Free Copilot Chat Tutorial: Master M365 AI Without Paid License

Hello and welcome to this complete tutorial on  how to use Copilot Chat. The version that we're   working with here is the free version that is  included with your Microsoft 365 work license.   This is not the consumer version and this is not  the paid license. This is what you can do just  

by virtue of having a Microsoft 365 license at  work. I will go through all of the capabilities,   the skills you need to make the most of it,  including how to write good prompts and some   good use cases. We're going to have a look  here at how to use secure chat at work and  

how to build agents and how to use AI in the  apps you use every day like Word, PowerPoint,   Teams, Excel, and OneNote. If you get value out  of this, please share it with others. I think   this is a really valuable capability that people  don't know much about. Many of you have probably  

got it and don't even know you've got it because  all of the attention gets put on the premium paid   license. But there's a lot of value here, too.  So, come along for the ride with me. Uh, you   should be able to follow along with most of this  with your own documents. Pause where you need to,  

jump ahead where you need to. Let's have some fun.  Let's start with what this is and how you access   it. First thing you can do is go to your browser  and go to this website here, m365.cloud.microsoft. And this is the hero experience, the landing page.  This is actually the Microsoft 365 landing page.  

So, you'll start here with the chat, but you'll  see you've got some other things along the side   here. The ability to search for documents that you  are working with, as well as the apps button here,   which allows you to navigate around to  all of those Microsoft 365 experiences,  

but I'm going to be focusing here on mostly on the  chat and the agents we'll come back to later. Now,   this is where I generally work with it, but  you might find this in other places as well.   In particular, you might discover it in Teams. So,  if you're just going about your business in Teams,  

you may have seen this co-pilot icon pop up here  in the chat. If you click on that, you will find   the same experience. So, what I'm looking at  here and what I'm looking at here are the same   thing. You'll just notice that when I'm in the  web browser, I've got this nice menu on the side  

that allows me to access many things, plus more  screen real estate, which I do enjoy. But this   is the same thing inside Teams here. So, the idea  here is that this follows you around wherever you   go. We're going to have a look at where it is in  the apps later in the tutorial, but it's following  

you around wherever you work in the context  of your work. Now, this is the free version.   I'm going to emphasize as we go along the parts  where there's a cut off between the free version   and where you would need a paid license. So, first  point here, if you're accessing this in Teams with  

a free version, and you've just found it inside  here, you can do everything I'm showing you in   this video, but this is not actually Copilot  in Teams. It's not doing meetings, summaries,   any of those things. You do need the paid license  for that. It's just a way of accessing that free  

chat experience in the context of where you are  in Teams if that's where you happen to work.   I should also note I'm using a demo system here  which means I'm logged in on a web browser and   I'm using the web versions of everything but these  things all work in the desktop versions as well if  

that is how you happen to work. The core thing to  understand here is that this is secure enterprise   grade chat. It actually uses the same capability  that sits underneath ChatGPT, but instead of   you going out to ChatGPT, which honestly is a  consumer tool where you can do what you like,  

but you should not be putting secure work data in  there that can train the model underneath. This   is actually bringing AI securely into your work  environment. So you're bringing the AI into where   you are already logged in securely rather than  going out and taking your work data somewhere that  

might not be secure. So that's the core thing to  understand when you are logged into this. If you   needed to do it the first time, you might have  been prompted to log in with your work account.   You'll see I'm logged in here. Obviously I'm not  Amy. Amy is my demo account. I'm Lisa. So Amy is  

my demo account here. But we are logged in with a  work account. And you'll know you're in the right   place. And you can check on this by selecting  this little shield icon at the top here. And this   tells you about the enterprise data protection,  which means that when you are using this chat  

experience, Microsoft does not see your data. They  do not have any eyes on your data. Unless you're   explicitly filling a feedback form and saying,  "Yes, please share this thing." They can't see   what you're doing. And anything you do here is not  being used to train the underlying model. So the  

concept of what's going on here is that we have  a large language model basically something that's   been trained on a lot of data from the internet  from anything it can get its hands on basically   that understands how words work together how words  what they mean and how to string them together and  

can use that to generate responses for you. But  anything you do here does not go back into that   underlying large language model. It is safe and  secure here and stays within that work tenant.   Now there is a concept in here of copilot memory  in the sense that it understands what you're  

doing. That is not the same as it taking it back  out into that public large language model where   other people might pick up on things you've done.  We'll come back to memory. So this is safe secure   enterprise AI. Let's get started with the most  important skill that you need to understand to  

get the most out of this which is prompting.  Now, I am deliberately not using the phrase   prompt engineering, which is a whole science that  goes on because as a person at work using AI, you   don't need to understand prompt engineering. That  sounds very technical. I promise you this is not  

technical. This is just effective and clever use  of language. So, this is a prompting skill set and   I'm going to take you through a framework here of  how you can get more out of this. So, what we're   going to do to start with is work with an existing  document that we've got available inside one  

drive. this user is logged in and so this co-pilot  chat can actually work with the documents that   you've got in your one drive or you can securely  upload documents in here to work with as well. So   I'm going to start with the place that most people  start with which is a prompt that looks something  

like this. Summarize this file, right? We're going  to do better than this but I just want to show you   the concept. So what you can do here is select the  forward slash key that sits on your keyboard down   on the bottom right next to the shift key. the  forward slash key. And what that will actually do  

is show you the documents that you as the logged  in user have access to on your one drive. Now,   if you want to be able to access SharePoint, all  of the documents, all of your context at work,   that needs a paid license to be honest. That's  the thing you're paying for. Like that's the  

highest value thing you're paying for. But what  this can do is go to my files and I can select   certain files to work with. So that's what you can  do with the free version, which still has a lot of   value once you know that it's there. So let's  say that we've got a bunch of stuff that we're  

working on here. This is actually showing me sort  of all of the recent things that I've been working   with. I can actually just start to type it in.  So let's say I've got my uh promotional plan for   this Contoso Tea and I can come in here and select  that. So I'm saying summarize this file and there  

it goes. Now, if you use a very, very basic prompt  like summarize this file, you're only ever going   to get a very generic response because if I asked  all of you watching now to summarize this file and   let's um let's actually take a look at what's in  it so that you've got a sense of it. Put a link  

to this document in the description if you want  to follow along. I've just borrowed this from a   Microsoft learn site. So, use your own document or  you can grab this one if you like. There's a lot   going on here. So, we've got an executive summary.  We've got taglines for tea, various things there.  

Now, if I asked every one of you to summarize  this, I would get a bunch of different things.   Some of you might give me a sentence, some of  you might give me four bullet points, uh, someone   might put it in a table, I don't know, like the  output could be different, but it's a very general  

prompt. So, the thing that you need to think  about when you're writing a prompt to get more   out of the co-pilot chat experience is to think  of it more like delegating a task to a person.   So think about if you've ever delegated a task to  someone or if someone's ever delegated a task to  

you. It helps enormously if you understand the  context of why am I doing this task? What is it   being used for? What else is going on around it?  So if you take that mindset, you'll find the exact   same thing will happen here. So the framework to  think about here and you don't have to write this  

full prompt every single time is that there are  four elements of a good prompt. The first one is   the goal. And the goal often looks something like  this. Summarize this file. It's just like what do   I want you to do? And for most of us, that's  where it starts and finishes. That's what we  

end up using it for. But these other things are  the things to consider that make a big difference.   And this one here, context, is the thing that  makes the biggest difference of all. Why do we   need this? What are we going to use this for? So  instead of summarize this file promotional plan,  

summarize the key points listed in this file  in order to prepare me for a meeting where we   are going to be discussing the budget. Summarize  the key points in this file to help me generate a   marketing plan that will fit in with the rest of  our product line. Do you see what I'm doing here?  

The context of this matters a lot because if I'm  asking it to summarize in the context of a budget   meeting or the context of creating a marketing  plan, those are very different things. So if   you give it that context, it's going to use that  context to give you a much more relevant response.  

So we've got our goal context. The next one is  the source. So in this case, the source is the   document that I'm providing this file. So that's  good. I've shown you how to do that, but you can   also add examples of other things you want it to  use or other outputs you want it to use. Summarize  

this file for help me prepare for a budget meeting  and please follow this example that I've used in   the past and you can link that document in there  as well. And the fourth and final part of this is   the expectations. So, we've got the goal, the  context, the source, and now the expectations.  

What kind of output do I want? Now that can be the  length of it can be the style of it. It can be I   want a one pager. I want one paragraph. I want  two sentences. I want three bullet points. But   it can also be how you want it to be presented.  Please present it in a table. Please present it  

in a way that's suitable as a speaker notes for  a presentation. Please give it to me as an agenda   for a meeting. So it can do all of these different  formats. Now these four things overlap a little   bit. Like I said, you don't have to do all of  these things structurally perfect in order every  

time, but these are the things that I want you  to think about as you write your prompt. Why do I   want this? What's the concept? What other examples  can I give? And what is my expected output along   with the goal? And if you follow those principles,  you will get so much more out of this. So,  

let's take a look at what happens if we do  that. So, if we go back to our original summary,   we've got the overview, objectives, tactics,  implementation. I'm going to start a new chat. So,   you can actually keep going with the chat. I could  ask it for follow-up questions and keep iterating  

on that prompt. Starting a new chat basically goes  sort of clear the context and start again. So,   just think about as you're doing it, don't go  in and kind of just completely change the topic   halfway through. It's really good at doing that  sort of follow-up conversation, but if you want  

to start again, that new chat button up in  the top right there is the way to go it. So,   I'm going to say summarize, and I'm going to give  it that same file. So that was our promotion plan in order to prepare me for budget conversation  with our executive. So there's my context. I've  

already got a source in there. I could put an  output in there, but I'm not going to do that.   Please give me a two sentence overview and  a summary of the key costs involved and the   justification and business value for each. Do  you see the difference between a prompt like  

that and summarize this file? So if you're doing  something where summarize this file is enough,   that's okay. I tend to find at least more  than half of the time adding these other   elements in here is going to give you a much  better result. And here we go. It's given me  

the two sentence overview. It's given me the  key costs and the justification. So, because   I gave it that context, I'm in a much better  position now to prepare for my budget meeting. I hope you can see that that prompting skill is  really a language skill. It's a delegation skill.  

Don't be afraid to get in and try it. Don't be put  off by the idea of prompt engineering. This is a   thing that will help you discover value. Copilot  Chat also keeps your conversation history. So,   you'll notice that I've got some things down  the side here. This is a collapsible menu. So,  

if I collapse that up or down again, that  will open or close those conversations. So,   I can go back into the previous one that I was  in before. Or I can view all conversations,   and that will allow me to go in and pick up on a  previous thread. Let's say I was going to write a  

LinkedIn post earlier and I want to pick up  that conversation again. I can go back in,   select that, and continue that conversation.  Or I can choose to delete conversations or   rename them if they're things that  I want to come back to. Generally,  

you can get up to about 30 turns in a conversation  before it just kind of can't deal anymore. So,   that's about the limit of it before you have  to start something new. In addition to chat,   you also have an experience in here called agents.  I'm going to click under the agents menu. Again,  

this can be expanded or collapsed. This explore  agents capability. Now, agents is a word that gets   used a lot. Everything's an agent. So, you might  think of agents as highly technical things that   are, you know, reinventing the world. Yes, that  is true. But also we use the word agents in this  

context here to mean something that you as a user  can have for your individual productivity. And   the idea here is that if you've got something that  is very repeatable, something that you want to do   over and over again, instead of writing a prompt,  you can actually expand on that prompt to provide  

more context and more detail about all of those  things that you're trying to achieve. the goal,   the outputs, the expectations, the context,  all of those things, and save that as a set of   instructions. So, think of this as sort of that  uplift from instead of just writing a prompt,  

this is something that I want to do over and over  again. And I'll show you an example in a moment   where I can basically save those instructions and  each time I need to do that task, I can jump into   my agent. The agent knows the job. I've given it  essentially a job description and it's able to  

perform that task for me. You want to think of  these agents, these personal productivity type   agents, as like little subject matter experts  in one very specific thing. Don't try to get   them to do all of your things together. These are  designed to be they work well when they're very,  

very specific. So, in the explore agents, we've  actually got some agents here that are provided   by Microsoft. So, I'm going to start with one  of those and then we are going to go in and   have a look at an agent that you can create  for yourself. So, you can choose whatever you  

like here. if you're following along, but  prompt coach is one of my favorites. So,   you just choose this and select add. And this  is a an agent that has got instructions in the   background that helps you write a better prompt.  And it actually knows about that framework that  

I was just talking to you about. It's called  the GCSE framework, goal, context, source,   expectations. So, this is actually already got all  of that background information in it and it can   help you write better prompts. One of my favorite  hacks is using AI to help you get better at using  

AI because guess what? It knows how to use itself  better than better than better than you do. So,   once you've got that agent in there, you go over  to the right hand menu here and you can pin it so   that it stays there under that agent menu and  you can come back to it later. If you find you  

aren't using it, you can unpin it and it will fall  down the list and that's fine. But that will allow   you to sit there. Now, before I show you what  this does, I just want to go back and contrast   something so that you can see what this does.  If I go into my um let's say we go back into the  

summarize the promotion plan here and if I said at  the end here, how could I improve this prompt? So,   this is just in the general co-pilot chat.  It's just got general knowledge, the general   large language model. It hasn't got any specific  instructions here about prompting, but it knows  

some stuff. it will give me a reasonably good  answer. It will say um your original prompt   is clear but it could be improved. Here are some  suggestions. So it's giving various things. Okay,   that's not too bad. If I do this in the agent.  So what I'm going to do here is just grab this  

and copy it. And I'm going to come back over into  my prompt coach agent. And what I want to do here,   it's got these quick prompts that allow me to like  this has all been set up. So when you create your   own agent, you can also do this. But I'm going  to choose this one. Analyze my prompt and provide  

feedback on how I could improve it. And then  I'm going to put my prompt in here and watch   the difference in what this does. Because in the  background, it's actually got a bunch of things   that say you should review the prompt using this  framework. You should respond in this particular  

format, in this particular way. So, you'll see how  different this is. It repeats my original prompt.   It provides me with an improved prompt. It tells  me the changes made and why and it is actually   using in the background that framework. So that's  the concept of what the agent can do. There are  

other agents in here that are similar and I would  encourage you to have a play around with these   depending on what feels what feels useful to you.  There's one that can help you with brainstorming,   something that can help you with improving  your writing and various other things in  

there as well. Now that we've had a look at  those out of the box agents from Microsoft,   let's have a look at how you can create your own  agent. Start over here in the next tab down by   clicking on create agent. And that will open this  experience. You will see there are two tabs at the  

top here. Either describe or configure. Depending  on when you're watching this might change where   you land. I want you to click on the describe  tab if that's not where you already are because   this is the easier experience as a beginner to  start working with this. So, you're just going  

to describe in natural language the use case for  your agent, what it is that you want it to do. So,   I want you to think about things that you do  regularly, things that are repeatable, things that   might need a lot of context involved in them, or  things where you've got information on a website  

maybe that you want to use or repurpose. So,  examples of things here could be you could create   an agent that could help you with the clarity  of your communication to say, "I need you to   write in a certain style using a certain type of  voice from a certain perspective or for a certain  

perspective." You might need to have an agent  that can help you with your career to say, "This   is the next stage I'm aiming for in my career. I  want you to give me guidance based on that very   specific job role and the objectives and review  what I'm doing." All sorts of things like that.  

The example that I'm going to show you here is  using a website that is a Microsoft 365 glossary.   So let's say that I am working in a role where  I'm dealing with technical people all the time,   but I'm not technical myself. And they are talking  in all of this kind of jargon. And you know what?  

Every single industry has its jargon. So this is  one that I think is quite a useful helpful thing.   So I am actually just going to grab that URL. This  is a public website that we're going to be able to   use. And what I'm going to do is say I want an  agent that can help me understand Microsoft 365

365 terminology and jargon in a way that will make sense  to me. Now let's say I am a   uh let's say I'm a writer. Let's  say I work in marketing actually.

for a technology company but I am not technical  right so I'm giving it some context here same   principles of the things we thought about before  with the context and the expectations and so   what it's going to do is first thing it will  do is suggest a name for my agent and you'll  

see we'll immediately pop up on the right hand  side here that we've got a bunch of suggested   prompts we'll come back to that so Microsoft 365  marketing guide so if you are happy with that,   you can just say yes, go ahead. I am going to  call this one my M365 jargon buster because I  

want to want to have a bit more fun with it. You  can call your agent Joe if you like. Uh you might   forget what Joe is because you are going to create  lots of different agents. The idea here is that   you will have lots of different sort of subject  matter experts. So call it something that makes  

sense to you that if you create a dozen of these  that you will still remember what it is. So you'll   see on the right hand side there that that's  changed. Let's refine what you call what your   agent will do. Should you share any specific ways?  Do you want it to use analogy, short definitions?  

So to say it should always use simple. No,  let's not say simple. It should always use   clear nontechnical language. I am a professional  knowledge worker. It would help to have analogies.   that would make sense to me or use cases that  would help me understand. Now, you'll see my  

typing is terrible here. It actually doesn't  matter. I've misspelled some things. It will   get it. It's able to handle those things.  When I'm doing this and I'm not on camera,   I am glad to correct my own spelling mistakes  though because I actually really hate doing that.  

that sort of show you that you don't have to type  it perfectly because it's interpreting my natural   language and restructuring it underneath. Okay.  So, would you like it to reference any publicly   accessible websites? Well, yes, I would. So, I'm  going to put that one in there and it will now  

bring that knowledge in. It will refer to that.  If I don't do that, it could probably still do   the job with its general language model, but I'm  actually grounding it, pointing it to a knowledge   source that I want to use. So now we've actually  got a problem here because the limitation on this  

is that we can only go two levels deep in the  URL. There's a bit of a catch here, but I think   the Microsoft learn site of Microsoft 365 would  be fine. So we'll say um that's fine. We'll go   ahead and use that. So just be aware of that one.  Sometimes that doesn't work. This is now going to  

have a lot more knowledge down the tree than just  that jargon site. If you have control over where   something sits in a website, you might be able  to change it. But if you're using a publicly   accessible website, uh this is sadly just one of  the limitations. All right, let's come into the  

configure tab here. So now we have a chance to  see what it's done, make any changes. You don't   have to do this part. You can just create it,  but I want you to understand what it's doing.   So first of all, here's the name. These are all  editable boxes. I can go in and change the name,  

change the icon if I like. I could use Copilot  to generate an image that I could use as an   icon if I wanted to do that. It has a description  and a set of instructions. So the description is   like its overall purpose and the instructions are  based on what I described before about basically  

this is the job description. Now you can have  up to 8,000 characters in here. You'll see I'm   only using 456 of the 8,000 characters. I can  go a lot further with this. And again you can   actually use Copilot to help you with improving  instructions and writing instructions. You could  

have something here that says, "Please review this  in the following order and give me this output." A   little bit like we saw with that prompt um coach  before where it's giving you a structured output.   So I could say um you know, please always provide  the structured output with the definition and then  

an analogy and then an example use case and then  a clear definition. So you can go quite far with   this. I'm going to keep it quite simple for now  because it doesn't have to be complicated. The   last thing, the second last thing rather we've  got here is knowledge. So you'll see that with  

the knowledge here, I could enter a URL. So if I  hadn't already put a website in, I could do that   or I could put more websites in if I wanted to do  that. There's the one that I've got. Um, we could   restrict usage. So at the moment, it will use its  general knowledge and other sort of things as well  

as this. I could say only use that. So let's say  we only want to kind of go with the legitimate   source that I've put in there. That's as far as  I can go with knowledge in the free version. If   you want to do things like connect to all of your  documents on SharePoint, connect to your emails,  

connect to Teams, upload documents, all of those  things, that's available in the paid Microsoft   365 Copilot license. So, this is where we cut off  with the free version. I can have instructions and   websites and that's it. If you want documents and  other all of your work data, again, that's what  

you're paying for. If you want to do something,  if you're sort of more in an administrator maker   role rather than a knowledge worker role, and you  want to publish an agent that is for the whole   company to use, let's say we want something that  is all of the SharePoint documents and all of your  

policy documents for employees to ask questions,  you can do that, but you need to go a little bit   further outside of this experience. Copilot Studio  full version allows you to create something kind   of enterprise wide with all the required security  and controls, but then you can publish that back  

so that your users can consume it here. There's  licensing associated with that. That is absolutely   not free. So to be clear, in the free version, as  an individual user, I can create something that   helps me with my personal or team productivity,  instructions, and website knowledge. Anything  

more than that, you are into paid licensing and  different tools in the stack. You've also got the   option here for it to create documents, charts,  and code. This enables something called a code   interpreter. You don't have to know code. Just  gives it an extra capability in the background to  

be able to do those things. This use case doesn't  want it. But if you doesn't warrant it rather,   but if you've got an agent that you want to be  performing more sort of analytical type tasks,   toggle that switch on. And we could also choose  to let this agent create images. I don't really  

want it to do that, so I'm not going to toggle  that on either. But you have those options there.   So now we have got a whole lot of suggested  prompts here. So based on my description,   it's put in these suggested prompts that the user  could just click. But in this case, these things  

are very specific examples of what we could do.  So I could use them there, leave them there as   examples of what we want to do. But realistically,  I'm creating this for me. I want to just be able   to go and say, what does this mean? and it just  knows the instructions in the background. I'm not  

going to use any of these quick prompts. You don't  have to have them. So, you can edit any of these.   You can have between zero and six of them. In this  case, I don't want any of these in the agent. But   you can see here, you can just go in and edit or  change them. So, some use cases, they're really  

useful. In this one, I'm just like, no, I just  want to be able to go in and type that in. Now,   I can test it here. I'm just going to go ahead  and create it. You'll find that as you use it,   you might want to come back and improve it. You  can absolutely do that. This is not over the first  

time you create it. You've also got the option  in here to share it with your colleagues. So the   share button here will give you a link that you  could use to share for specific users or anyone   else. Hear what I said earlier about this. All  this will do is share a link for them to be able  

to use it. They can't edit it. It's not like it's  being deployed properly across the organization.   This is meant to be just sort of for your own and  for your team's productivity. So let's say you   know the I find this useful and I want to share  it with the rest of my marketing team. That would  

be the use case for that. So for now I'm going to  make it only for me and go to my agent. And there   it is. It's over the side like we saw before.  I'm going to pin that because I want to keep it   there. And I am going to ask it. Let's go back in  and find something that was in the glossary that  

um that we might want to have a look at here.  GDPR. There we go. There's a good one that comes   up in marketing all the time. what is GDPR? Now,  it's going to use it's possibly not a Microsoft   specific term, but it's going to kind of give  me all of these things. So, this is a regulation  

um from the European Union. What does it mean for  marketing? Because remember, I've given it that   context of that's what it means. If I just used  the general chat instead of an agent and said,   "What is GDPR?" I would get a general definition  of it, but it wouldn't necessarily be giving me  

the Microsoft related context or the marketing  related context. Let's try another one here. What what does global administrator mean? Here's a simple way to think about it. I feel  like this is quite good. like it's not it's  

not kind of dumbing it down too much for me, but  it's explaining it to me in non-technical ways   in marketing terms. So, you'll see what it's  doing here. It's not just giving me general   answers. It's actually framing it according to the  instructions. Now, I've kept this one very simple,  

but if I wanted to do more with it, I can come  back in here and click on the three dots next to   the pin there, select edit, and then that will  come in here, and I can come back in and edit   those instructions and click update. the more  you put in here and the further you go with it,  

the more tailored it will become and the more  sophisticated it will become. So that's how   you can use agents in the free version of Copilot  Chat. To round out this section on Copilot Chat,   let's take a look at the concept of memory. Now I  said earlier Copilot does not share your data with  

the large language model. It's not used to train  the large language model, but it does understand   the context of how you're using it securely within  this environment and what you're asking for it and   the things you're working on and the type of role  that you have. And you can use the memory function  

to help it just personalize and tailor things  a little bit more for you the way you like.   So, if I were to say for instance come in here  and just sort of show you the starting point of   this is to say draft an email um to a customer  outlining a new proposal or something like that.  

You're going to get a very generic voice. This  is one of my favorite things, least favorite   things. Dear customer, I hope this message finds  you well. I mean, if you ever want to give away   that something's been written by AI, that's it.  So, what we're going to do is make it sound more  

naturally like me. So, I'm going to tell Copilot  something about the tone and style that I prefer   to work in because this is definitely not it. So,  here's an example. My tone is professional yet   approachable, collaborative, warmth, friendliness.  You do whatever is right for you here. I'm going  

to say, "Please remember this when I ask you to  write emails." key phrase here. Please remember   this. So, watch what happens here is we get this  little piece here that says memory updated. So,   you'll see, whoops, let's jump down a bit. It  goes in and actually kind of takes that and  

you can use this for other things as well. You can  say remember that I always like to start with this   grading. It doesn't have to be just email. This is  a really nice easy example. It can be other things   about this is my role and this is the focus I have  and this is the type of approach I have. Now, you  

can see from here that I can actually go straight  into manage memories from here. But the other   place you'll find this is to go into the three  dots at the very top right into settings and then   into personalization. Now, you have control over  whether you want to use this or not. You don't  

have to use it. There are two things in here.  We've got copilot memory. So, you'll see that that   memory has just been added. And if for some reason  I didn't want that there, I can get rid of it.   Now, it will only pick up on things where you  either explicitly ask it to remember or where  

there's an implication. I prefer this. I want to  do that. Those types of things are the things it   will pick up on. It's worth going in and having  a look for yourself at what it's picked up on and   clear things. Now, this is my demo environment. In  my real work environment where I work with a lot  

of clients, it actually has quite a long list of  you're working on this with this client and this   with this client because there's a repeated use  of those scenarios. and at some point I won't be   working on this with that client or I don't want  it to remember that because it's unhelpful. So  

I can actually go in and remove those memories.  So you've got control over doing those things as   well. If we go back into the custom instructions,  this is where you can just directly type something   in here. So the memory is sort of picking up  on the things that come out of the chat and the  

way you use it. But again, you can put something  in here about your personal interests or goals.   You can also change the tone of uh of how co-pilot  responds to you. So you could say, "Please always   um greet me in a in a warm and friendly tone. Take  on a persona of something else." So we could say  

take on the persona of a trusted advisor and use  a friendly tone or something like that. But you   can actually ask it to do whatever you want here.  So, this is sort of like your influence over the   meta instructions of how Copilot will interact  with you. Uh, and then you can save that and you  

can always come back. And yes, I know I spelled  advisor wrong. We can uh we can go back over that.   So, we'll close that down. So, let me just take  this exact thing again. And this is a terrible   prompt. You know that now because you watched  the rest of the video, didn't you? But still,  

for the purposes of explanation. And so you can  see that the whole tone of the thing is far more   like something that I would write and far more  aligned to what I would do. So if you want to   have that more personalized experience in the free  version, you can do all of those things with both  

the memory and the instructions. Let's now take  a look at what you can do inside the Microsoft   365 apps with the free version. So we're going  to go through in turn Excel, Word, PowerPoint,   Outlook, and OneNote. I'll put bookmarks in.  Skip ahead if you've got a favorite in here. So,  

the core concept of working in the Microsoft 365  apps with the free version is that you have access   to this same chat experience. It's following you  around all of those apps and it works with the   document or the context that's on the screen.  This is a big difference from the free version  

to the full paid Microsoft 365 copilot version.  The full version allows you to kind of get into   all of your business data or affect changes to  the things. This is actually just working with   being able to sort of answer questions and work  with the context of the open document. So that's  

where the limit is with the free version, but you  can still do quite a lot with that. So let's come   in here and we'll have a look at co-pilot. There's  always a co-pilot button on the top menu bar. I   am working in the online version here. This is  the same in the desktop version. So you'll see  

that this is essentially the same experience  as that other co-pilot chat. Importantly,   the navigation panel here allows you to get to  that conversation history. So that is something   that we will come back to. So what I can do here  is to say for instance create a chart showing  

the order total value by shipping method. Now,  in the chat, it understands the context of the   open spreadsheet. So, it knows that I'm talking  about that spreadsheet. But with the free version,   I can't actually interact with the spreadsheet.  So, it won't create the chart in the spreadsheet,  

but it will create the chart in the chat and I  can copy and paste it in there. It also stays in   the conversation history because I'm just working  in this side panel. So, that chart will sit there   and I can actually bring it across into other  applications. Stay tuned. We're going to show  

you that. So, what we've got here is the chart  by shipping method. Is this a Oh, here we go.   Click and collect express standard. All right. So,  that's not bad. And we've got some key insights in   there as well. But that is now exactly the same  chat. If I go back over here and just refresh  

my Microsoft 365 chat experience, you will see  that that is now in the conversation history that   is now sitting there. So even though I've done  that in the context of that Excel spreadsheet,   I can access it from anywhere that I can access  chat which includes going across all of these  

different apps. So you can do all sorts of things.  You can ask it here for key insights. You can ask   it to create charts. You can come in here and say  let's copy this and paste it into the spreadsheet   if you want to put it in there. But importantly  with the free license, it is working in the chat.  

The other thing you might like to know here is  you can expand this out a little bit because it   does get a little bit hard to read along the way.  So, let's take a look at that same concept inside   another spreadsheet here. So, we've got um bring  your pets to work day. I'm going to come back to  

this in a PowerPoint deck. We can open co-pilot  and I can say in the context of this little budget   that we've got for a proposed bring your pets  to work day rather than creating a chart I can   say please write a summary of the costs involved  in this. So I don't have to specify what this is  

because it knows what it is. So we've just got the  cost breakdown and that's all in there as well.   All right, we're going to park that one and come  back to it. Switching contexts here. We are going   into Word. So, I've got a document open and I can  open Copilot on the side. Now, you can use this to  

generate content from scratch. You could come in  here and say, "Can you please generate a document   for me that's a marketing plan for such and such,  but it's not I tend to find asking it to generate   things from scratch without reference to other  things isn't terribly good. You can also bring  

other documents in here. So, what I've got here  is the product description, the key features. So,   what I could do here is help me come up with  a an effective product marketing tagline for   this product. So, again, I'm interacting with the  document. So, here we go. We've got some tagline  

ideas based on this. And again, I could chat with  it on the side. I can copy and paste it in there.   But I can also bring another document in. So I've  got some other things related to the Mystic Spice   Premium T. You'll see this little plus button  appears everywhere in the Copilot chat. So you  

can either upload a document that you've got  just somewhere on your desktop or if you go   into the add work content that will actually give  you access to the things you've got in the one   drive that you are logged into. That was a really  clunky way of saying it. Your one drive. So, I've  

actually already got the promotion plan. So, what  we could do here is to say um which one of these   fits best with the promotion plan. So now I'm in  the context of this chart. It's looking at this   document. I am now looking at my taglines and then  saying based on this which is going to be the best  

and it's actually brought those things together.  So I'm working across multiple documents here.   I could also say compare the contents of this  document that I've just uploaded with the document   that's open and what's missing or what are the  differences between them. So you can actually  

sort of work across multiple things. But with the  free version, it's working with what's here and   what you explicitly add in that little plus button  in the chat. It doesn't as a starting point have   access to all of your work. That's in the paid  version. So let's take a look at PowerPoint.  

And this is my co-pilot chat in PowerPoint with  the co-pilot button here. I can come in here and   open the navigation panel. And you can see that  I can access those agents I was using before. But   remember, I asked when I was in Excel to write a  summary of the costs from Excel. So I can actually  

open that chat here and bring that in. So let's  copy that. And with the free version of Copilot,   I can't actually in the chat ask it to make  changes directly to the content. That's the   paid license. But let's just grab another slide  in here. So, I'm going to maybe just at the end  

of this section here, um, add a new slide. And I'm  going to call this one cost breakdown and paste   the text into there. So, there we go. That puts  it in there. Now, you don't need to watch me do   PowerPoint formatting. You can do that. But let's  say I wanted to add an image in here. So I could  

ask Copilot here to generate an image showing a uh  a dog with surrounded by muddy. I honestly think   sometimes that image generation is a little bit of  a gimmick, but if you are trying to do something a   little bit fun in a work context and you want  to be able to generate something specific. So  

all of the other this is actually an AI generated  PowerPoint presentation. It uses stock images, but   you can use the copilot chat to generate images  and to do that as well. So we will give that a   minute. And here's the big reveal. Now obviously  I could have been more specific in my prompt with  

what type of dog and what type of image I wanted.  Uh we are getting a golden retriever, I think,   because that's always apparently the generic dog.  Just while that's coming through, remember if you   do a copy and paste, always get rid of the top  and tail because it pulls in that would you like  

me to do the next step. And honestly, I put  the cost breakdown up at the top there. So,   just whenever you do that, just the top and  tail of the AI will catch you out. And so,   look at our beautiful rich golden retriever  coming through here. And actually, I think we've  

got a Labrador. And I know that this image is AI  generated because honestly uh I have a Labrador   and there is no way that Labrador would not just  be eating said money. All right. So again, what   we can do is copy that and bring it across into  the deck. So there is always this sort of little  

bit of work of doing these things. Is it a golden  retriever or Labrador? Someone needs to tell me.   I think we might have a bit of a hybrid in here.  So but you can see how you can sort of work side   by side. And then don't forget in addition to  co-pilot, designer is part of PowerPoint and  

designer is sort of offering me something here  that helps me with that formatting in the end.   So use these tools in conjunction with each other.  Let's take a look at Outlook. Now the copilot in   Outlook here can only work with the email that's  open in context. So this same principle with the  

free version, you can only work with the things  that are open in the context. You can't work with   the deeper data that requires the paid license. So  you can't do things here like summarize my inbox,   identify action items in my inbox. You can only  work with the open email. It also can't access  

your calendar, all of those things of all your  email and your calendar. You do need the paid   license. So to be honest, this is fairly limited.  I think this is probably only useful if you want   to use it to help you draft emails or manage like  if you've got a really long thread, I think it  

could be really helpful. So you could say identify  action points from this thread and it knows what's   in there. So you can see this is one that's kind  of gone on and on a little bit. So we've got sort   of the you know what's going on here. So you could  say like there's two people involved here. So  

that's pulling those action points out. I could  say draft a response and remember that I used   the memory before. So now it still knows that cos  I'm still in that same chat experience where it's   using that memory. And so you'll see it's using  that same language that I said, professional yet  

approachable. You'll notice there's also no I hope  this email finds you well. We're starting with the   hi and the cheers. So again, this is something  that I would need to copy and paste and actually   put into the email. So it can help you in very  limited circumstances there. I think the highest  

value thing here is that you can use it in the  context to help you draft a response to the email   that's on the screen. You just have to do the  copy paste, but use that memory trick to get it   to write to sound like you rather than sounding  like a machine. If you don't like the em dashes,  

you can also in the memory say please never use em  dashes. So you can go ahead and do that as well.   And OneNote, which is often the most underrated  thing. I live in OneNote in my real work and in my   content creation. This is an example of something  I've put in there that was actually sort of a  

timestamp of a video I did a couple of years ago  uh of an ask me anything. So, I thought I'd see   what we can do with that because it's something  real. So, what can I help you capture? So,   please create a list of the key topics covered in  this list. And you can see that it's working with  

what's on that page there. So, this is sort of  my index of all of the topics we covered and it's   actually sort of grouped them in a certain way  and turned it into something that I could use. So,   same sort of thing. I can interact and repurpose  in there. Couple of final things I want to show  

you here. Back in this main experience, as I said  at the start, this is actually your Microsoft 365   app. So, even though the co-pilot chat is the  main experience here, you might also like to go   into this search bar. The search bar allows you  to search obviously for things that you've got,  

but as you search for things and find them,  you've actually also got little suggested   prompts down the side here. So, for instance,  we could create an FAQ out of that document,   and that's like an instant prompt that pops up and  does something that I wrote before. And look at  

the way that it's written the prompt in a really  good way for me from that single button click. So,   that's something that's worth going in and  just having a bit of a an understanding of   and ideas of things you can do. You've also got  suggested prompts that you can use here. So,  

underneath the main thing, you've got a bunch of  different suggested prompts. You can open this see   more and go in here and see some other ideas of  what you can do or go into the prompt gallery here   and see even more ideas. There is a thing here  called your prompts. So if for instance I come in  

here and let's go back to the FAQ and let's say  I really like this. I could save it by clicking   that ribbon icon and then when I go back into that  prompt area that's something I could use again.   So I can start to build out and kind of save the  prompts that I want to use. Or of course, don't  

forget you could turn some of those into agents  if they are highly repeatable things. So there's   a lot you can do with the free license. Obviously  with the paid license, you are getting that sort   of grounding in the full context of all of your  data at work as well as some premium agents and  

other premium connections. There is actually a lot  more in there, but as a starting point for free,   I hope you can see there's a lot going on in here.  If you can get yourself across some really good   prompting skills, use your documents in there,  work across those apps, basically just try and  

use it in the context of where you're working,  I think you'll find a lot of value in here. And   it's certainly a very good place to, you know,  start to get used to using AI at work in the   tools that you're using every day. If you don't  or don't yet have that full co-pilot license,  

there is absolutely plenty of value here. Please  let me know in the comments if you came the full   distance of the tutorial with me. I always love  hearing that. Thanks again and happy Copiloting.