/ Insights / View Recording: Automating Personal and Team Tasks with Copilot and Copilot Studio Insights View Recording: Automating Personal and Team Tasks with Copilot and Copilot Studio July 11, 2024Curious what tasks people are ACTUALLY automating with Copilot and Copilot Studio? Join us to explore the real-world applications of these powerful tools.What You’ll Learn:Practical Applications: Discover specific tasks automated with Copilot and Copilot Studio.Efficiency Boost: Learn how automation saves time and enhances productivity.Step-by-Step Guidance: Get a walkthrough on setting up and utilizing these tools.Unlock the potential of automation with Copilot and Copilot Studio. Transcription Collapsed Transcription Expanded Nathan Lasnoski 0:18 OK. Hello everyone. We are proud to be here to talk about automating personal and team tasks with Co pilot and copilot studio. I am so excited about this topic because it is so much about something very passionate to me, which is enabling every person to be able to be more with AI. It is about every person in your organization. This is not just about an IT person or a data scientist or an AI engineer. It’s what every individual and their ability to leverage AI to be able to force multiply with what they do, and we’re going to talk about how you think about that, how you frame it in the greater context of adopting AI in your organization. And we’re going to do a bunch of demos that show some of this in action. So let’s introduce ourselves a little bit. My name is Nathan Wisnoski. I’m concurrencies chief technology officer. Do with concurrency for 22 years. Been working with data science for the in AI for the last nine and to speak frequently on this type of topic. I would love for you to connect with me on LinkedIn. That is my QR code. You can just scan that and you can find my LinkedIn profile. I release a weekly newsletter on AI, and this week’s is actually about how to be the best version of yourself with AI. So I think that really well aligns with the things we’re talking about today. I would love for you to read it. I love for you to engage and talk and give me your perspective on on what you think. So nice to meet you today. Glad to have you here. Super interested in in having conversations and in the jail. I would love for you to introduce yourself too. Would have would have a slide for you, but I love for you to introduce yourself because you’re the star of the show. Ajay Ravi 1:50 Thanks nude. So my name is Asher Levi and I’m working with concurrency for the past five years, mainly on the local node side of things and the business apps. So compiler studio is something which Microsoft all have invested heavily over the past year or so, so really looking forward to this our session and showing some what are the some of the capabilities. Nathan Lasnoski 2:07 Uh. Awesome. Thanks. It is. Jay spins this stuff up so quickly to to create value for companies like we’ll have a meeting with a company in like the next minute. He’s got something really exciting. Ready to go to show them. So I’m thrilled to be able to have him present some of the things that he’s done today and also for us to talk through the picture. So let’s get started. So today we’re going to be covering 3 things. We’re gonna cover framing, AI adoption, and if you’ve been to some of my presentations before, you may see some things that would be familiar. And they’re familiar because we want everyone to understand that part of the framing. So we’re gonna talk about framing AI adoption. The context of Co pilot and copilot studio and also where you need to start thinking about other technologies that would complement them as well. So we’ll hit both of those today and give you a lot of content along the way. So as you are thinking about of taking advantage of AI within your organization, we always want you to be aware that there’s two big domains that you need to be taking advantage of, and these are not mutually exclusive. So the first domain is adopting commodity capabilities and this is you’re leveraging things like M365, copilot or copilot.microsoft.com or something that gets released in Salesforce or some tool that’s put into your ear piece system. This is about adoption and that adoption is a new skill that every one of your people are going to have to interface and the ability to delegate to an AI agent, the ability to perform tasks in partnership with AI is not a skill that most of us have or you have experience in our lifetimes. So it’s an opportunity for us all to learn something new. So commodity adoption is a big part of what we’ll be talking about today, but then directly adjacent to that is this idea of force multiplying the very mission of your business with AI. And we always seek to find that opportunity. Seek to find that opportunity for the strategy of the business, the existing strategy of the business to be accelerated as a result of AI and to deliver new value to customers. So we’ll talk a little bit about that today, but we’re gonna be a little bit more on the commodity and adoption side. Then we will be on the mission driven side. They’re both kind of fit in the picture. So for those of you that are technical, I think this is gonna be really useful. Slide for you for those that aren’t, you might get a little deep, but just kind of take it as it goes. So the two big things we talked about on the left commodity and mission driven are the two big boxes, OK, these ideas of like I’m adopting something or I’m building something probably another way to say it build versus buy, right. But there’s different sort of levels of that, right? So as you go through the just general AI capabilities, think of like this top area as the I’ve got a person out there using chat GPT. Well, the problem with using chat GPT of course is in other kind of AI tools that are built on making money off of selling your information is it’s not private, so we need to use private tools for our businesses and that’s for things like Microsoft Copilot that every office, every Office 365 instance has available to it. Take advantage of that. If you do nothing else, just give your employees a healthy option. That is comparatively and safe and has commercial data protection on it and is not used to train other models. All that kind of good things that you need to be secure in your organization. Just start there. But then as you start wanting to leverage AI to be able to advance your organization more, and especially ground that information that’s specific to your organization, you start using tools that have that grounding built-in, like M365 copilot. So you might say, hey, I want to write a faster email or I’m drafting a presentation or I’m creating a marketing asset, maybe I’m summarizing a meeting, these kinds of things are a great fit for M365 copilot, and I’ll show you some examples of what I personally use later on in this presentation. But then you get to a point where you’re like, well, I I I want to get very specific answers to very specific questions for internal tasks or automation that a person’s performing. And that’s where some of the things will demo today around Copal Studio really come into play. The accuracy and precision of those kinds of use cases is dramatically higher than what you would achieve by just turning on MG65 copilot and asking a question about your health insurance policy like you’re going to have a very specific document that sits behind that question. So but studio fits in the space of automating team and task activities that we can take advantage of. So we’re gonna show a lot of different interesting things about today and then the box, we’re not going to spend as much time on in this presentation, but as like super awesome. Important is as you get to this point where like this is now become so critical to our business that it needs to work all the time. It needs to have extremely high levels of accuracy. It cannot answer it. Customers question wrong. We’re using it for demand planning. We’re doing a next best action model. We are building a chat bot that our our caregivers are using. This probably a custom L scenario you might be using some of the tools from Copilot Studio as parts of that picture, but there’s probably a higher degree of custom that’s fits in that picture, so it’s part of the continuum. I think the reason I really wanna show this is that your strategy is gonna have all this stuff as part of it. It’s gonna have commodity tools. It’s gonna have fully custom. It’s gonna have everything in between as you advance your company down the road, so that makes up a multi agent architecture. So if you think about like things that are happening from copilot, that should intentionally be part of a picture that relates into tools that everyone’s building and compiled studio or power automate. But then also might be fitting into a fully custom solution you built that’s running your business and all of that should be thought of as a canvas. That all relates to each other as part of the same picture. Avoid the idea that like I’ve got this point solution over here. Got this point solution over here. Never will they talk like that’s not the future of AI agents. The future of AI agents is almost in a sense of like I don’t have employees that just kind of sit on their own islands. They work with each other and they have to collaborate and communicate to be successful. So do AI agents as we grow the business? So, umm, where we’ll be getting today is like how is the way I interact with these AI agents from a personal and team task automation changing. So if we reverse time about a year and a half, OK, prior to the chat GPT moment when we Googled for something like we basically like Googled in a way that gave us a list of sites that we would go to to find an answer. And sometimes we’d ask a question and would give us the answer that question. But usually, like that’s not basically our pattern like we would go ask for something and then if the chat GPT moment we got comfortable with the idea that, like, hey wait, I can interact with a sort of a I chat agent and it feels and acts sort of like a human. There’s all these like gaps around it, but like the interaction is much more consistent than what I would have experienced before. So we got comfortable with this idea of like back and forth experience with the chat bot. But then we like said, well, wait, I wanna do that. Not on the information from that was frozen in 2021. I wanna do it on the information that’s centric to me my information and that’s where this idea of retrieval augmented generation came in. It’s basically the model that exists from 2021, plus your data, and your data is really what’s answering those questions. So we then got comfortable and you may be in the process of starting to implement some of these patterns. Now we’ll certainly show some of them today the idea of a rag pattern. Like I’m asking a question or I’m wanting information and it’s retrieving something from a document or a business system that’s grounded in that data and retrieve. But then what? We’re going to talk about today is how can I start to ask you to do things for me, like, not just return information, but are there opportunities for it to actually perform a task like order something or perform this action or preparing materials for my briefing or like go in perform this onboarding activity that I would previously had to do. Like when can we get to a point where I truly am delegating activities almost like an intern that’s growing up? That’s where this idea of copilots really comes into play. And this is like an overused term at the moment and it’s got a lot. It’s a very loaded, but it’s this idea that like agents work side by side with us and then eventually we get to this idea where they’re fully autonomous, which we’re really not at at this point. I’m not having chat GPT run my airplane right? We have some even a fully autonomous cars were tons of investments are going in, still have flaws. So there’s a continuum we’re working on and we’re still kind of moving in that direction. So prior to like showing you examples I I I always start with like the examples and I moved to governance. I’m like governance always gets the second half of the story. So I wanna start a little bit with governance, because for me it’s the really about like the key part about this. Like, if you’re convinced that using things like copilot or copilot studio, or building apps and enabling that is the right thing to do, the most important thing from my perspective is to govern it well, to enable and governance to to sort of balance those two things. So when you’re building these lanes of adoption, adopting commodity capabilities, building mission driven factories, it all has to fit in this human first AI picture. There’s a gentleman that I’ve worked with, Todd Mclees. He’s an expert at human skills development and adoption, and I’ve really appreciated his framework because it’s been very much about what are the new skills we have to learn. But growth mindset adopting in delegating to an AI agent, skills development, creativity building operating at top 1% of our own human capability, there’s a lot of transition here. It’s not just like what we even had with a smartphone, so this idea of a center of excellence or a pattern that exists to move down the human first AI path, it’s really important. But that’s also tied to, like starting with. Why so? The best companies? They create this understanding of what am I applying AI to that impacts the strategy of my business and they map it and they maintain it and they understand it. So this is an example of 1 like for a variety of companies and you can see different types of scenarios in this this slide, right? So reviewing contracts, so like this idea here of I’ve got contracts coming into my organization. I’m reviewing the PSA and MSA. I look for these 15 things like what are the net terms, what does the, what’s the, what happens if there’s a mistake like one of my retribution options. Like there’s these things that I wanna look for in those Msas and PSAS, well, I can have an AI agent actually looked for those things for me. And then they can report back and give me that information. So I can start to get to this point where like I really understand where is the value and how do I map that and how do I then go achieve it? And that’s where I think even scale even at commodity, this is where there’s a lot of opportunity. So we’ve seen people turn this into idea, registry apps and power apps or other tools that enable them to be able to maintain this over time and not to have this just be like a one time spreadsheet but be a vehicle for capturing, understanding or ROI within the business and being able to then track that forward into achieving value across the business. And that even gets to the point of like, hey, that’s built in power apps. I can start to chat that I can ask questions about the ideas I can interact, even creating that for many businesses had become an opportunity to be able to leverage the tools and to learn. So it’s been very fascinating from that standpoint. And then getting to a point where you can even do true ROI analysis and this is an example of using an ROI analysis to be able to determine how autoquoting will improve the productivity of business. Actually one of the more common use cases has been all my salespeople produce quotes every day. How can I optimize their process of producing quotes like because I know the more accurate and faster I get my quotes back out to my customers, the better success we’re gonna have. So knowing that that’s one that we see a lot, we’ve started to develop our ROI calculating tools and enable us to Fast forward through the process of understanding how we can gain ground for a company and things like sales, quoting and other use cases because it’s not interestingly enough, if you do this enough, you find everybody kind of has a lot of the same problems like sales, quoting, demand planning, preparing documents and assets, HR like customer service. These things are very repeatable. A lot of it happens in many companies, so that’s something I think has been a really nice asset. So let’s start talking about like, OK, we’ve got governance like we know we need to see OE, we know we need an idea register. You know, we need to maintain that over time. We need to have like processes or moving from DEV to QA to prod. Like all those things should exist. But let’s talk about what are actual ways people are using some of these tools. So we’ll start with commodity AI, specially with M365 copilot and I’ve used this slide before, but I think it’s a good starting point, which is every commodity AI is just boosting us. So I was at A at a company yesterday and they started adopting him, not empty, 65 copilot, but GitHub copilot and and he said it is clear as day like the productivity difference between individual who has the ability to use GitHub copilot and not it’s easily saving them a day of Ohh software development work just in that week because they didn’t have to go out to Google and find it. They didn’t have. They just started typing what they wanted. They created a virtual plan and was grounded and then materials and built the assets for them that they could then do the remaining 10% for. And I think that’s really where the I think the GitHub copilot use case is like a precursor to what we’re experiencing with M365 copilot. They get them. Copilot is ahead of us because people have been using it for a while. And you’re already seeing the results objectively in the productivity of development teams that are using GitHub copilot to achieve ends in their organization. M365 copilot is the same thing like there’s always kind of healthy skepticism of like, is this really going to benefit me? And then you find the thing you have the name saying of like, well, it doesn’t really do this well and it doesn’t work. When it went to find this document, that’s all true. I have experienced all those same words that most other people have, but I also wouldn’t give it back. I find that like out of one out of four times, it can’t find the document I was looking for that I needed to go find the other three times it did, and those three times meant I didn’t have to spend time finding it. So, like trade of work, summarizing means even like this is one that for example I’ll show you an example of this. Just making sure we turn that stupid thing on is like one of the most critical things, because like you, you’re so used to not having an AI agent take notes for you and then you’re in a meeting where it is taking notes and just summarizes and you’re like, shoot, why didn’t I just do that for this meeting? So that’s part of the adoption process as well. It’s just like getting past some of those roadblocks of making it productive, so creating content for me, this is about like statement of work generation. So I have statement work templates. We have to create them. If I can give it the asset, give it input assets that like are complementing it like a PowerPoint, we use to present the initial draft that can create a good start on my statement. Work for me before I even have to put energy into it. And if you’ve written statements of work before, they’re not creative work like it’s a contract. So if I can have something take something that load off, the better RFP same kind of idea. I’ve answered the same RFP 72 times. Can I have something create the response that RFP for me rather than me having to do it myself? These are all great opportunities for using things like Copilot organizing with customers. I hate flying time like I don’t know if you guys ever use fine time, but like, were you like, have a poll? Like there’s a meeting poll people have to, like, pick times that there is say it’s like a bridge to nowhere. Like nobody’s, it takes so long to get people to actually respond to it. And by the time they do like half the means have the times have passed and stuff, it’s not great. I like to give people like, hey, here’s my availability specific times that I have like action oriented, right? But I always have to peruse my calendar to find those things, so I finally figured out a prompt and it took a while. Like for me to figure out the prompt, I had to kind of commit myself to it. I like. Here’s the prompt that will like affectively generate my openings in a preference that I like and like if I had an assistant which I don’t like. If I had an assistant, it’d be like they already kind of know my preferences. And then, like, here’s sometimes you should schedule and. These are things you can book over and all that kind of stuff. I’ve finally gotten to a point where I can actually delegate that to an AI agent to give me the times and then send them back without having to double check it. So like, that’s an example of like where it saves me time on on regular basis. Meeting summaries just have been absolutely huge, like the idea, especially when you turn it on but has been a real value associated with using copilot. This is an example of it from like a legal contract review. OK, so like starting with summarizing the email conversation like ohh my gosh, I’m the legal guy. I just got pulled into a conversation that has 17 different, you know, back and forth with the customer. And now I’m asked to do something like what do I do with this? That’s where the summarizing email conversation gets really handy. I just came back from vacation and I’ve got all this stuff. Summarize it, then drafting the content. Maybe that’s useful. Maybe it’s not, but this is the this is the one that I think really is valuable, like combining copilot with copilot studio to say. All right. Here. Here’s the document. Look for these 15 things. Give me the summary and then I’ll take a look at it. Point me in directions to go. The difference between doing that with copilot and not doing with Copilot is stark. So great example of like where that can accelerate it. Moving that then into comparisons where changes have happened, catching up on the meeting and you can see how as I gain these skills, I can start to get acceleration on each of these subcomponents of the process. But the thing about this that I wanna really stress to you is people just generally don’t do this on their own. Like you turn on the license, they’re not going to know how to do it. They probably won’t force themselves to do it. You need to train them on the chat. The task to be done like the job to be done, the activity that they’re gonna perform, it can’t just be train them generically on using copilot. It’s not gonna work like you have to. You can train them generic kind of Copilot. Just need to give them. Like I’m in a domain that has some really bad official opportunities. Train them on that. Figure out how to make it work so legal is a great example of this. So building on that, let’s talk about copilot studio, because this is really where like you think about like Microsoft’s messaging from build. What they’re really trying to do, they’re taking these tools and they’re they’re bringing them to the masses. So some of these are like, not masses, things, but some of these are. You’ll see examples of things that are So what is copilot studio? It is an automation business process platform. It’s goal is to automate automate for every person in the world and do so in a way that cars with your unique skills. So there is a wow like holy cow, this is a complicated architecture document diet diagram. It’s you again. May not be a tech person, and that’s OK, just stick with me for a second. So copilot studio architecture it’s about, I’ve got a lot of ways that people can consume the things I build in copilot studio. Think a chat bot in teams. Think a delegation that happens from an app think. A meeting that I’m in or a trigger that happens in app that I’m that I’m monitoring or a SAS application like. Like the HR platform, tons of channels can be used to be able to interact with a intelligent agent and then with that facilitates is the ability for me to use conversational language and escalate to topics that then follow into different actions downstream. So that means that like I can have the Copilot Studio platform sort me into different lanes. So like I may have a bot that can do 72 different things and it sorts me in the lane or lane two or lane three lane 4 based upon the conversation I’m having with them that then sends me into work that can be isolated into different flows and that’s where we’re going to talk about here. But what’s important, too, is it’s all built on this. This idea of this data verse and the data verse is basically this like SAS data platform that exists across copilot, copilot, studio, and other things. That sort of Dynamics platform and so on. But then also it’s ability to interact with fully custom use cases and things like Azure open AI. So if you’re building this and you say, oh cool, this is really working well for me. And then you’re like, shoot, it’s not able to get that quite tuning to what I want to do. This really needs a blunt punch out into something much more sophisticated, and maybe you have a full data science person who’s working on it. You might punch it out straight into an Azure open AI use case or a full data science use case. So I think it’s what’s exciting about this is it’s a framework that is going to have a ton of investment that you can build on and leverage the work that people are putting energy into and still build custom things in and around it as you need. So some examples of Copilot solutions. This is a just let’s start with the most basic scenario ever. I’ve spoke at Summerfest Tech a couple weeks ago, and this example of like all right, cool and building my first copilot studio solution. And I’m gonna plug in a public website like, let’s just start with that. And I mean I could use files or data versus or SQL tables or whatever I want to use, but let’s just start with a public website and start by asking questions like who are the keynote speakers? And am I speaking at it and it’s framework is already built to, like, respond in an intelligent way. It’s responding with good information and B it’s giving me links to where that data came from at that source location, whether it’s just file or a SharePoint site or a public website. And for many people, that’s like all they want to do. I just I have a document. I want to surface it. I want to be able to allow people to answer questions against that document like boom, like you can do this really quickly with copilot studio solutions and build on it to do really sophisticated things as a like extension of what you’re doing. So at the most basic level, that’s like a starting point of doing something interesting. Then we get to something. We’ll talk more about later, which is doing like real task automation, which is like employee onboarding. So I’m like gonna drain this totally Ajay is gonna demo a solution. That kind of starts to talk through this, but you can see here we’re not chatting with the employee onboarding platform to like get this to go. It’s detected. We know someone existed. We’re gonna do things as a result of that in existence. I’m gonna follow up with them with SMS text. I’m gonna respond to certain things. I’m going to give send them documents. I’m gonna like all these things that we like are annoyed by having to do. We can start to delegate to AI agents to perform for us in a really intelligent way. So employee onboarding isn’t good example of that because it just represents a a process that there’s a lot of good things that we need to do and we want to make sure they’re done well in single time and we want the humans to focus on where the real energy should go. As part of that process, so this is an example that I’m gonna skip this one because I want Ajay to kind of just demo it for the most part. The other example is doing things with service now, so we actually happen to be a service now partner as well, and one of the things that we’ve been seeing value from is taking service now and leveraging them now assist platform and Oracle Plot studio and putting that as an interface to interact with the Copilot experience. So our to interact with the IT group rather than having every single call, I have to go to it. What if I could like put chat bot that answers a lot of those questions or lets them understand the status of the incident, or open a new incident or solve their own problem, do so in a way that really gets them moving down quickly and that’s what you can do with this. So there’s an auto box plugin for service now. I can plug that right into a chat experience, or I can use that now assist platform to be able to or both to be able to make for a great experience for an individual who’s answering questions or dealing with problems that they need to get answered by it. So it can be a nice little starting point. Also, returning data from the knowledge database you can see here that I’m I’m asking for some of those. Some of those questions to be answered now. In this example you can see that we’re actually creating something. OK so maybe I need to create a purchase order. So in this case what I’m doing is we’re asking a question. How can I help you today? And we’re indicating I want to create a purchase order, but then based upon that it says I can help you with your email address. Can you confirm the device and it’s prompting me with the things that I want along the way and then actually performing the task for me? And that’s really again you can see an employee onboarding, you can see in this example we’re using the dynamic capabilities that platform to actually cause it to do something for me that I would have had to do much more manually in myself or someone would have to do like in the tool to do like what if I could just say hey or chat with this experience and it will do all the work for you, just answer its questions. Umm and I hire an intern in. They can do that work because whatever right then just opportunities for me to be able to really think about how I can delegate to that then AI agent. This is also true for like documents that are uploaded in your SharePoint environment. So if you have like of course data that exists in your internal environment, you can surface that for Q&A experiences as well as even updates to those documents as necessary in a way that’s really appropriate and why I think this is so important is because this is like one of those really key like HR onboarding kind of use case right or sorry, HR Q&A use cases that you probably wouldn’t build custom like you’re it’s not it’s important but it’s also not like important enough that like you’re going to build something completely custom for it you’re going. Have surface a series of documents and assets that are very specific to that process. Put a chat bot in front of it. Get in front of all your internal information workers and your non information workers, manufacturing workers or whatever, and it already provides value and that good is good enough. Is a perfect fit for it. Then you might do one that has to have more tuning and maybe still even then you can get the right tuning from the copilot studio platform, or you start to inject other custom capabilities into it. So alright, so at this point I love to hand over to Ajay and Ajay is gonna walk through like an experience of doing some onboarding with power platform, power automate copilot studio just to kind of show what the capabilities look like interface and different parts of this. So go ahead, Ajay, take it away. Ajay Ravi 29:49 Thank you, Nate, for providing all those details. So let me just quickly share my screen. OK, I hope you can see my screen. Nathan Lasnoski 30:03 Yes, Sir. Ajay Ravi 30:04 OK. So we’re going to quickly take a look at some of the things which you can do using power platform like Nate mentioned or the copilot studio power automate and some the some of the power apps. So we have spent uh it quick HR chat board. So using this and HR person can kind of like do a bunch of different actions like you know all on board and start creating records and in our system and then also you know you can look for the record status and then you know send off a letter at personal task and all those things so you don’t have to like go to a bunch of different places to perform a lot of these things. You can just use a copilot studio in order to kind of like to a lot of these things. So this is our ECHO Pilot Studio view. How which? You can see over here. And then what? I’m going to do is I’m just going to like, I’ll start a normal conversation. So I’m just going to type hi and then what I want to do is to start onboarding and your user. So I’m just going to like start start typing on boot there. Yes, Sir. So it pretty much asked for the employee name which you would like to onboard. So I’m just going to type my name for now. So what it doing in the back end is is going and creating a new record in an HR system in this example or a power app or like in Dynamics 365. So right now you can see that our new user has been created and then it is. It also automatically triggered some automation in the back end for manager approvals and also provided me with the reference ID which is associated to the new record that was created. So if I go to the system and then just refers to now you can see that in your record got created just now and then you can see the employee ID associated to that record as well. So from just a chat experience, you can directly create records and or different HR systems allowing you to integrate with different business systems. And in addition to that, if I quickly show you how some of the back end stuff works, and so you can see that an automation got triggered all setting it to the manager or for approval and starting of the onboarding process as well. So this way you can create records in an HR system and then start various approval of processes. The next thing which I’m going to show is also if I want to get a application status for any of these existing record from the HR system. So I can do that as well. So what I’m going to do is I’m just going to look for the application status for this particular or record which just image ones. So I’m going to what I’m going to do is I’m going to type. Yes, I’m just going to type as application status, so it will pretty much ask for the employee ID. So I’m just going to type. The employee ID. So what else is doing? Is it is checking or the HR system and then it pretty much picked up the ID information and then provided what is the current status. So this way or you know from the copilot studio you can interact with an HR system and get all various data points and then the status is and all various things. Now the next thing which we can do like for HR person if they want to add personal tasks to their you know priority list right? So I’m just going to type. I want to create task. So I’m just going to type create, offer letter are Ajay. So it’ll asking how pretty much what is the priority. So I I want to like keep it as high priority. So what it is doing is it is going ahead and then creating a new record in A to do and then it will be assigned to me. So you can see that if I go here you can see that in new task got created. So this will easily allow users to add items to their high priority. I’d like different lists so that way you know they don’t have to manually go to various system to do that, but instead my chat or Copilot or experience you can directly create all these reminders and all those things. So what I’ve done is I’ve said now any high priority items as like you know that due date set as like 5 days off from today. So you know that’s been automatically set so you don’t have to manually go to different system and create all these things, right? And then another thing which I want you to do is so if I want to. Create another task so sent training. So now all I want to create another reminder which is like a medium priority. So I’m just going to, like, ask the desk medium and then if I come to all my to do so now you can see that our under task got created and it has a medium priority and then the due date is set to like 10 days from current date. So this way or you can use copilot studio to kind of like automate and then create various tasks for personal items as well. So this way you can do a lot of things and then in addition to that another thing which I would like to do as an HR person is maybe I’ll send an offer letter to the user, right? So I’m just going to type. Create offer letter. So what it is? Strict wilting is a couple of additional information, so I’m just going to type my name and then. The designation. So basically what it is doing in the back end is it is going and creating an open letter and then sending it to the user. So if I quickly show that experience. So now you can see that and offer letter got generated and then it emailed to my, you know, personal email and then you can see the different or dynamic values which I entered. So this way you can kind of like automate the offer letter creation process and another thing which I would like to do is maybe send some training materials to the users so. So I can just provide the name. I’m just going to uh, provide any email address as well. So now what it did was it was it kind of like automated and sent it personalized email or with some of the training materials, right. So you can see that an email got triggered and then just sent to my personal account with all those training material based. The user needs to or kind of like review before onboarding or like you know before starting the first day right. So this way you can kind of like automate a lot of different processes or using just a single copilot studio and then does interface and even just kind of like look at how the automation looks like. So this is a simple example of how the flow looks like. Uh for sending the training materials, you can see that we have pretty much triggering than automation from the flow and then we have customized the email and then it will accordingly look for some dynamic values and then send the appropriate or email to the new user. So this way you can kind of like automate and improve your onboarding experience using simple local no code solution like our copilot you. Nathan Lasnoski 38:14 Awesome. Ajay Ravi 38:15 But bad with my Shentel. Nathan Lasnoski 38:15 Thank you, Sir. Ajay Ravi 38:16 Yep, I will stop sharing. Nathan Lasnoski 38:17 I love it. Umm. I’ll bring mine back up here. There you go. You can see my screen OK. All right. Ajay Ravi 38:32 Yep. Nathan Lasnoski 38:33 Thank you. OK, so I a common question that comes up is how in the heck is this licensed? Like what weird thing? It’s complex like it is. It can and will be complex sometimes. Umm. Here’s the kind of like net net of it, every person in your organization that has empty 65 Copilot already has access to copilot studio as a. As a developer like, they can build them and use them like unlimited for themselves. OK. But then as you build copilots, those copilots can either be used by a people have Copilot, M365, copilot, or they can be built by it for you. Anybody the anybody can use them and they’ve made this really attractive. So it’s essentially for copilot $200.00 per tenant per month for 25,000 messages, and that sounds like a lot. It can be and it also can’t be. So if you consider like a message being that a token, umm, a message essentially is like 2 tokens per so like 2 messages per interaction. So that means like 12,000 interchanges on, like asking a question and getting response. So I think that’s for like many of the copilots people are building like that’s a pretty low point of entry. Umm. If you have one that’s like doing tons of interaction, of course that that’s getting increase, but I think it’s a pretty attractive costing model. There are other models that they have that are tied to like other bulk licensing and so on, but I think that’s a good way to just think about getting started. Like, OK, I’m not having to invest in some like across the board license for my entire organization just to start building with copilot Studio. Like I can build them for $200.00 a month and you can do a free trial like just get started. Like if you have an Office 365 tenant and you’ve been turned on to be able to like use the capability, it’s not like blocked for you. You can start turning it on tomorrow and like playing around with it and getting value. Of course, there’s like a lot of things you want to do well, and that’s where we come in. But that’s the licensing I think is pretty attractive for this. So lots of interesting things coming. I’m not gonna talk about all these, but I think it’s important to know that like Microsoft is putting in an enormous amount of energy behind copilot studio like it is a tool that they see as the future of work. Copilot, copilot, studio, copilot, and 365 like A You’re seeing them. Use the term a lot, but like the ability to, not just. Like return information, but actually do things is really at the root of what they’re trying to accomplish here. So tons of opportunities. I’m gonna hit a couple of the ones that I’m really excited about here in a second. So things that I think are really valuable one is the automation center. Essentially, the ability to see all the things happening across the automations that are built around either the robotic process automation or the logic apps that are and not logic apps are the power automate that are being executed in the context of the broader ecosystem. And I think that’s important because just having the ability to govern, this is an important angle for the IT organization. The thing that I am really starting to get excited about is these next two. So AI flows and AI ARPA, so AI flows, essentially the ability to describe what you want and have it create the flow for you and your mileage will vary. I’ve already been playing with it. The ability to say like I want an AI flow that will do this, this and this with this system like sometimes works really well, sometimes doesn’t. It’s usually based on whether there is a out of box connector for that job and the data the data exists already for it, but I think the future of that is exciting and why I’m so excited about is because really where this is going is describe what you want and it gets created by the platform. Now it’s not totally there yet, and I think we got a ways to go, but you can see where they’re moving toward that. The English language essentially becomes a programming language. The idea of describing my ask and this is the same in the power apps platform as well. The way we started creating that idea Registry was create me an idea registry that has these fields and these controls and these things and it built that interface and capability for us like we actually have to start from that. It actually build it. So I think that’s really gonna lower the bar again for more people to be able to have access to this kind of innovation. The second thing, and this this is so early, so like this isn’t something I would expect tomorrow, but it’s something that just kind of reminds me of why this is so important is there’s, if you’re familiar with the term RPA robotic process automation, there’s a point in time where this was all vogue, right? Where, like you’re essentially automating common clicking processes that people performed, and sometimes it worked really well. But a lot of times it was just. It was too fragile. It was so based on where the button was and how one had to be clicked and the pixels on the screen and stuff, it just didn’t work very well. So what? They demoed it. Build was his idea of essentially teaching an AI to perform a task in a similar way to which you teach a person sitting next to you. Like you, you like do the task and you describe it verbally as you’re doing it and it’s like capturing all these notes and it then can adjust the platform as the interface changes. And it can do things that a little bit more like generic in nature then like it’s similar to a way to which like we notice like the move from chat bots to chat GPT like the stark difference between what we experienced before and now. I think this is gonna happen with the RPA experience as well and like how much I can truly say. All right, cool. I’ve onboarded a new employee. I’m going to train that employee to perform this task for me. That’s really exciting to me. The potential future of that now again, this is not really something that you have today, but it’s just giving you an insight as to what the potential future is gonna look like as we keep marching towards this destination, we’re at the infancy of its capabilities and the ability for us to, like, get in and Start learning the skills. This is what the other kind of inside I had the other day was. We are going to be so much better off if we have started relatively early knowing how to interact with AI platforms as new capabilities drop. Then, if like 2 years from now, all the stuff exists and we’re expected to like drop into it and start being productive like we’re gonna need some help. We have skills to develop, so part of this is our skills development along with the platform. As we sort of learn along with the innovation that’s happening, so it’s there’s our onus on us to be part of that process. So OK, so where do we go from here? Hey, I was told we weren’t gonna get through this content like we’re already at 11:45. We’re pretty much done. This is great. We’re going time for questions. So where are we go from here? There’s three things that I suggest every organization do. So the first thing is you need to start with. Why? Like everybody always needs to start with why this idea of just start playing with tools like great. But the problem I’ve seen when people just start playing with tools is they tinker around with stuff but never makes it to production. You don’t want to be that organization, you want your solutions to make it to production. So that means we need to be very intentional, intentional about our training, intentional about our onboarding, intentional about helping people to be productive, tooling, uh, seeking to solve the right problems or secure the right opportunities and strategizing on that is really where we need to start. So that’s understanding the best opportunities for value, getting executive alignment mapped to outcomes. That’s the thing that we do a lot with organizations to help them to perform that task. Like, where do we go from here? How do we gain ground? How do I get Executive MGMT? How do I executives know that this is gonna help us be powerful and change the way their business operates? That is a very important task for us to have, like combine the top down with the organic bottom up. So you really get actual value and then that moves into like evalue in a scenario and sometimes people hit us right there like they’ve already got the scenario figured out and we’re now have the onus of digging deeper like let’s make sure this is the right thing to solve. Let’s understand what you have now, what the potential future could be and understand the ROI behind it. And you saw some of the assets that we bring to the table as part of that. And then the third thing is like, let’s build stuff, let’s build and scale our ability to build S like our build with motion, right, the essentially the idea that like I will help you build things and you will learn to build things and we will build things together and we’ll scale that across the organization and responsible way that is really the net of call to action is like actually get outcomes from this. And part of that’s the enabling the experimentation. Yes, part of it’s solving the right problems. So our goal is to help you do those things and we would love to participate in any one of those stages, but especially we love to start with strategizing with you to help you make the right choices. So understanding that that we’ve got all that indicated, definitely fill out our form as you’re leaving today, but now I love to take some questions. We’ve we’ve like. Amazingly, have like 12 minutes left, and of course everybody needs a break and stuff, but we would love to answer some questions. So if you got more, start dropping them in the chat and Jay and I will start to answer. Well, that’s a great question. Thank you, Mickey. So the question is, would there be tiers or levels for managers, leaders, administration, general users? Or is it for all regardless of focus levels? My my initial response to that is absolutely it’s for everyone. It’s a question of access, what they should be doing with it, what they shouldn’t be doing with it and the kinds of jobs to be done like, I don’t know if there’s an idea called jobs theory. Essentially, what is the job they’ve done? I have in the organization and the main thing is that each of those persons jobs are different from each other, right? They all require different sets of skills and they would have different activities. They would probably delegate to an AI agent, so like I met with an executive team and the the CFO was like I wanna ask the AI agent to give me my weekly summary of these things. And I was very different than like the line level person that wanted to be able to interact on these particular other data points. So I think it’s not so much like art. Is it relevant to anyone, these people? And it’s more about like how do I make sure I’m solving the right problem for the right person and that’s why I meant back to that persona mapping idea that like when you roll things like this out, you need to be very specific about what their job to be done is and how I make sure that person is getting value versus like just generically hoping values gonna exist. One other thing I would note is that there’s a really interesting crowdsourcing tool called prompt buddy that is used to essentially crowd source Copilot M365 copilot prompts. So like at the more massive level, like just everyone, they can share things that have worked for them. That’s been one of the biggest challenges in general. It’s been like what are you using it for and what does this other person using it for and how do we like all benefit from that? And that’s uh, that’s a tool that, like, helps with that engine is being able to navigate that. So hopefully Mickey the answered question. If not, I can follow in that. Could you clarify the $200.00 a month for using customized Copilot, specifically deploying for users in your organization who don’t have their own copilot licenses? Yes. So the idea behind that is that you could deploy a Copilot for $200.00 a month that no one else in the organization has any other license for it. So think about it even like, let’s pretend they have 0 licenses. It’s a it’s a external chat bot. You’re a customers are interacting with it or employees that are on your manufacturing floor that don’t eventually have digital identities and you’ve put it in front of them as an experience that facilitates that essentially pays for that based on messages. It’s consumption based billing rather than a like a license based billing. Next question, is there a large amount of effort needed to put in to ensure that you have clean data before rolling on M365? Copilot concerned about giving users access to Copilot and then generating content on stale data and SPO, such as old policy documents, bingo. Huge. Huge point. Umm, so we didn’t really cover much of that today, but there is a very prescriptive adoption process for him to 65 copilot and one of the first things you’re doing is preparing the estate. And you know you’re not going to undo 10 years worth of old SharePoint data in like 3 weeks like we just have to, like, reconcile it like it exists. So there’s things like oversharing tools and scoping back what the semantic index has access to. See, everybody remembers when they turned on delve and all of a sudden people could see what their executives are working on. Like there’s a certain amount of initial work that you do up front, especially with your pre pilot group to be able to scope that back. So we do have like a prescriptive process of what that looks like. I don’t have it in this deck, but that isn’t really important thing. Umm, in terms of like the policy documents and so on, that’s exactly where things like copilot studio fit in really well. So like rather than like empty 65 code just is not gonna be good at that job. Like there’s too much diversity of content to answer very specific questions, right? All the time. So that’s where I’ve got a policy document or I’ve got an HR document. I’ve got this thing. I’m gonna put that specific content behind a Copilot that I build like in in copilot studio for example, or have more custom one where it is specifically looking at that content and nothing else, and then you’re pointing your team to maintain that content or going forward. So rather than it being like, here’s the 10 years of SharePoint content, do your best answering that question. It’s Nope, this is the content that we’ve rigorously performed for that chat bot to answer these questions. The last thing on that topic is umm, as you are preparing the content for like broad Andrew 65 rollout, that’s for things like just Brett or information governance in general come into play using policies and tagging associated with that content is really important. It reinforces why that was a good idea in the 1st place. If you were doing it and why you need to do it if you didn’t in the past, OK, next question, how would an organization approach integrating Copilot solutions into a large number of applications that were built before the chat? GPT moment. Ooh, that’s a really interesting question. Umm, I think it goes back to that idea about the like seamless garment. Like everything you do should be built on architecture of how everything else should be talking to each other. Like, it’s not just a one off. Don’t avoid the. Maybe the one off kind of get you going down the road, but like the one off should be the foundation for building a a ecosystem of related things that all work together well instead of like a whole bunch of completely unique chat bots that don’t talk and you might have a bunch of like legacy applications. And what we’ve been finding is like a lot of those legacy applications can be made as data sources like whether we’re synchronizing data to some database or providing or it’s are like available. It doesn’t, and it can be a source for the experience that enables that activity, but not necessarily like building it right into the platform itself. So I think that’s something needs to be impacted a little bit more, but there is just tremendous opportunity to be able to like innovate on older platforms even like I mean frankly like a lot of the platforms are building under like, yes, we’re hundreds and stuff. So like the data source itself is like these old as dirt boat anchors that existed and we’re building on this next question if anyone who has access to M365 copilot license automatically has access to Copilot Studio, what is the Copilot point of the copilot studio license? That is a great question. I think it would be if someone did not have access to M365, copilot would be the main point of that, like like and that’s really the case. Like a lot of organizations haven’t rolled them, just like Copilot across the entire state. Umm so that that would be my only take on that. Ajay Ravi 55:11 Uh, I can add to that name. So so you need to organization need to purchase the tenant license which costs like the $200.00. Nathan Lasnoski 55:12 Please. Thanks. Ajay Ravi 55:19 So that way you know and users who don’t have access or like you know have license they can use utilize the Copilot studios. So the $200.00 per month for 25,000 messages is the tenant license that you need to purchase for the organization in order to use the copilot studio. Nathan Lasnoski 55:37 Thank you. Perfect is the prompt buddy you mentioned built in a Copilot or separate app. Ajay Ravi 55:38 Umm. Nathan Lasnoski 55:43 It is a separate app you wanna drop that in there as I can do it later. Just Google prompt buddy, it’s on GitHub. Umm does an end user employee need to Copilot license to use a chat bot as an end user that has been built by using copilot studio? No, it does not. It just consumes it when you pay that $200.00 user license and then it’s it. I mean, that’s the starting point for the 25,000 messages, right? It can be bigger than that, but there’s no other license. You. You’re not going to pay extra like you could roll it out to someone with zero licenses as long as you’re paying that license. How does the licensing work if we are connecting to applications like Salesforce data query data, there’s two things about that that you need to be thinking about. The first thing is there’s of course power automate costs and maybe as Jay can speak to that a little better than I can. But the second thing is there’s also like data verse storage costs. So if you’re storing data in the data verse that is of substantial nature, you might have to pay for the storage of that data and the data verse. If it goes over certain levels you know about like the costing works for power automate interacting with things like Salesforce, Ajay. Ajay Ravi 56:54 So there are like bunch of different uh licensing model for power automate as well. So some other standard connectors, if you’re using within copilot studio, so errors of free calls. You just need to like purchase the Copilot Studio, but if you’re using some of the premium connectors like maybe I’ll Salesforce or some of the data worth applications so you can always purchase the by user license for the power automate that allows you to integrate and interact with a different system. So it costs like around $20 per user per month. Also, there are like few different licensing models there as well. So if it is a premium connector or like you know premium of functionalities, so you can always purchase power automate for utilizing and leveraging some of those things. Nathan Lasnoski 57:40 OK, here is the prompt. Man, that’s a big, big link. There’s prompt buddy if you want to get access to that. Can you provide a prompts that you use for things like your calendar availability? Sure. I’d be happy to. That’s a great. I should just do that in like the newsletter or something O so yes, I will do that. That’s a. That’s a great ask and I would love to do that. So thank you for asking. OK wow, this has been awesome. Wow, what a fun session. Thank you for spending all your time with us. We are really looking forward to following up and talking about how we can make this real in your business and thanks for all the questions that makes these so much more interesting. When people ask lots of questions, so thank you for that. I love the engagement and really excited to talk more about this, so I have a wonderful day and thank you for joining us today.