Right. They can perform a very, you know, a very narrow set of tasks based on data, but it's static data, it's data that you upload. So with autonomous agents, they're, they're working 24, seven, they can be triggered automatically, right? So not a human going in and pressing the go button. And then they have access to your real time data. And in most cases, these autonomous AI agents by design require little to zero human input or interaction after you set them up. All right, so yeah, you got to set them up and you can go cut these agents off at any time, right. And then a lot of times, like I said, if the agent runs into a problem, there will be kind of a fallback or you know, to the human. All right, so let's, let's go over in a little bit more detail. So autonomous AI agents right now, they can handle tasks from simple interactions to fully autonomous actions like sending emails or managing onboarding. So right now it is an integration in Copilot Studio where users can create customer agents in Copilot Studio, like I said, launching in November 2024 for specific business needs. And also Microsoft is launching 10 different pre built agents. So these pre built agents are for routine tasks like supply chain management or expense tracking. So now let's hit rewind one month and you might be thinking, Jordan, I've heard about these, I've heard about these agents, these Copilot agents. Actually you talked about them, right? Yeah, I did. So in September, Microsoft had some pretty noteworthy updates to its Copilot, you know, set of software. Right. Because Copilot is now sprinkled everywhere in Microsoft's operating system. So we had these, what were called Wave two announcements from Microsoft in mid September, so about five weeks ago and I talked about it on this very show, I went over it. So you might be confused. Okay, was Sadi and just giving us a new timeline? No, these are completely different. These autonomous agents in Co pilot Studio are different than the co pilot agents that were announced at Wave two. So I'll oversimplify it here. Right, but essentially what was announced with the Copilot agents, right? And I don't know, Microsoft, maybe, maybe we should give them different names, right? It's like, oh, the new co pilot versus the old co pilot, right? Because now we have this copilot V2 for you know, users, you know, on the front end accessing Copilot in the browser, right. You have this new UI UX. So this is different. So the wave two copilot agents announcements, think of them more like GPTs, right? Customizable versions of GPTs. So this is what copilot for our live stream audience. I, I have a visual on the screen now. So the wave two announcements essentially were this. It was kind of like GPTs. Right. So on the left you can kind of chat with the co pilot agent to help you build an agent. Yeah, I know that sounds weird, but that's how it works, Right. And then on the right it kind of builds it for you. And then you can test it in the big, kind of the big announcement with these Wave two agents. And what separated these Wave two co pilot agents from the GPT builder that had already existed is you could share these co pilot agents right in the now what's called biz chat. I believe they'll be integrating into pages, into Copilot pages. Right. But that is not what was announced yesterday. I know it's confusing, right? So what was announced yesterday, completely different. So let's just go over, we're just going to say this. The September co pilot agents versus the October. Right. So the September 2024 copilot agents in wave two were designed to assist with tasks but required more human inputs. They helped automate specific workflows but were not fully independent. They were not autonomous. Right. Like I said, more like GPTs that you could, you know, share and use throughout the Microsoft ecosystem. And then the October 2024 copilot, those are the ones that just came out. These are the autonomous agents available in Copilot Studio that are autonomous, meaning they can act with minimal human intervention or none, and handling tasks like client communication and decision making on their own. All right. I don't know if anyone else is confused. Hopefully breaking it down by Wave two September copilot agents. This is what was available. And now this October announcement, which will be rolling out in November. But if I had to put it like a three to five word, right, wave two was advanced GPTs. You can share with your team what was announced in October yesterday. AI employees. Big difference. Big difference. And I don't think, even though this is all over the news, right? If you follow AI, you've probably heard about this. I still don't think this is getting really enough play. I don't, I don't like. I think this is a huge deal and I don't know, maybe because it was part of this AI tour, right? There wasn't a lot of visibility. You know, where Microsoft normally, if they have a big conference, everyone knows about it. I don't know. I'd say most people, unless you read everyday AI every single day, you probably didn't know. Microsoft had their AI tour eventually in London yesterday, but I don't think it actually got as much play as it should. So right now, agents can range from simple prompt in response, right. To fully autonomous. And they can execute tasks like we talked about, sending emails, employee onboarding and lead generation. Here's the important thing right now, because again, one of the downsides, and we'll talk about those a little bit more. One of the downsides about autonomous AI agents, number one, it's your data, right? Because you always, you always need to know. Okay, well, I don't want an autonomous AI agent on my behalf emailing customers or members of our internal team with old documents, right? So right now, the AI agents, the autonomous ones can draw context from Microsoft 365 graph systems and then also, sorry, Dataverse and Fabric. All right, so essentially some different data channels or groups of programs inside of Microsoft. Right? So Microsoft 365 Graph Systems of reports, the data verse inside Microsoft and Microsoft Fabric. Okay, Speaking of Microsoft, got to take a quick, quick coffee break here. All right. Gotta shout out our sponsors from Microsoft. The Work Lab podcast from Microsoft is made for leaders who want to understand how work is changing. Effective leaders adapt. They stay on top of trends. They embrace any edge that they can get. They learn from the ways technology is transforming other fields and how it's enabling organizations to work more efficient, efficiently and productively. So for real world lessons and actionable insights to help you stay ahead, check out the Work Lab podcast. That's Work Lab. No spaces available wherever you get your podcasts. Yeah. The new season from WorkLab has been fantastic. Can't wait for the new, the new drop this week. All right, so let's get back into autonomous AI agents and talk about why now or why is it different? So I think this is one of those instances. It's almost like the boy who cried wolf with AI agents. Because I'll say, going back to 2022, right, in early 2023, we were hearing about AI agents, and I think a lot of people just started to kind of turn their attention when people talked about agents. Agents, agents, Right. A year ago, 18 months ago. You know, you have to, you know, tip your hat to some of the early, the early players in this space, like LangChain Right? And every single big company, every single big company, open AI meta, Google, Amazon, right? Microsoft, obviously, Salesforce. Every single big company over the last six to nine months has prioritized how big of a deal it is for them. And one of the their main goals for many of these companies that I just named is having autonomous AI agents, right? I'd say, you know, if we were having this conversation four years ago, the rush was better large language models. So now we're in this phase of, okay, well, now what? What does it mean if we just have, you know, incremental gains on these large language models, right? The GPT models, the Claude's models, Geminis, right? It doesn't mean a whole lot if you just keep getting these models better and better because consumers expect more, right? Whether we would have this same conversation three years ago, the expectation now from the average business consumer is, okay, great, you have these smart, you know, these smart large language models. Now I want to see business value. I want to be able to have these models do some of my work, right? And not me, have to coach them through. That's where we are now. But why now? You know, I found this interesting. Sadia Nandela in his keynote kind of mentioned these three different things, and I want to break them down simply for you about why will it work now, right? Why couldn't it work a year ago, two years ago, five years ago, AI agents. So a couple of things that Satya and Adela talked about was one, having a universal interface, right? And his new catchphrase is, you know, having copilot be the UI for AI, right? I don't even hate that. When I first heard that at the Wave two announcement, I kind of chuckled. Then I'm like, wait, that's actually not bad, right? The UI for AI, the user interface for artificial intelligence. But he talked about a more universal interface, right? So what that means is now AI and large language models are great at not just understanding natural language, but also multimodal inputs and multimodal outputs. And that changes what's possible, right? You no longer have to be a prompt engineer whiz to get the most out of large language models. You can upload a screenshot of something, you can upload a PDF, you can upload an Excel sheet, right? And it'll say you can tell a model, hey, what is this? And can you create a visualization out of this? Can you tell me what this means? Can you convert this into a different format? Right? Large language models are now better with a universal interface and understanding of how we work, right, no longer has to be these long drawn out text prompts. Number two is reasoning and planning. So interestingly enough, and I don't think these details have been released with these new autonomous AI agents. Well, what's powering them? Right, presumably it's something from Open AI. We don't know yet, at least as of, you know, late last night right before I, I went to bed, couldn't find the specs. I'm sure we'll see them in the coming days and weeks. But is this the O1 model powering this? Right, so the OpenAI, what was first called Q then was kind of, you know, internally nicknamed Strawberry, but we saw this new reasoning model from OpenAI that has really just blown the, blown the top off of all these benchmarks, right? But a model that can think like a human, can process things step by step like a human, right? It has this kind of built in chain of thought. So interestingly enough, Saudi Nandela referenced the O1 model when talking about agents, but didn't say that's what's powering them. Right, because we know Microsoft also has been investing a lot of time and money into essentially, I won't say backup plans, right, because we talked about some, some reported, you know, renegotiations that are happening right now between Microsoft and OpenAI. So we know that Microsoft also has invested heavily on its own internal models, right? They've been creating some impressive smaller language models, the kind of inflection AI aqua higher. Right, but number two is reasoning and planning. Okay, because what, what do you need, right, if you say, oh, I need an AI employee, you need that AI employee in theory to be able to reason kind of like a human and you need them to be able to plan out tasks. And with the new O1 model and you know, a lot of these other big companies are working on models that can reason that's available at least as of now and then last but not least, memory in context. Right? Because now today's state of the art frontier models, they have bigger and better memories, they can better understand you and your business and retain that information and can access your up to date information. So those three things that kind of make this, this why now moment, it's the universal interface model's ability to reason and plan and then better and improved memory and context. So let's take a closer look. Okay, I just want to look at this one example that Microsoft talked about yesterday and they talked about a few examples of, you know, some of their clients that had already had access to these new autonomous AI agents. All right, so one, one quick one here. Is this McKinsey, right? Maybe you've heard of this small little company called McKinsey. All right, so they went over this McKinsey use case, essentially when they would get emails, right? And some big company, they come in and they say, hey, I want to work with McKinsey. I want to hire you. Here's, here's some information about our company. All right, so for our live stream audience, let me see if I can make this, make this a little bigger. There we go. That didn't really help. All right, so here's an example, right, of an email that the McKinsey team might get, right? So this is an example. They shared this. You can go watch the whole video demonstration. We'll be linking it in our newsletter, but I'm going to break it down so you don't have watch it if you don't want to. But essentially, you know, this is an email from a company saying, hey, you know, hey, here's my name, here's my role, here's what we're trying to do. We really want to engage with McKinsey on this project. Here's some details, right? Standard email. And you probably have a, you know, a form on your website or a dedicated email that handles these inquiries, right? We have one. I mean, we're a small business. We have something like this. I'd say most businesses have something like this. So I like this use case. I like this example. Yeah, I think what Chris said, maybe I need a work lab mug or maybe an everyday AI mug. All right, so pretty simple example, right? You get an email, there's a lot of information, and then one human will generally read through it. These get piled up in their inbox. They have to think, okay, who should I forward this to, what should I respond? What information in a long email do I have to pull out, maybe copy and paste something, right? So this example that Microsoft went over is setting up an agent, right? So as an example, this autonomous AI agent built in Copilot Studio has a set of actions that it goes through a set of steps, right? So we talked about, hey, what's one of those, one of those three things that why AI agents can happen right now? Well, it's AI's now ability to reason and plan. Okay? So as an example, some of the steps in this AI agents workflow is to check the engagement info, right? So this example from McKinsey, all right, to check if they've had previous engagements with this clients, right? Because if so, you'd probably forward it to the, the lead of that team. It needs to check the industry type, right? What industry is this company? Identify relevant expertise, etc, right? So there's many different steps that this agent goes through. And here's the thing, y' all in co Pilot Studio, I'm excited, I'm excited to get my hands on this. I probably have to see. Maybe I'll have to pull a favor, right, See if I can get some, some earlier access so we can talk about it with you, all right? It's like, man, Microsoft is one of my partners. I gotta be able to get some of this stuff early so I can talk about it with you guys, right? But essentially the way you build this, the way you build this agent, right, even three, five years ago, you would think, okay, well, this requires a developer. You need a, a team of software engineers, right? You need to know how to code. I don't know, you need Python or, or Ruby or next JS or JavaScript, whatever, right? You would think, okay, for an AI agent to go through and do all of these things, that's a lot. But here's the example Microsoft gave of how it was actually built this agent that McKinsey used to, that goes through and it handles all of their kind of engagement leads and inquiries built with natural language, right? I'm not going to read this whole thing, but as an example, right, we kind of talked about the, the first step, first couple steps here. Checking the engagement info. Checking previous engagement. So literally text like this. All right? Analyze the incoming email you received and extract the following information. A client's name, the client engagement scope, industry start date company name number 2. Check engagement info Use check engagement info action to verify that all necessary engagement information is provided in the request. B. Oh, sorry, that was A. So that was A. And then B. If all necessary. If all the necessary is not provided in the request, right? They didn't even use the word information. If all the necessary is not provided in the request, send an email, right? I don't know if, if this was done on purpose, but there's like grammatical errors even in this kind of natural language. And maybe that's what Microsoft is trying to showcase here, right? Like, hey, it says it should say send in email. It says send an email to the client to request all the information and stop further execution. Simple natural language, right? And then you get that agent that goes through and completes all of these tasks for you. So like I said, we'll, we'll leave the, the link to the video if you want to watch that in today's newsletter. But here's McKinsey's reported results. Because you're probably thinking, okay, well, not, not that big of a deal. This is a low level task, right? Well, not necessarily, because they said that it handled more than 3, 1300 inquiries. It cut the lead time by 90% and that's huge, right? If you're thinking sales, you just always be selling, right? Always be closing, always be selling. It's so important to get back to people as quickly as possible. So cutting lead time by 90% by using these autonomous AI agents that McKinsey bought or, or built inside of Copilot Studio. Huge 90, cutting the lead down by 90% and then also reducing the admin overhead by 30%. All right, so again, to recap, to recap that, because I think this, this use case is a very, very simple one, okay? And also what makes this autonomous? Well, it's a trigger, okay? It's a trigger. So maybe your company has built something like this with a little bit of duct tape, right? Because technically this type of AI kind of automation or autonomous workflow has already been available or it's, it's already been possible, right? So you could use something like Zapier. Like Zapier with some AI and with some different connections, right? But essentially you had to kind of duct tape this and build it a little manually and use some third party tools. But now this is all under Microsoft's umbrella, right? You're not having to use any duct tape. You're not having to do any, you know, third party API calls, You're not having to rely on other services. It's all autonomous, right? So the big thing here that happens is that trigger, and that is one of the things that separates these new autonomous agents from, you know, different agents that were already available in Copilot Studio is you can set the trigger. All right? It runs 24 7. And if there is a problem, and Microsoft gave an example, right? The example they gave, like, oh, the person that this was supposed to forward to, that person no longer worked at the company. So then essentially a user would get a message inside of Biz Chat, right? Which used to be called Copilot, but now it's called Copilot Biz Chat, right? You can think of it as essentially, if you haven't used it, kind of like, kind of like teams, right? But you get a little message down there and it says, hey, there's an error in one of your flows. And then you go in there and you can just Type in natural language and say, oh yeah, this was in the flow. It said to send this to, you know, if anyone was in the accounting industry. That goes to Bill. Bill is no longer with the company. That actually goes to Jane. So let's send that to Jane and then you could go in there and update your workflow. Right. But that's the power of an autonomous AI agent. That is, number one, it works 24, seven, it's triggered depending on an action that you set. And number two, when you can still keep human in the loop. Humans in the loop. At the right, at the right time, at the right point, at the right moment for the right purpose and work with your dynamic data. That's huge. Right? I think we're going to be seeing a lot more of these case studies like this one that we went through with McKinsey. All right, so that's a wrap, y'. All. I could talk about this for another hour. I think it is extremely exciting. But I wanted to take a show today just to go over this in more depth in part because like we talked about, we just were hearing about co pilot agents last month. So even myself, I had to do a lot of digging and I'm like, okay, what's different with these autonomous AI agents that were just announced at the AI tour in London and what makes them different? And the ability now inside of co pilot Studio, I don't think we can overlook what this means for business. Obviously a competitive move here. You know, Microsoft is competing with a lot of different people. You know, I don't know if Salesforce's kind of announcement at Dreamforce and essentially saying, hey, we're an AI, we're an AI company now, we're not a sales company. And coming out with Agent Force, which does a lot of these same things, right? It looks like Agent Force in the demo that they did, same thing. You can build an agent in natural language, set a trigger, you know, it can connect to real time data. So I don't know how much the kind of Agent Force announcements or movement influenced Microsoft, if at all, to really ramp up its autonomous AI agents. Maybe it was because of the O1 model, right? I'm assuming the O1 model is powering this. We don't know yet. As soon as we find out, we'll report that to you. I'm sure that there will be some more specs either today or later this week. But I think this is important. This is one of those big stepping stones in our journeys of AI first companies and being a generative AI leader in your workplace right now. I think you have to understand this is a big step, right? Maybe you're Microsoft company right now and you run everything on, you know, windows. You use Copilot365. Maybe you don't, right? Maybe you're a Google company. You know, I'm assuming that Google will be responding in turn. Google's already announced at their IO conference, you know, they're working on AI agents. You can actually build some agents right now. So I'm assuming that Google, with whether it's a couple weeks, couple months, couple quarters, will probably have some big AI agent announcement. But here's the thing. This is no longer theoretical. This is no longer one of those things that. Okay, well, in a couple of years, this is Microsoft said November willing, it pushed back. I don't know, it looks like it's ready in production. Right? Microsoft shared use cases of companies that have actually already been using this technology. So I don't think you can sleep on autonomous AI agents. They're not some figment of our imagination. They're not some part of some wild AI fantasy. They are available now. Well, in a couple of weeks anyways. All right. I hope this show was helpful. If so, please let me know. Please share this with your network. I know this might be your little secret. This Everyday AI, right? Like, oh, this is how I'm the smartest person on AI in my company. Share this with someone. And if this was helpful, please go to your everydayai.com Sign up for our free daily newsletter because we'll be recapping all of this, right? Maybe you were out walking your dog or on a treadmill or on a treadmill with your dog and you missed a couple of points. Don't worry, we'll be recapping this and everything else that you need to stay ahead in the world of AI in our newsletter, go to your everydayai.com so thank you for tuning in. Please join us tomorrow and every day for more Everyday AI. Thanks y'. All.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everydayai.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.