#107 TAG talk about Use Cases for Microsoft Copilot

Hello and welcome back to another episode of the Process and Automation Podcast with the Automation Guys. Yeah, as always with me here today is Arno. Hello, Arno.

Hi, Sascha. How are you doing?

Good, good. So yeah, back with another recording on our own. So the last couple of months has been We have been busy with interview episodes, but yeah, today we are, it's just us, just us, but equally good, I would say. Of course. Yeah. And here on the podcast and offline, we have the joy to talk with many experts in the space and also have plenty discussions with, yeah, with, with our members, listeners. And we find out that things stand out across all of them and obviously lots of things stand out for them to be successful with AI and automation in general. So it's the for today, as we, it's just us talking about it. So obviously we can pick another very dynamic topic. So we would like to talk about Microsoft Copilot and yeah, Microsoft Copilot, what it is and how Copilot works and what you can use Copilot for. So yeah, you probably heard about Copilot in the last couple of months. So obviously when, when ChatGPT came up, everyone was talking about generative AI, gen AI, and Microsoft of course, working very closely with OpenAI and doing their thing there, but they have their own sort of version of ChatGPT and generative AI. That's, that's Microsoft Copilot. Yeah, so it obviously is a significant, significant step for Microsoft in the world of AI-powered productivity tools, especially, we know Microsoft Excel, Word, PowerPoint, all, and email Outlook tools. These are, these are the productivity tools for everyone out there. I would say 90% of the world is using those. So yeah, Microsoft Copilot integrates machine learning capabilities into all those applications, and that will enable everyone to effectively work smarter. So it's one thing we say here a lot, work smarter, not harder. So that's why we're talking about this today. So it could be drafting documents, could be looking at data, could be creating a presentation, and yeah, of course, managing your emails, or when you are a developer or sort of getting into code, so it can help you writing code. And yeah, it is definitely the assistant out there everyone should look at, especially we working with those tools every day. Yeah, adding Copilot to it absolutely makes sense. And yeah, Arno, yeah, so that's the topic we talk today.

Yeah, thanks, Sascha. Obviously we all have heard The buzz around Copilot, not a day goes past where you don't hear that term Copilot being used, you know, in marketing emails you get to join a new webinar or event that is showcasing the capabilities of Copilot. But the reality is that it has actually been around for a while, the underlying technology. Of course, Microsoft has sort of put that all in one place called the Copilot, but under the hood, it, it still uses artificial intelligence, AI, which includes sort of machine learning technologies. And basically what it is, it, it functions as an intelligent assistant with inside, uh, your Microsoft 365 applications. So you've mentioned Word, Excel, Outlook, PowerPoint, So sort of the day-to-day applications that you'll go to if, if, if you do your work, but it goes slightly beyond that in that actually Copilot also looks at your wider Microsoft 365 estate, including all of your data, all of your documents, all of your emails, all of your users, line of business data. So these are data that's stored in Dynamics, for example, or in your generally in your Dataverse database. It uses natural language processing, or NLP, to actually understand and interpret and generate human language in such a way that's meaningful and useful for us to use. And this first fundamental sort of, I guess, concept of NLP actually allows us to sort of interact with Copilot using a sort of natural and conversational language. Of course, behind the scenes, we've got machine learning, and these are algorithms that help Copilot learn from the data that's inside your Microsoft 365 estate. And it also is a continuous learning process, so it improves its performance over time. Obviously it looks at all of the data. So, you know, we've got a lot of interactions, email being sent, and that can be a vast amount of data over a period of time. And what it does, it looks at that, it looks at sort of contextual data as well, exactly where you are inside the Microsoft 365 environment. So if you're in Word, it knows you're in Word, so you're probably going to be able to, probably going to write a document. If you're in Excel, you're probably gonna do some, some analysis, data analysis. So sort of learn about that, that sort of contextual environment. And then what it does, it help you make more sort of relevant decisions that's based on your data, the context. And then large language models is, is, is sort of sophisticated language models. And, you know, OpenAI's ChatGPT, which stands for Generative Pre-trained Transformer, is used and it's trained on these diverse datasets. And what happens there is it can generate sort of coherent text, understand context, and also looking at, you can almost call it kind of creativity in generating responses. So what happens there is that it actually generates content on data, but it sort of adds a, I guess, a creative flavor. So it's doing suggestions, it can summarize stuff. So I think that, you know, we obviously got the sort of the machine learning, which is the algorithms. We've got the large language model that actually can sort of do the generative text generation. And then the other sort of technology is that Copilot sort of built on top of the Microsoft Graph API. And the Graph API is an application programming interface that connects all of the various data sources in Microsoft, such as your emails, your documents, calendars, SharePoint, all of that together. And so Graph is really the— or the Graph API is really sort of the gateway to the data. So all of these three things then work together And it provides you then with a sort of a comprehensive AI-powered digital assistant that's context-aware and can also make decisions or make suggestions on decisions or help you make decisions quicker or, you know, sort of create some creative content for you, for example, you know what I mean? So it is something that's very useful. We all use ChatGPT at the minute, but Copilot just looks at your work. 365.

Yeah, I think that's, that's a big, big differentiator. So last, last year was all about experimenting with ChatGPT and with similar things, Gemini, Claude, and all these. And yeah, that was all about where we upload a document maybe, and, and it was sort of in the public space, which was not good either. I, I don't want to know how much public or secret information actually sits in the public models and databases of those, but I think the real, the real benefit really is using your own data. And this is, this is not, not just the benefit is nearly, I think nearly this is, this is the, this is the big tick box for corporates to actually start using it and adopting it. If, if it's sort of somewhere in the public area, it will never, never, never work. No one will adopt it. Healthcare will never adopt it. Finance will never adopt it. All this confidential information. So yeah, it really makes sense. It is now in a secured. In your Microsoft container, basically. Yeah. And then really gets exciting, I think. And I think then you can actually start using it properly. And for that, I thought it would be nice to go through some specific use cases, some practical examples how Copilot can assist you in your day-to-day job. So yeah, we mentioned earlier Microsoft Word, of course, is a big tool we use every day. And In companies, in the different roles, we have to draft reports. So, and this is how Copilot can assist you by doing it. So you're just outlining maybe a few key points for your report, and then Copilot can then look at more sort of detailed content based on these points. And this is where it does make sense, of course, looking at your data. So it only makes sense, of course, asking Copilot by just giving a little outline if it works with your data and understands what you want, and otherwise it would come up with some fantasy stuff, isn't it? So, but by, by knowing your data, it can create these points based on similar documents and content you have created in the past, and this is super powerful. So yeah, so, and, and Copilot can also, when you start writing, identifying sort of, okay, errors of course, and can make any sort of suggestions in terms of In our company, we want to write in this kind of format. We want to write in this kind of terminology. So it can really make sure that you're writing better quality reports and documents overall, and it can format the documents. You don't have to worry about getting the right sort of formatting in. Sometimes it could be super tedious and we know we screwing up a lot with that, trying to catch up with all the formatting. Yeah. but that can all happen automatically as well. So, so this is just an example how you can use it in your day to day. Yeah. So it can create as well come with an overview. It can come up with some extra findings. So yeah, very powerful tool you can use every day. And it works very similar, I guess, Arno, with other things. Yeah.

I mean, I mean, that's a very simple example where you want to write a report within SiteWord, but I've also mentioned Excel earlier, so Excel can actually integrates with Copilot, and Copilot can help you analyze large datasets by suggesting different formulas, for example, or creating pivot tables and generating visualizations. It can also help you generate test data, for example. So if you have got something you want to model and you want to create test data, it can do that. So in the past, that can be quite tedious. If you want to have test data, you want to visualize and perhaps push for a model, that's a very useful use case for it. It also helps you to simplify data manipulations as well. So it's a very useful companion to have within inside the DANA data analysis environment, especially using Excel. Like I said earlier, it's private within inside your Microsoft tenant. So it only looks at your data, your context, doesn't go outside and doesn't publish any, any, any data outside of, you know, your, your specific tenant. We're looking at kind of creating powerful presentations. Milekan Adrian and creating really compelling presentations. It could suggest layouts, for example, generate slides for you, design elements, and things that's got a visual appeal. Now, that might not sound like a very difficult task, but if you're a person who have to create sales presentations, perhaps just change the context slightly because a different customer and different industry, it could take the best practice from your presentation library. And we all know how hard it is to find a presentation. We think, oh, I really have a nice one. I can't remember which customer it was for. It's almost similar. So Copilot can help you just to navigate that, get to that information quickly. And that's, that's where the productivity comes into play. If we look at things like Outlook, for example, Copilot can help you manage drafting email responses, prioritizing messages, for example, even scheduling meetings. Of course, it's got access to the Graph API. So it's got access to, to people's calendars. Calendars, so on and so forth. It could also sort of, if you write a really long email and you want to summarize it, you know, it can sort of create a little summary for you, including sort of quick key points, perhaps next steps, action points, these things that you would usually have to do yourself. I've seen it used in Teams meetings, for example, where transcript is there. I turned on Copilot. Then listens the whole meeting, creates action points. It actually creates tasks as well for you to action or for the task members of the call. So very, very, in terms of productivity tools, very useful in that sort of space. And when you want to create something and you've got a wealth of information that you sit on, but you don't necessarily know where everything is and you don't have the time to go and unpick it and read all the documentation again in order to get the best bits. Copilot's just there because it's got that information. And as people get more used to working with Copilot, they'll get more used to using the right prompts, for example, to query Copilot and talk to it to tell it this is the sort of thing that I'm after. Make recommendations. Of course, it's not going to create the end-to-end proposal for you, but what it's gonna do, it's, it's actually gonna suggest the best way of writing it. You can then take that information, tweak it, and I've seen in the past writing proposals myself, it does speed up absolutely tremendously. So you can do a lot more in the same amount of time, focus more on the content and create, you know, true human creative thinking rather than putting hands on keyboard to put text on the screen or on a document, which is just sort of normal reference type text you'll do anyway.

Yeah, very good. Yeah, you just mentioned going through like lots of documents there earlier. So, um, of course we can't really go through like hundreds of pages every time again. So of course we can, but it's so time-consuming. Just imagine we talked about Word earlier. You have like sort of a shareholders agreement or some other terms of business, that kind of stuff.

Yeah.

So you can ask ChatGPT, not ChatGPT, sorry, Copilot, of course, to come up with, yeah. Yeah. What are sort of the key points here to maybe be worried about or what is the highest risk here? So it obviously can help you to identify and point out a few things. Of course, it's not replacing immediately sort of your legal advice, but it can help you already. Just to understand the whole thing a bit better and much faster. So, yeah, but after all, so of course we want sort of these productivity gains. I think that's huge by using these technologies. So Copilot and others. But yeah, you can then sort of focus more on strategic and creative aspects is what we always say when it comes to sort of automation and AI. So we take away all this mundane, repetitive, time-consuming activities so you concentrate more on the value add. Yeah, a huge driver is always, is always improved accuracy. So Copilot is there on the side and can understand context and provide all these sort of relevant suggestions. So, and that, that will ultimately reduce the errors because it's nearly there. Like, yeah, like your, your helper. And if you make a mistake or if you, if, if, if you enter something wrong, it might highlight that for you. Yeah. And overall it's huge time saving and Copilot will become more and more powerful over the next, I think probably on a weekly basis nearly with all the advances Microsoft puts into the platform. So yeah, it can handle more routine tasks. It can, yeah, take away more of those tasks away from you. And then further down the line, even more end-to-end automation, I guess. But for now it's just helping you there in those kind of tools mainly. And yeah, huge benefits.

Yeah. And, you know, you look at the, it's very user-friendly to use in terms of the experience using this capability or this tool because it's got this natural language processing capabilities. It's very easy to interact with it. So you can speak to it in over voice, even if you want to, you don't have to type a sentence or if you type a sentence in your own style of English, for example, or whatever language you interact with, it understands your intent. And it just means that the barrier of entry using this technology is really low because it's so easy to pick up and adopt it. So you could have people from sort of various technical expert levels or levels of expertise pick this up. So you could have a developer, for example, that asks it to write a piece of JavaScript to enable a ribbon button in Dynamics, you know, and it will tell you how to do that. Or you might have a marketing assistant, you know, wanting a piece of content based on your business performance or, you know, your annual reports last year. And it might be quite hard to get that sort of information. Copilot can dip in, get the highlights or the noteworthy information, present that to her, for example. Cool. So it's, you know, it's just the ease of, of sort of getting it in front of people and using it. Of course, I mentioned earlier, it continuously learns. So the AI models is always learning, and it always improves itself over a period of time. You know, this assistant of yours becomes sort of more used to understanding what your preferences are, and it will then in turn deliver more accurate suggestions. And, you know, it seamlessly integrates with your other Microsoft applications. So there's no additional software you need to install. Once you're in Microsoft 365, you've got Copilot. Of course, you need a Copilot license. There's different licensing levels. So that's one thing that just to be aware of. But to get started, there are sort of basic access that you get that you can use. You get Copilot Studio as well. Copilot Studio actually helps you to create Copilot Copilot's targeted for specific things. So it is, it's very interesting where we are at the minute with, with, with, with the technology. I think, like I said earlier in the program, we are hearing the term Copilot a lot. I think Microsoft understands the strategic importance of it. The technology is there, and the fact that it looks at your data and understands your data is extremely powerful thing. And I think the adoption of this will just explode in the next sort of couple of months or next year or so. So I guess if you're out there and you're even a C-level or you're a departmental head or person working in a specific department, check it out, come to us. We'll give you a Copilot demo, sort of just to demystify it more. And it's an easy way to get started with, with, with AI. And then from there on, you could sort of start to adopt more of the kind of advanced concepts. Of course, there's security that needs to be taken into account as well across your estate. So these are the finer details that Copilot, Copilot Studio can allow you to set. So you start your journey off in the right direction. So all of the governance is in place. So what you create for your business or for the users that can use Copilot is effectively sort of a very productive environment where it's all governed, it's all in a nice framework, and they can just get down to business and use it rather than worrying about security. So those are the type of things that we can help take care of, for example.

And yeah, you just mentioned that. Yeah, it's so important that everyone understands that Copilot was in the Microsoft 365 sort of universal ecosystem. Allows you to be 100% in control. And I think that's so important for all these initiatives or people thinking about initiatives, because that's, as we mentioned earlier, sometimes, or very often, a stumbling block for wider adoption. Yeah, with that in place, you can trust Microsoft to then know how to deal with security and all those kind of topics. So yeah, it's a great platform to look at initially, and And yeah, it'll grow. It'll allow more access to other systems, which makes the automation and the use of the data even more comprehensive because then it just plugs into all your backend systems more and more with ease. Yeah. So as Arno said, yeah, have a look at it. If you need some help, reach out to the both of us and we can do more of those sessions with you. So Arno and I, we start as well Q&As regularly on the Automation Guys community. So if you're not part of the community yet, that's definitely something to check out. Yeah, for that you can visit us on our website, theautomationguides.net, and there you can get all the information to join the community and join all these Q&As we do. So very valuable to get the both of us answering all your questions. It might be Copilot, it might be RPA, it might be all sorts of stuff, all things process and automation actually. So yeah, if Copilot is an interesting topic for you and you would like to know more about it, yeah, reach out to us. We get a lot of feedback from the community and that drives a lot what kind of topics we cover here on the show and drives as well what interview guests we get here on the show. So yeah, let us know. We would like to hear that. And if you like that episode, yeah, give us a 5-star rating. Below 5 is not good. So of course all your feedback is welcome, positive or negative. And yeah, we will be back with another episode very soon. And until then, let's automate it!