Getting Started with Microsoft Copilot

So welcome to the Microsoft Copilot class. Today we're going to be learning about how to use Copilot in the various different Microsoft accounts, Microsoft apps, and how it works. How it works might be different depending on the type of account you have and whether or not you're paying for Copilot.

So we're going to talk about all that sort of stuff. So the whole idea of using AI in general is to make ourselves more productive, to try to make us more creative, and to enable us to do things that are beyond our normal capabilities because AI can do things that we might not know how to do, but we can ask it to help us in doing those things so it can make us more effective, be able to expand our capabilities. If you've taken our ChatGPT class, and multiple people took it yesterday, there'll be a couple little repeat things here just in terms of fundamentals.

Some of you are new and haven't not taken that ChatGPT class, so there'll be a little bit of review in terms of general AI because there are some similarities between things like ChatGPT and Microsoft Copilot. And just so we understand the relationship here is OpenAI created the technology that powers Microsoft products. So OpenAI created ChatGPT, and Microsoft invested in OpenAI, and they're calling their version of ChatGPT, they're calling that Copilot.

So Microsoft Copilot is similar to OpenAI's ChatGPT. The difference is that ChatGPT is basically a website, or you can also install an app now that you can use, but basically the app is very similar to the website in many ways. And that is, you're basically having a chatbot.

It's a virtual assistant that you chat to. And fundamentally, that's the way that Microsoft stuff works, but the big difference is that Microsoft is embedding that into all of their applications. So instead of just having the website, think about having that chatbot in all of the Microsoft apps.

AI Classes & Bootcamps

• Live & Hands-on
• In NYC or Online
• Learn From Experts
• Free Retake
• Small Class Sizes
• 1-on-1 Bonus Training

Named a Top Bootcamp by Forbes, Fortune & Time Out

Learn More

So having it built into Excel, having it built into Word, and Outlook, and Teams. Everywhere you go in all the Microsoft products, you get Copilot built in. So that lack of friction of having to go over to ChatGPT to the website, do something, and then come back.

If you wanted to write something, if you're using ChatGPT, you have to copy it, copy an email, go to ChatGPT, say, can you respond to this? And then you have to copy the response, paste it into Outlook. When you're doing this with the Microsoft Copilot, you're in Outlook, and you just say, respond to this. Or you say, write this thing.

So you're in the app. And so that lack of friction means you're going to use this a lot more if you're using Microsoft apps. Now, not everyone uses Microsoft apps.

So if you're using Slack, for example, instead of Teams. If you're using Gmail for mail instead of Outlook, and Microsoft hosted email. If you're using other competitors like Dropbox instead of OneDrive, you're not going to get as much of an advantage out of Copilot.

Copilot is what I like to call the walled garden of AI. Microsoft is wanting you to take and put all of your stuff with them. They want you to use their OneDrive instead of Dropbox.

They want you to use Outlook instead of Gmail. They want you to use Teams for chatting and for meetings instead of Zoom and Slack. So they want you to use all of their services so you can get the most out of Copilot.

And they're using Copilot as a way to say, come to the Microsoft side. And then you get Copilot everywhere. So when we think about the AI products that they have, they're basing it on the tech from OpenAI.

But that doesn't mean it works exactly like OpenAI's chat GPT. There will be similar things. You can do text chats.

You can create images. But the big differentiator is that the Copilot features are built into the Office apps. That's the big differentiator.

Being Microsoft as they are, there are some confusing things like there's a million different websites that you can access this through. Their thought is we're putting Copilot everywhere. So wherever you happen to be, if you're using Bing, which is their version of Google, so their search engine, if you're in Bing, you can chat right there.

If you're on Office.com because you're using Office on the web, you can chat right there. You can also go to the Copilot.Microsoft website. Are these all the same? Yeah, basically, they're all the same.

They're just different websites because they're just putting it everywhere because they have all these different websites. And they're like, we're just going to put Copilot everywhere, even though it's the same stuff. And we'll go through these.

They want their money back from their $10 billion. So they are pushing Copilot very hard. I know when I first was looking into Copilot, I wasn't sure, does my Office include Copilot? Because I didn't understand what the deal was.

And so I looked and I couldn't quite tell. It seemed like I didn't have Copilot. It seemed like it didn't work.

So I contacted Microsoft and they said, no, it's not included with your traditional Office. It is a separate subscription, a separate purchase. And then you can purchase it for $30 a month as part of your company organization.

We'll talk about pricing more later on. But essentially, if you're an individual person, you can get Copilot Pro, the paid version, for $20 a month. Or when you're part of an organization and they're getting it for all their employees, for them, they charge $30 a month as part of the organization.

So when I asked about this, I wasn't ready to buy Copilot at that time. I kept getting calls, emails for multiple days. They're like, do you want to buy it? Do you want to buy it? Do you want to buy it? And you want your $10 billion back.

Okay, I get it. They were very aggressive in their sales. So eventually, we did purchase it.

Not for everybody at the company, but for a few people, just so we could start to test it. And especially for myself to be able to teach this and start to use it. But their thought is, we're putting this everywhere.

We're putting it into the Windows taskbar. We're going to create a Copilot button that's always there in the taskbar. We're adding the first new key to the keyboard in a long, long time.

We're adding the Copilot button to new Windows computers. We're going to make people put in the Copilot button. So they are seriously behind this.

They're embedding it everywhere that they have it. Figuring that they want to try to get their investment back. And they're thinking this is the big new change that is going to change all of their products.

Behind the scenes, it is driven by what's called a large language model. Basically, there's a lot of training that has been put into this AI. And this is OpenAI's training, as far as I know.

What I don't know, they haven't really talked about it. They're powering Copilot based on Chat2BT's technology. I don't know if Microsoft has done their own different training or not.

I doubt that they've done their own separate training. They're probably using OpenAI's training to do the models. But I don't know.

They haven't really been forthcoming with that information. But essentially, whether it's OpenAI's training, whether it's Microsoft's training, they input lots of text, things like Wikipedia. They probably took all of Wikipedia's information, trained on that.

And they look for patterns to understand human language. And they train it on a bunch of voice as well, because they have text where you can talk to it by texting, or you can use voice as well. So they trained it on a lot of voices to listen to language with different accents and different languages.

And by training on all of this vast amount of knowledge, they start to see patterns. They start to understand human language. And basically, this all kind of came about by thinking about predicting the next word.

So that when you're typing on your keyboard and it predicts the next word, they wanted to do that better by looking at language. And they realized if they just keep doing it, they can talk. And they're like, hey, this thing generates.

That's why it's a generative pre-trained transformer, because it's pre-trained. They give it a lot of information. And it generates text.

And it's just basically a prediction engine. It predicts what comes next, and next, and next, and next. It's a bit more complex than that, but pretty much it's a predictive text generator.

And just that's how it can write things. So the interesting part about these LLMs, or large language models, is that while humans can guide some of it and give it positive feedback, negative feedback to help it, a vast amount of that training is done autonomously, on its own. So these algorithms kind of teach themselves.

So the computers, we can just give it a lot of information, and they find the patterns. They kind of figure these things out. When we have these AI models, it's how we've trained the models, and how we've kind of taught it to work.

But a lot of this stuff, they kind of do on their own. They've created algorithms, how it can kind of teach itself. And that way, we don't have to be telling it, this is the right answer, this is the wrong answer.

We've kind of taught it, kind of like a person, like with kids, we give them a lot of knowledge, and they learn that, and then they can start to generate their own responses and answers. We kind of did that with computers. But in terms of like, does it really know any of this stuff? It's looked at things, it's looked at patterns, it tries to pick out patterns to respond with, but it might not really truly understand what it's saying.

And when it gets things wrong, but it thinks it's right, we call those hallucinations. Oh, the AI is hallucinating, meaning it's getting it wrong. And you might say, you're wrong, and it's like, oh, sorry, here.

And it might do the same thing again. You might say, are you sure about that? And it's like, yeah, yeah, I'm sure that's correct, because it sounds right, but it doesn't really know even that it's wrong. So one thing we always want to double check is, is AI doing the right thing? We don't just blindly trust that it's correct.

It can be very good, and it can be correct, but it can also be wrong, even when it thinks it's right. Now, when it comes to privacy, think about that training that they're doing. They train a lot of stuff ahead of time.

And when it comes to Microsoft, because they're targeting businesses, Microsoft does not use any of your data to train those models. So all your data is protected. They don't train any of their stuff.

So ChatGPT has innate built-in knowledge. And when Microsoft looks at your company data, all the learning that it does based on your company data is all private. Unlike ChatGPT, where some of their accounts, you have to go turn off training, some you don't.

It depends on which account you have. With Copilot, they never train on your data. You don't have to go in and turn it off or anything.

They just know that they don't want that stigma of, are you leaking private information to the AI to train? So all your stuff is secure and private, specifically for your company. And if you want to read their privacy and protections and all of their security policies, I've linked out to them here if you want to read more about it from Microsoft themselves. But keep in mind, Microsoft is not focused on the regular consumer as much as they are business.

So Copilot is more of a business offering versus, I think ChatGPT is kind of an all type of offering, not just business. I think Microsoft is more targeting the business side of things. Could an individual just kind of normal consumer use Copilot? Sure, of course.

But they know there's more money in businesses. So I think that is their main target audience is businesses. So they default to being private by default.

Now, as far as features and pricing and what we're going to be able to do for today, this will depend on your account. So there is Copilot and Copilot Pro. These are for individuals.

And then there's what they call Microsoft 365 Copilot. So they've rebranded Microsoft Office into Microsoft 365. So most people still think of it as Microsoft Office, but that's Microsoft 365.

So with Microsoft 365, just by itself, Microsoft 365, that is your Office applications. So that's your Outlook, Word, Excel, PowerPoint, those kinds of things, right? Now, if you want to add Copilot to those apps, notice you can do that with the Copilot Pro, which is $20 a month as an individual. Or if you're in an organization where your organization has licenses that they assign to various users, they do that at $30 a month.

Notice the free Copilot, just the regular Copilot, which you can get for free, does not integrate into the applications. So the free Copilot is basically just the chatbot where you can go to the website and you can chat with it. It's not going to have as many capabilities.

And to be honest, for free, I would rather use ChatGPT because you're going to have more capabilities for free using ChatGPT rather than Copilot. But when you pay for it, you can get it integrated into the apps. So I think for myself, if I was choosing free versions, I would just go with the people that make the technology.

I would just go with ChatGPT if I'm just doing it for free. So to me, really, Microsoft, when I'm choosing Copilot, I'm choosing it because I'm paying for it, getting it integrated into the Office applications. Think about that if OpenAI creates the technology, they get all the new tech first, then it can go to Microsoft and then they can start to integrate it.

So if you want to be getting more features on the cutting edge, you go to the people that make the technology, which is OpenAI. But what ChatGPT and OpenAI don't do is they don't integrate into all the Microsoft apps. That's where we have to pay for Copilot to get it integrated into all the apps.

But notice that even the free version, you still get your data protection, meaning they don't train on your data. The other thing is this grounding part. This is kind of like the training part of things.

When you pay for the organization side of things, they're going to be able to look at your organization's data and be able to learn from that data. And it's a different way to interact. We're going to see that there's going to be a web part and there's going to be a work part.

And that's what we get through this, where the work part has access to things like your OneDrive and your files and your emails. And you can interact with that company data that's stored in Microsoft because you're using Microsoft Teams, you're using Microsoft Outlook for your email. And that's going to get us, that we can interact with those, that work data in a different way.

If you are generating images, you get more images, of course, when you pay. You only get a few for free if you're generating images. And we'll see how to do these, of course, later on.

But for the most part, I kind of view Copilot as really the ideal way to interact with Copilot is you have to pay for it. Yes, there's a free version, but the free version even looks very different. The free version is really kind of more geared towards consumers with very simple needs.

And it can't do a lot of the stuff that the paid version can. So the web grounding part and the Microsoft Graph grounding, the web grounding means that it can go and get information from the web. I don't know why they call it this grounding part, but so if it's web grounding, like you saw here, that means it has access to the web.

So there's the pre-training and then there's the, let me go out to the web and find something. So if you say, how's the stock market doing today? If it can't browse the web, it can't understand how the stock market's doing today. So all of them can access the web.

And the Microsoft Graph grounding, that is that it can go in and connect to your Microsoft 365 services, your emails, your files, your meetings, your chats, those sorts of things. And so being able to access those, notice that Microsoft Graph grounding, that is in the paid organization side of things for the $30 a month. So keep in mind that if you're using the free version of Copilot, the interface is actually different than the paid version of it.

And we're going to see that. So what you see and what I have might be different. And we'll talk about those differences when we get to that.

They also, because this is a chat bot, essentially Microsoft, you might see that it's, they call it biz chat or business chat. That is for people that have the Microsoft 365 Copilot. That's the paid, the one that's above the pro, the one that's for organizations.

That's the level that I have here is the organizational Microsoft 365 Copilot. So the free version, this is how the free version looks. It's got this very rounded corner.

It looks very kind of supposed to look futuristic, but yet friendly for people, that it might be scared of AI. And so they just made it super friendly, very simple. And you can message the chat bot.

You can talk to it if you want to by clicking the microphone. The paid version looks entirely different. So the free version here, so like, hey, Dan, it's great to see you, friendly and you can message it.

In the paid version, this looks more a bit like chat GPT if you've used chat GPT. And in the Microsoft 365 Copilot, you have the web versus work. Remember that the Microsoft graph grounding parts, that's the work versus the web, right? So in your Copilot Pro, you don't have the Microsoft graph grounding.

That's the work part of things that you'll only get the organization. So you do get the web, but do you get the work tab? That is only if you're going through for the Microsoft 365 pilot. So some of you must have the other part and not part of thing plan.

So my company pays for this, then Copilot is connected to all the company files and stuff. So I can ask about company names. And then you can answer by chatting.

Right, you have access, like you can pull things from OneDrive, for example. So you can start to do that through the work side versus the web is more the general AI's knowledge, not specific to your work, right? So it can browse the web and uses the general AI knowledge, whereas the work tab uses your company. It can access your company files.

We're going to see it as we do it. So if you just have one thing, basically you're just going to have the web. If you don't see the work tab, it's because you don't have the organization accounts and you're going to be in the web tab if you don't have both tabs.

So if you have a business account or work, is linked only to email or anything? So if I use someone else's email and login, that's not going to work because it's connected. Then I'll get their email and all that. Right, so yeah, so Copilot account are particular to a person.

So at an organizational level, so basically you can't share, you cannot share a co-pilot account. I know people want to save money and they want to share a co-pilot account, but when you sign into Microsoft 365 products, you're signing into your accounts. And so in an organization, every individual user needs their own accounts, is $30 a month per person.

And when you're logging in with their email, that's their stuff, right? Their files, you're logging into their accounts. You're not just sharing a co-pilot authorization, you're signing into their OneDrive, for example. You're signing into their email and it's only going to work like, and we'll see this later, but for Outlook, in Outlook, it's only gonna work for the email that like, let's say you've hosted your email with Microsoft.

It's only gonna work with that organizational email. It doesn't work with all email addresses in Outlook. So if you've added multiple email addresses in Outlook, let's say you have a personal email and a work email, it's not gonna work with your personal email because it doesn't authorize Outlook to use co-pilot.

Your email account itself individually is co-pilot enabled. So it's only gonna work with that one email address. It doesn't enable it in the app because they're doing some server side processing of stuff.

So they need to host your email to be a Microsoft email account. So like if you're hosting your email, companies can host with Gmail or they can host with Microsoft as far as who's providing the email address. You need to go to Microsoft to host your email so that they can have it on their servers and they can process it and so forth.

You can't use Gmail to host your corporate company email. This is the walled garden that I was talking about where they're like, you bring all your services to us to take full advantage of Outlook and Teams. So don't use Slack, don't use Dropbox, don't use Gmail.

Put all your services in us because that's where you get co-pilot working because co-pilot won't work with Zoom. They have their own AI. It won't work with Slack.

It won't work with Dropbox. It works with all Microsoft products. And so it all has to be hosted with them.

And so this is where they just want you to keep bringing more and more stuff to them. And I understand, I mean, how can they integrate into all the other things if they don't control those things? So, yeah, I mean, it's an ingenious plan and I understand why they have to do that, but it does kind of make it that if you want to truly take advantage of everything co-pilot has to offer, you need to go all in on Microsoft as much as possible to take full advantage. Now, you can still use it in just Word or Excel or PowerPoint.

You don't have to use it everywhere, but the more you use it, the more you want to use it. And so to really fully tap into everything that co-pilot can do, you need to be fully invested and do as much as you can in the Microsoft world of things. Yeah.

So, and we'll get in, this is just an overview here and we're gonna see how all of this stuff works, of course. This is just to give us an overview so we can understand that once we dive in, kind of what's the framework here. Now, there's also, because this is Microsoft and they need to confuse you a little bit, there are two co-pilots.

There is Microsoft co-pilot and then there's GitHub co-pilot. And I was disappointed that when I got Microsoft co-pilot, that it was not GitHub co-pilot as well. GitHub co-pilot is for coders.

I do both office stuff, I also do coding as well. And so Microsoft co-pilot is their general AI that's for everybody. People that want to use it in all the programs and use kind of the chat feature of it.

It's kind of like chat GPT integrated. It's specifically for coders and I use the coding app called Visual Studio Code and I would love to be able to get AI stuff in Visual Studio Code, but I cannot use co-pilot, Microsoft co-pilot to do that. I have to buy a separate subscription to GitHub co-pilot.

So they're like, well, you can pay us for your Microsoft 365 subscription to get all of your office apps. You can then pay us for Microsoft co-pilot our co-pilot to those apps, but to use co-pilot in your coding app, you can pay us another monthly subscription. And I think that's another like $20 a month just for coders.

So there's two co-pilots. The vast majority of people are probably going to be using the regular Microsoft co-pilot of course, but if you're a developer and you write code, there is a separate GitHub co-pilot and they are two different products because GitHub co-pilot is just for coders and does not give you the regular integrated chat bot into all of the different apps. They're different products that serve different goals and they're separate subscriptions.

So they're like, we want all the money. So just to be clear, because I was excited to get co-pilot and start to code and I'm like, oh, no, that's another subscription, different thing. So we're going to be learning Microsoft co-pilot, not GitHub co-pilot.

All right. So there's a variety of ways that we can go and interact with this. We can use office.com slash chat.

We could do Microsoft 365.com slash chat. Either those work or co-pilot.Microsoft.com. These are all the same thing. It gives you the same experience.

Whichever one you want to type in, you can go to any of these websites and we want to log in and we can try sending a message and chatting with the chat bot, with the co-pilot chat bot. And we can maybe ask it things that it can do. So I'm going to go to co-pilot.Microsoft.com. But even if I went to those other websites, they would all still work the same.

So co-pilot.Microsoft.com. They actually redirect. And there's technically even another website you can go to. I can give you every single website that you can go to because there was just too many of them.

But they all, whether it redirects, whichever one you want to type in, they all go to the same fundamental technology. Yours might have a light background. I just chose the dark background.

You can change that theme if you want to. So if yours is a white background, still fundamentally the same way that it's working. Waiting for this to load.

Now, if you have the free version, yours is going to look very different. You'll have that kind of consumer version of it. You can still chat with it in the same way.

You won't get all the same features of it, of course. So I'm just going to be here on the web tab. The web tab is kind of the general tab where, and if you don't have the work tab, you're going to be in the equivalent of this work tab.

Basically, this is the general knowledge, the pre-trained stuff. And it can also search the web if it needs to search the web as well. So I can type in a question.

Just waiting for this to catch up. Tell me 20 things you can do. Okay, and I can send that message to the chat bot there.

Think of this as your virtual assistants that you're talking to. And you can ask your questions. So if you've ever used ChatGPT, this is basically same type of thing as ChatGPT.

And it can give you 20 different things that it can do. And if you type in that same thing, you're probably going to get a different list of things. So go ahead and type that in and see what you get.

There might be some similarities, but it's probably not going to be exactly the same because it is generating its own response, custom for me, which could be different than what it generates for you. So it can answer questions. It can provide summaries.

Can generate things, stories, poems, essays. Help with coding. Even though it's not integrated into my coding app, I could still give it some code or ask it to write code.

Offer studying tips and educational resources. Assist with math problems. Create to-do lists.

Recommend books. Explain complex concepts in simple terms. Provide travel tips.

Help with learning languages like Spanish, French. Offer mental health tips and self-care advice. It can be your therapist that never makes judgments and you're not afraid to admit things to it.

No privacy concerns. Assisting with job search and resume writing. Yeah, you could put in your resume or put in your work history and say, write me a resume and it could help with the resume.

So now one thing that's different here is there's some view prompts down here. So as far as ideas, because it's just a blank slate where you can ask it anything, Microsoft I think knew that people just don't even know what's possible. And they're like, we wanna help people because they know that they're going to kind of an average office type person.

And not everybody's great with technology and computers. And so they wanna help out everybody that's using Copilot. So they have some prompts that you can view.

They click on view prompts. And here they have some suggested prompts, but they also have different things. They're saying these are examples of things that you can do because you might not think of these things.

You might be wanting to create something. So I could come in here and say, show me prompts about creating things. And these are things that it could create.

And these are just some examples. Create a Gumby style clay model image of the polluted earth. Okay, that would be interesting.

Going on holiday, write some funny out of the office email responses to use while I'm on vacation. So they're just kind of giving you some ideas of things. Have a laugh.

Suggest five office friendly jokes. Visualize your data. Help you to code.

Because you're creating visuals, you're creating text, you're creating code. So just saying, hey, I know you might not think of all the things you could do. So you could turn these things on and off and you could say, well, manage.

How could you help me manage? What are five to six ways to beat procrastination? And we could click on this prompt, of course, to do it, but it's just trying to help you to know what's possible. So this might be something that you wanna browse through just to get ideas of things. So summarize, like get an overview.

Summarize a certain report because we can upload files to it. So these are all just ways for you to learn different things you can do. And these, to be honest, these are just some of them.

There's a see all button here. And when you click the see all prompts, that'll take you out to their website. And you can see all of the prompts that they suggest.

Now, is this every prompt you could ever type in? No, because you can type in anything you want. But again, this is lots of different suggestions because they're trying to help train people in AI. They know this is new.

So they realize that a lot of people don't even know what they can do with AI. And so they're just saying, hey, there's a lot of prompts you could try, trying to get you ideas of things that you could do with AI specifically for co-pilot stuff. And so this website has even more than that little dialogue did.

They break it down into the kind of job you're in. So if you're in the sales industry or finance or marketing, you could say, what's specific to me? Let's say you work in HR. You could say, what's specific to me that I could do here? Draft a planning memo to leverage and implement hiring best practices.

Draft an FAQ about a certain lead policy. So these again are just examples. And they're saying, hey, you could do that in Word.

There might be other things that are in PowerPoint or something like here's a PowerPoint, create a presentation about training materials for new employees. So these are just some examples of like, this is the department I work in, give me stuff specific for me. Or if I turn that off and I go into tasks, these are the tasks.

But notice there's even more than I saw in that other dialogue there. Or if there's a certain app you work in, say like, yeah, I do a lot of stuff in Word. So show me stuff specific to Word.

Or you could check on all the different apps you use and then just see things specific to you. So this is helping people to know what's possible. And this is worth just browsing because there might be things that you didn't realize like, oh, it can do that.

And you just might not have thought it up. So this is part of what they're trying to help solve. Where, you know, these are subtle differences.

Like if you're used to chat GPT, they don't really give you prompt ideas. It's kind of up to you to figure out what you can do. Microsoft is trying to help with that.

So because they know that they're, their average user is an office user, they're trying to give things specific to people in industries, people doing jobs, trying to help them with that. As you create your messages here, you can give it a follow-up here. Like for example, if you're like, okay, that was cool, 20, but maybe you want more.

You could say, give me 10 more, but specifically for me as a, social media marketing manager. Right, because maybe I do social media marketing, right? Just as an example. So you can give it a follow-up and it'll come back and it'll give you 10 more, but specifically for social media marketing, right? And so that gives you those.

So if you want to keep along the same chat, you can keep doing that. On the right here, we've got our chats. If you don't see those chats, there's a button to open up that panel.

So if you don't see those, you can open up the panel over there and you can see your past chats. So you can click on a chat to go back to it. If you want to get rid of it out of your history, you can click the little dot, dot, dot, and then click delete.

And it will delete that out of your history in case you don't want it to be there. There's no organization that you can do with your past chats. It's just kind of listed in chronological order.

I still wish there was some way to organize and create folders or something just to organize this stuff. The idea is that if you're going to create a new topic, a new stream of thought, just like you have different chat messages for different people, you can create different chats because as long as you're staying in the same chat, you're kind of along the same line of thought. And so if you change topics totally, like let's say you're doing something on coding, for example, they're still thinking you're talking about social media marketing.

And so they're going to think, wait, you're a social media marketing person. And now you're talking about coding. They may kind of get confused about things.

So when you want to switch to a totally different topic, create a new chat. And that way, now you have a new clean blank slate to start working. And it's not going to get confused with the other things you were talking about.

Because now you have a new, we call this context. It's kind of wiped away the context and it's thinking just from scratch, doesn't have any preconceived notions, not kind of thinking about what you were previously talking about. It's a new text thread.

You can come up with whatever you want and it doesn't get confused with something that you were previously talking about. It doesn't remember the things from the previous chat in this new chat. They're separate chats with different chains of thought.

Now, as far as the different tabs, if you're on an organizational accounts, so like I said, I have the Microsoft 365 Copilot. That's the organizational, the 30 hour a month one. If you have Copilot Pro, that's still a paid account.

So you can do more than you can with the free one. But the idea of going up to that Microsoft 365 Copilot, that's for organizations. So that's going to add the work tab.

If you have that organizational level and you have the two tabs, you'll add the work tab. So we were in the web tab. So if that's all, if you don't have these two tabs, essentially you're in the web tab.

But if you have that organizational accounts, they'll add the work tab, which is useful for finding stuff within your company. So you can find other people, files, meetings, emails, and stuff within your company. Whereas the web is for searching the internet, general questions.

The web is the general pre-trained knowledge, not specific to your company, but the work stuff is your company. That's why when we get that organizational account, it adds the work tab. So for example, in the work tab here, and you might not have the work tab.

So if you don't, I'm still going to show you how it works. So you can still see it. And then if your company has that organizational accounts and gives you Copilot through that, that would add the work tab at that point.

So let's say I want to summarize a file, just as an example. We'll go through lots of different examples of what you can do in all the various apps, but let's say I'm using this. And keep in mind right now, we're just looking at the chat.

Later, we're going to see, we're going to go through each Microsoft app and we're going to see how this is integrated into each Microsoft app. This is just one way to interact with Copilot. So if I go to work, you're going to notice here, we've got people, files, and more that we can do.

And I could either click that and notice how it typed in a slash, or I could just type in the slash. So I could either click the button or when I hit slash, that brings up this little panel of stuff. And I'm logged into my organizational accounts.

So they know things like my coworkers, right? So I could see something involving a coworker. Maybe somebody sent me a file or something and I could try to find those things. I've got files that are stored in Microsoft OneDrive.

I could also upload files too. I've got meetings created in my Microsoft Teams. I've got emails that are stored in Outlook.

So it can access those things as long as it's in my Microsoft organization, right? So I'm using Outlook, I'm using Teams for meetings. I've got files that are in OneDrive. This is where I talk about that walled garden.

They want everything in the Microsoft stuff typically. Now you can upload stuff, but really Copilot is geared around all the stuff being in your Microsoft stuff. So if I want to say, summarize a file, summarize this, and then want to add a file, I could either hit slash or I could click this button, which will type the slash for me, either way.

And then I could say, hey, there's a file. And these files that it suggests are files that are like recent files in OneDrive. But you can also upload files as well.

Let's see, like these, notice how it's based on when I last opened these files. So they're going to put like the most recent files. So you kind of go back based on recency.

So like the most recent thing going back. If I look at this, let me just show you my OneDrive here. I'm going to go to my desktop here and I'm going to go to my OneDrive.

So I think that's up here. Yeah, so here's my OneDrive. And here I've got some Microsoft Copilot files.

So like there's an Apple report that's in there and got some other class files in here, right? Got some Power BI files. So if I want to access those, like for example, let's say this Copilot chat files, this Apple 10K report. When I'm back here and I'm adding a file, I can either hit this button or hit slash.

And then start typing Apple, give a second here. Should've chosen the, sorry, give a second here. I'm going to click on files here.

So these are based on when I last opened them up. So like there's that one. Or I could hit upload and I can upload a file.

It doesn't have to be in OneDrive, but these are from OneDrive. So let's say I click on that. So that's pulling that from OneDrive.

And then I can hit submit and it will summarize that PDF. You can summarize Word documents, Excel files, PDFs, text files, all sorts of stuff. So this is the Apple financial report, their 10K report.

And it's going through and going to create a summary of that. So I can read through that PDF. And I could also ask you questions about that as well.

So basically anything from the company that part of the ecosystem you can search from. From here. From here versus, oh, this was in the drive or in the calendar, like it's just everything connected.

Exactly, right. And so like, this is all in OneDrive. So they're based on, the thought is based on recency of like, oh, you were just opening this file up.

You probably want to interact with it in Copilot. So they're going to suggest, I wish they had like a nice like browse OneDrive. That's the one thing I don't like.

So like, if you want to say summarize, summarize this and you hit slash or you click that plus button. What I wish is that there was like a browse OneDrive in here to just browse it. Because finding things in here, like it's a very small list.

Yeah, and they were like, oh, if you just recently opened it it's going to be at the top here, right. But I wish they had like a, just a browse OneDrive. Now also, once we see this, there's the OneDrive website where you can also do some of this in OneDrive.

So keep in mind, just one of the ways. So if you want to do something in OneDrive then maybe we should be doing it in OneDrive, not here. So while I do wish there was like a browse OneDrive if I'm here, I can also go do this in OneDrive itself.

But in OneDrive, there's also the Copilot so you can. Right, so we can do that, right. So we can do this here.

So when you're just looking in OneDrive you can say, summarize it there. It might be easier to do it there. So I want you to see both ways.

The chatbot is just one way to interact with things. It is not the only way. And sometimes you're seeing here, oh, actually it's kind of like by not being integrated this way, oh, it'd be better if it was integrated which it is.

So when we go to OneDrive, you're going to be like, oh, that's just easier to be in OneDrive doing it there. So I want you to think that this is just one of the ways we can interact with things as a chatbot but it is not the only way. It's classically what a lot of people think especially people coming over from ChatGPT because ChatGPT is just a chatbot but that's not all of Copilot.

This is just one of the ways to do things with Copilot. And we're going to go through each app to see that. I'm sorry, if I have an opportunity to search and then like, I know, I haven't made one recently but it's got such and such a name, is there a... Yes, you can just start typing to search.

Yeah, so if I hit slash, give it a second here. So if I start typing in, for example, Garfield notice how it searches Garfield, right? So yes, you could just start typing to filter it down. Of course, if you've got a lot of stuff, like a lot of files, that might not be the best way to do that.

So you might want to go into Excel or into OneDrive where you have the file open. This is just one of the ways to interact with stuff. It's not always the best way to interact.

Now, yeah, anytime you're doing anything whether you're in the web tab, whether you're in the work tab, you can do follow-ups. So you can ask it to do something and then you could say, oh, can you make that small? Can you make it shorter? Can you make it longer? Can you make it simpler? You can always have those follow-up questions. And these things are true of any chat whether you're in the chatbot here or whether you're in a particular app as well, right? So if you don't like the response that you got and you want to say, if you're like, oh, maybe that was a little too much or maybe you want to focus in on something like, can you tell me more about risk factors? You know, you could say that.

Tell me more about risk factors. So you could do those follow-up questions and you keep on the same train of thought. So in this case, it knows that we're still working with this Apple 10K report because it has that context.

If we had asked that same thing in another chat, they would have no idea what we're talking about because it's in this chat that it understands what these risk factors are because we're talking about this 10K report. It has that context. So her question for anybody online that's listening, her question was like, if you ask it to give the title of the report, would it always give you the same title of the report? Because if I say this, it generates something unique every time, because it could choose different things potentially.

You know, if you did that same thing, you could get a different response. It might choose to highlight different things. That's the generative part of AI.

But if I said, give me the title of the report, it depends on how clear that is. If it's clear what the title of the report is, then it should give me the same thing every time. If it's not clear what that title of that report is, then maybe it's going to make up that name.

So I can't say for certainty, depending on what you're doing, as far as how accurate that's going to be, because it depends on its assurance, like how sure is it of its answer, right? So like, let's say it's a math thing. If it's a math thing, as long as it's figuring it out correctly, it should always give you the same answer, because it should be doing the same math. But if it's not quite sure of exactly what you mean, then maybe it could get things wrong, and it could start to make up different answers.

Depends on how creative it has to be at answering that, and how sure it is of what it's doing. So I wish there was a simpler answer as far as knowing if it's correct or not, but this is where we always have to get involved of, do we trust what it's saying? It's kind of how difficult is this to answer? If it's a difficult thing, then I might not trust its answer as much. If it's easy for it to answer, then I can be more assured of that answer.

It can sometimes be very confident of its answer, even though it's wrong. And if you ask it, like, are you correct? It might say yes, even though it's completely wrong. It doesn't really necessarily know, because it makes an answer that sounds right, but is not necessarily always right.

So we do want to always kind of double check it. So one thing that's interesting that you might notice here is that there's a button, edit in pages. What is that button, edit in pages? So in Chachapiti, they have something called the canvas, which is like an editable kind of interactive canvas to work on.

It's kind of like Google Docs, where you can interact with other people and kind of work on a collaborative thing. They call that pages. So with pages, we can go in and we can interact with the results and make changes.

We can delete sections or move them. We can edit the contents. And this is more of a collaborative type thing, edit in pages.

So maybe not something that we want to do in here, in this particular example, but maybe if I was doing some research and I wanted to write an article about something, and I want to go in there and be able to collaborate on this article. So maybe I wanted to write an article on the 10 best practices for effective communication. So I'm choosing the web tab because there's nothing specific to my work that necessarily needs to be used for this.

I'm not using files that are in my OneDrive. I just wanted to use this pre-trained knowledge if it needs to search the web to get that knowledge, not work-specific knowledge, right here. And, oh, this one does not have the edit in pages.

So maybe that's something that's only in the work tab. Okay. So let me try that again in the work tab.

I'm going to create a new chat in the work tab. I'll say write an article. Oh yeah, I think it's because you can work with your coworkers that it needs to be in the work tab.

Write an article on the 10 best practices for effective communication. Yeah, so it needs to be in the work tab because you can collaborate with coworkers. You can invite them to work.

That's why it has to be in the work tab. But in this case, it's still searching the web. Still searching, right.

It's not like the work is only limited to your work stuff. It still has all of its knowledge, but you add to it the stuff from your company. So think of the work as web plus work, right.

It still has all the pre-trained knowledge that Chachapiti has that underlies Microsoft Copilot, plus the work stuff. It also has access to that if necessary, right. And here they actually, so here they have the edit in pages because this is something that you can collaborate with your coworkers on, okay.

So now here I can see the different best practices. And let's say there's something here that I wanna, it's still working on this, but maybe there's something else that I wanna do in here. And I wanna have a little bit of a say in some of this.

So there's a little chat thing that I can go over here and click on, right. So let's say providing feedback. So you can comment on this.

You can give feedback to say like, let's say you're collaborating with other people here on this. So if you share this and you give a link, other people can come in and work on this, kind of like Google Docs. So you can share this with coworkers, so they can go in, didn't wanna go to that citation there.

And so they might come in and say like, oh yeah, I like this. So they can give you a thumbs up on that. They might comment that, right.

There's also, if I drag this, I can rearrange these things. Because maybe I want a different order for things. You can go in and start to interact with this.