Alright, so let’s get started. So I’m on my laptop, I hit the start menu, I type in Claude and there it is. So I click into that and it opens it straight up. And now this is Claude Desktop. And so this is the hub basically for all your usage of Claude. Now, this is the same experience you’ll get on the web if you use Claude through Chrome or Edge or Firefox and it’s also the same experience you’ll get out of the mobile app if you choose to download that so you look very you look exactly the same across all of those across all those products and so this is where you can come in and ask you questions and so you can ask you things about your company’s data so long as the connector is set set up and we always suggest that it is set up so I can do things like you know what emails haven’t I responded to what emails have I received overnight and give me a summary things like that and so you can go and query my mailbox I can also ask you questions about information in SharePoint so I might say things like you know what’s the company policy on bringing your dog to work and it’ll go and trawl through SharePoint to try and figure out you know is there a policy around pets and so you know it can do all those kinds of actions.

Now there are some some levers you can pull here the important ones really for me you know in this kind of like baseline 101 getting up to speed really is just the model. So here you can either select Opus or Sonnet. If you stick between those two, you’re going to be fine. And so really what this comes down to is, you know, Opus is a bigger model.

It’s a bigger brain, so to speak. And so if you need something particularly complex, use it. If you just need everyday stuff done, I’d select Sonnet. It’ll be just fine. But don’t stress too much about it, right? There’s a bunch of other things in there that I glossed over just then. Don’t worry about it. Go Opus or Sonnet, move on. This is where you can come and ask questions.

Now, the other part here to note is that this will search across any connectors that have been set up for you and also the web web if it deems it necessary. So if I asked it a question like, hey, Rob Hunter at Lanter, when did he join Lanter? There’s going to be information inside of our SharePoint that has information about that, right? There will be files on when Rob joined.

We probably did a profile on Rob at some point. And so it’ll go and get information there. But if it wasn’t, or it deemed it incomplete, it might also go and do a web search. But if it could answer the question fully from our internal info, it likely won’t go out to the web. And so the one to stop that, if you did want to just have fresh eyes on something, it happens to me quite a bit.

I work with a number of clients, and I’ve got to do that. And I’ve got to do that. I might just want to do some deep dive on their industry. And so I’ll say things like, you know, hey, you know, what’s ComBank? You know, what’s their latest innovation blog about? And it might find some things internally for the work we’ve done with ComBank. And so I want to turn that off.

I don’t want to, I don’t want to, you know, have it leverage any of our existing internal info. So I’ll click on that the plus button, come down here and turn off web search. Otherwise, I’ll just leave it on. Because if I ask it a question about internal stuff, it’s going to find an answer and not bother doing the web search. And so, you know, for most things, don’t worry about it.

Just let it have both. So here’s where I can ask the questions, right? So I’m going to ask a non-work question because this is my production account so it will surface very sensitive things I could say something like what was the score for the Sydney Swans latest game now it’s gonna run off it’s gonna figure out hey we don’t have anything to do with the the Swans internally it’s gone straight to searching the web perfect and And so, you know, it’s going to come back with some results, whatever it could find on the internet.

And so, you know, the disappointing loss to Fremantle. Great. So that’s the first problem. Now, here is where you can, you know, if you want to then talk about something else. So, you know, I want to know about the weather. New chat. Just start a new chat. Go up to the top left here new, we’re in a fresh window, ask it the next question. What’s the weather in Sydney today?

Again, it’s going to figure out, hey, we don’t have anything internally for the weather. We don’t know anything sacred. Claude has a nice plug in it seems, so it’s going to show me something a little prettier. But here is where you have the choice, right? If I were to wanted to dig into this a little bit more, I can keep going with this same chat and it’ll make sense to Claude.

So I could say things like, you know, what’s the week ahead look like? Now, this question on its own, Claude would have no idea what I’m talking about. What’s the week ahead look like? For what? But because I’ve been talking about the weather in Sydney, Sydney, I’m going to, I would be happy to bet good money. It’s going to say the week ahead for Sydney looks like blah, blah, blah, whatever the, whatever the results it pulls back is.

And so that’s where you continuing a conversation is quite, quite helpful, right? Cause you’re just adding more context, adding, you know, adding more information in there. And so you can see it’s, it’s correctly said the week ahead for Sydney and give me a bunch of, bunch of weather, weather details. Right. So very, very useful. You know, you can go back and forth with it. Now, obviously this is a contrived example, but you can imagine if I was saying something like, um, I’ve been talking to Rob Hunter about a series of, um, a series of talks we’re doing or series of videos we’re but I’ve been talking to him on Teams.

I sent him some emails. Like, can you go pull that stuff out and summarize it? Like, what have I been talking to Rob? Like, oh, I need a high level picture ’cause I wanna go and talk to Joel about it. It’ll go through and dig through all my emails and it’ll find a bunch of stuff. I then thought, oh yeah, shoot. There’s that file in SharePoint that also has a bunch of info.

In my subsequent chat, I’d say, oh, hey, make sure you reference that SharePoint file too. Like here’s the link or here’s what it’s called and have it go get that new information and enrich the answer it’s already given me. And so that’s where you can continue that conversation. But again, if I then wanted to talk about, you know, something entirely different, how I was talking to Phil about it at a marketing event we were doing, new chat.

Start the new chat. It keeps it clear for Claude what you’re after and allows that conversational piece to flow quite nicely. Now, if you want to refer back to previous chats, they sit here under the chats and tasks section. So you can just drop that down. You’ll see all of your chats. I have some stuff in there that I can’t show, but that’s where that lives. And so that is like, that’s the basics of using Claude for that prompting stuff.

Now, if I wanted to add certain files, I can drop and drag them in, or I can tell them where they live and it’ll go and get them. Claude has the ability to touch files on your local machine, on your laptop or your desktop, as well as network drives that you have attached. So you could give it the link to some file in the Z drive and say, hey, we need to do something with this and it’ll go and pick it up.

That won’t be a problem at all. And so, yeah, that’s the basics of the Claude app. Now do keep in mind that if you start a chat here in the desktop app and then you’re on the browser somewhere else or on your phone somewhere else, all those chats are shared. So you can access and continue those conversations anywhere and you’ll just be able to keep kicking them off.

And so very, very powerful, very useful and also very easy. You can just jump in, answer the question, see how it goes. Don’t There’s no great, great, you know, crazy guide on exactly what prompts to do. Just try and be a little specific with it. Like if you want a Word doc, tell it you need it to be output as a Word doc. If you’re gonna give it a document to use as a reference, give it the document as a reference.

Don’t break up your conversation, act as if it was a colleague. If you came to me and said, “Hey, Kane, I need this particular information. You would just want to say like, hey, Cain, here’s the scope. I need these three things taken in. I need you to produce a document that’s about this long, make it a Word doc. Here’s the style guide for what that Word doc needs to have.

Give it to me and let me go away and do the work. Treat Claude the same way. And so, yet. Don’t be too concerned with having the magic prompt, the specific prompt, but you know, just think about what you might need to get that task done in the way that you need it done and give it that instruction and see how you go. Now, on the note of prompting, if you get some bad results and you’ve got to go back and forth with a little bit, once you’ve got a good result, ask Claude say, hey, look, it took a lot to get here.

It took five or six, you know, back and forth to get an answer that I needed here. How should I have asked that to begin with? And it’ll come back with some details on how you can improve that. So don’t hesitate to use Claude to improve your usage of Claude. It’s absolutely worth doing. All right, so that wraps it up for this for the Claw desktop app.

Get in, give it a crack, and let’s jump into the next topic.