Claude AI Beginner Masterclass: Build Smarter Workflows with AI

1. Welcome to the Claude AI Fundamentals Course!: Welcome to Cloud AI beginner Workflows course. If you have already used Cloud for asking questions or generating content, it's now time to realize what truly makes it powerful, and that is with building workflows that save you time and become more useful every time you use them. My name is Hosakihui and I've spent years working with AI tools to improve productivity, automate repetitive task. Build systems that are efficient for both personal and professional use. Throughout this class, I will show you how to use Cloud as more than just a chat bot. By turning it into an intelligent workspace, you can use clod to help organize your information, help you with some tasks that you don't know how to do. Or help you with the repetitive ones. By turning it into an intelligent workspace, you can use it to streamline your daily task, help you with organizing your information, and help you figure things out that you don't know how to do. We're first going to begin by understanding how Claude is different from all the other tools out there and what makes it unique. Then we're going to jump into the cloud projects and how you can keep the context for one task in one space and build upon it for long term work. Next, we're going to dive into skills, which is one of Claude's most powerful features. Going to learn how to build reusable instructions and workflows that will save you time and help you get more streamlined outcome for your needs. From there, we're going to learn about connectors, which is how Claude is going to use external tools and their information to give you a more streamlined and professional work finally, you will learn how to put all of these features into one business flow. You may learn about all these tools. They're great, but at the end of the day, you want to see how it applies to your needs. So we're going to see how we get to use each one efficiently, which one you should use from planning to execution. We're going to see how we get to connect all these points so that you get to continue using this after the course is done. By the end of this class, you will learn how to utilize each of these features and use Cloud in a way that is consistent, efficient, and quick. Let's go ahead and jump straight into clot. 2. Introduction & What You'll Build: Hello, and welcome to this clot beginner course. By the end of the course, you're not only going to be using clod occasionally, but using it daily to improve your task and even help you get things done faster. By default, most people only get to use about 10% of clot's capabilities. But the goal of this lesson is to close that gap and make you feel confident with all the stuff that's available right on the screen. So we're going to start from the beginning. Is AI? What is Claude? How does it all work and what you could do to better improve your experience. Then we're going to dive into the different tools that Claude has and finally close it with a project where we get to apply our theory and actually get something done. This course is good for freelancers, founders, students or people who just need a little bit of help with the stuff that they are juggling will not be mentioning or going over any of the coding capabilities that Cloud has just because that's a whole different realm, but I'm going to focus more on the other tools that don't get as much attention. You also want to set up your expectations before starting this course. Now, Cloud is a tool, not really magic. You still need a lot of human judgment to get a good result. So what Cloud can do is actually take your concept, whatever you feed into it, use its strong reasoning and learning capabilities to generate outputs for you. But you as a human need to still review that content, make sure it's correct, it's accurate, and it's exactly what you need. So don't just assume that you're going to press a button and something magically perfectly comes out by the end because whatever you decide to do with Cloud, there's going to be a lot of back and forth when it comes to finalizing that product. As you move on with the course, I do encourage you guys to have Claude open. You can follow along on the screen activity that I'll be doing or just do something straight after a lesson. But I encourage you guys to apply what you see onto your own account and see what Claude does for you. So that way, you're able to see Claude as a capable partner. You know, it's a new partner. It comes in. It doesn't know much about you. You need to give it direction, context, background, and that way, it's going to learn all your patterns, learn all the facts. Get better at giving you what you need. At first, when you tell it to make a presentation, it's going to make something very generic. But once you feed into it your brand's identity, the brand mission, the tone, the language, the audience, it's going to tailor that presentation into something that's actually of use for you. Now I hope you have an idea of what this course is going to be. Just to recap, I'm going to be going over all the fundamentals and basics before we actually dive into the tools and apply our knowledge into a project. I will not be going over the coding capabilities of Claude because I do want to shed some light on the tools that are not getting enough attention. Now that we have that idea, let's actually learn what makes Claude different. There's so many AIs out there. What makes this one stand out? 3. What Makes Claude Different?: Now you may have heard of Claude either on the news in passing or you've never heard about it at all. Claude is made by enthropic. They are mission focused on safe and genuinely useful AI. It's not just about pressing one button and getting something in return, but something that spends time learning your patterns and giving you useful outputs. It is known to be a little bit more careful, more wise, and a little bit more refined when it comes to its responses. It will leave you with the feeling of confidence rather than confusion, and these are just some things that the company is known for amongst its users. There are three areas where Claude really stands out. The first one is writing and editing large bodies of text. So it picks up on the tone of your company. As I said earlier, it picks up on the language that you're using, what method you used in that paragraph, how to improve it, how to make it shorter or longer, in a more efficient way. Second place is its reasoning around really complex problems. So if you give it a document where it's completely messed up, nothing makes sense, Claude is really good at working around that and giving you something that is more organized and makes more sense. Thirdly, speaking of files, it's really good with handling any sort of document. You could feed it a PDF, a link, a screenshot, an image, and it's going to thoroughly look at that document and integrate that knowledge into whatever prompt you're writing. So it actively engages with that content and stores that in its memory. And the memory is something I'm going to be talking a lot more about in the further lessons. But let's just say that once you upload a PDF onto your cloud chat bot, it's going to be basically sprayed over all your future conversations unless you stop. Every AI chatbot has models. So these are different versions that the AI tool interacts with you as a user. So Cloud has a family of models and not just one thing. You can think of these models as basically categories where each one is good for one thing, and then within those categories, there's different versions of that model. So we have high Qs, Sonnets, and OPS. So what are these exactly and how are they different from one another? First one high Q, it's used for really quick task. You know, maybe you're asking Claude about the best restaurant in the area. That does not use a lot of power. It just does something quick. It gives you a quick response. You can talk back with it, but it doesn't need that much details, that much reasoning, that much thinking. It does a simple search for you, based on maybe user reviews, based on available sites, and gives you a quick output. The second one Sonnets. This is basically used for everyday task. It's basically people's default model because it uses just enough reasoning for something to be outputted quickly, but it does not take forever and it's not really used for more complex reasoning. This could be maybe writing an email, writing a file, getting an improvement on that file, et cetera. The last model is the heavy duty one that's used for more complex situations, such as the example where you have a really big confusing problem and you don't know how to approach it. You feed that into Cloud and then you go back and forth with it to come up with a genuine solution. That third model, Opus uses the most amount of context. It uses a lot of credits. It takes a little bit longer, and basically Cloud gives you all three models, this whole family so that you can either choose from them or have it assigned to your conversation automatically. Right now, these three things, you don't need to memorize them. When you're starting out your chat, you could just work with whatever is default. Right now, you can see that I'm set on SontFive Hi. But if I go over here, we can see the high Queues, we can see the Opus, and there's actually a new one called fable. So this is still not fully integrated. I think they recently released this. But the three main ones that I just talked about, they're right here. If you go over here, we have more models, but you could just stick to the ones that are right here. Now, as you can see from the main screen that we're looking at, Claude is a lot more than just a chat bot. A lot of things on the sidebar, things that you may not have seen on Chat GPT, Gemini. We have projects, artifacts, scheduled, and even a customized feature where we get to make our own tool to better aid us in that project. We have a chatbot right here. This is your classic chatbot, cows, things that are going to help you work on whatever it is that you're doing. And then we have the coding feature. So right off the bat, it's a lot more than just a chatbot. And looking over here, you can see that there's different tabs for writing, learning, coding, live stuff, and Claude's choices. There's also some suggestions down here, and we even have a design tab that I'll be showing you further in the lesson. Now, the biggest mistake that you can make with Claude is just staying within the chat bot, you know, just going back and forth, talking about things not really utilizing all the different tools that are right here. Now, I know it may seem intimidating when you first look at it. You may think water artifacts. Why should I even waste my time with them? What if it ruins my project? But that's the whole purpose of this course. We're going to be tackling things that are unfamiliar to you and help you feel more comfortable in this new workspace. Now, I do want to remind you if you do not know this already, this tool, like all the other AI tools, is not perfect. So when it comes to real world facts, when it comes to, you know, news, research information, it still has the ability to make mistakes. And that's why it's so important for you as a human user to have your judgment and make sure that all the information that's coming out of here is accurate. So it may include a detail about something political, but it may be completely wrong. You as a user. You need to be careful about that when you're going to be publishing this information anywhere. So take all of the knowledge that Claude gives you, most importantly, when it has to do research with a grain of salt. Now, unless it's something that does not require any research, such as rewriting a document with a different tone, then you don't really need to worry about it. The concern comes when it has to do research. It has to fetch links from online, look at the market prices or something like that. That's where you need to double check that information. And another thing, Claude will not remember all the details across the chats unless you turn on that feature. So we're going to go over that later as well, but just know that it's not automatic. It will not remember every single detail about you unless you wanted to. Alright, so now we know what Cloud is, what makes it stand out? Let's actually go ahead and look at how we can make an account and how we can choose the correct plan. 4. Creating Your Account & Choosing the Right Plan: So now that we know what Cloud is and how it all works, let's actually see how we could get the program and how we can start by setting up our account. So just head over to cloud.ai. It's right here in the URL, and you're going to be seeing this page that you have to put in your email. You could use your Google, your outlook, or whatever you prefer. And then you're going to be asked to verify that email and that's all there is to it. To be brought to the main page, and that's going to be the same thing that we saw in the previous lesson. So I'm just going to put in my email, and as you can see down here, you can download the application for your computer. That's what I've been using so far and what I will be using basically the rest of the course. But you can also use Cloud in the browser itself. So I'm just going to log in real quick and then I'll be back so we can continue. Okay, so I just signed in, and as you can see, it's not different from the application, the desktop app at all. This is basically the same features, and we have everything we need right here. If you wanted to change things or view things about your account, you just simply click right here and you get to head over to settings. Here you can see your general information, your account, billing, usage, and all the other things that you need. And then you could just use this as you go on. Now, if you use the same email for the browser and the desktop app, everything's going to be synced. So you could use the browser for a different device, but have the application on your main device. You, you don't have to copy everything over. Now let's talk about the pricing and the plans. When you first sign in, for the first time, you're going to have the free plan. That's basically going to let you use the chat feature, but there's going to be some limitations regarding the models, the responses, and you may get to use some of the good models for a short period of time. Also some limits with attachments, and overall, if you want to use Cloud for something more detail, you would want to upgrade into a different plan. But the free plan is perfectly fine if you're just testing things out and you just want to see how it responds to very simple requests. So let's head over to our account and then go to subscriptions. So I am currently on the P plan. It's $20 a month billed monthly, and then 17 if you pay yearly. So you can see that we get all the free things. The additions are clot code. We have cowork available. We have the design tab. We have higher usage limits, and then we get the different models, as well as the ability for clot to carry memory over to other chats. You're using the free plan, Claude will not remember every detail if you jump from one chat to another. But if you have this either P or max, whichever you want, you can organize the chats for different parts of your needs, but Cloud is going to know every information that you fed into it. For example, it's going to know your job, but you could have different chats regarding different projects within your job. Can also get clawed for teams and enterprise. That's a different payment plan. But if you need, you can always come over here, see what's available, or just chat with one of their buying agents to see which plan is right for you. Which one did you buy? Should you stay on the free plan, the Pro plan, or the Max? That really comes down to how often you plan on using this tool and how much you will be using from it. So if you want to use it for just very simple chats across your mobile, your computer, you could stick with the free version. But just know that after a certain time, you're going to get either no responses or very weak responses because the models are going to be downgraded if you step out of that li. The Pro plan is where everyone usually ends up, and then maybe they choose to upgrade to MAX. But P lets you use all of the available tools, the cow, the design, and it also uses a lot of models. Now, the Max is going to be a higher level if you constantly need clot to help you with your coding. Really big projects, maybe one for, like, a campaign, and you constantly need Clot to be careful, very detailed, and help you throughout this bigger process, you could upgrade to MAX. But for the purposes of this course, I think you should start with P. And after the course, if you feel like you need more, then you can upgrade to MAX. And the good thing is that you can always cancel your subscriptions. So if you paid for multiple months, but you changed your mind, you can always cancel it with Clot. So you have your plan right now. I highly recommend you guys at least get the pro for this course just so that you're able to follow along with all the tools that we're going to be using. If you stick to the free version, you will not see a lot of these options, a lot of the outputs that we'll be going over. So it's only $20. I hope you guys are able to. So just go ahead and make that purchase, upgrade your account to the P plan, and we can continue learning more about Claude. Mm 5. Navigating the Claude Interface : So now let's take a look at the interface. I'm going to jump back to the desktop application and just see where everything lies and how you can get started with some simple task. So on the left side, you can think of this as the homebase. This is where most of the tools are kind of combined together, and you can get started with the type of task that you want to do. So the first thing is the bar, you can collapse it. You can search through your different chats, your different projects, and look for things. You can start a new chat right here or with Command N. When you click on that, you're going to get your classic chat bot. You can write a simple prompt, attach things, either files or photos or a project. I will be talking about projects later. But as you can see, it's a lot more than all the other chatbots have to offer. We have something called skills, connectors, plugins, and there's also a research feature. By default, it's on the websearch meaning that it's going to look at very accessible links on the Internet. So maybe Wikipedia, Yelp, Instagram, that sort of places. You can have certain tags be attached to your prompt. So if you do write, it's going to focus more on tone, language, the quality of writing. And here are also some templates for you to get started. And as I hover over these, you can see that if I click on one of these, it's going to write me one of the prompts, and I'm telling it exactly what to give me, and it's also going to ask me one or two key questions. Now, this is something I have not seen in the other chat bots where it kind of quizzes you to get a better idea about what you want. So this is a very interesting thing that Claude does. It kind of has a quick aop quiz, in a sense, before it actually gives you the output. That's something that it made for me right here in my prompt. I asked for a case study, but it's going to ask me to basically choose the correct answer or the answer that I want before it even gives me any sort of output. What are the case studies for? Let's say marketing. Do I have any source material? Let's say I don't. Once it has these answers, it's going to combine your prompt and these answers together to give you something decent. Right now, interestingly, it's going to ask more questions. So it's not giving us random outputs, random paragraph, just for the sake of giving us a result. It's actually taking its time. So say just a few random things. It's going to think, and I'm using Sonnet right now. So that's the everyday level model. It's going to work through this. It may take longer because I'm not on high Q, the faster one, but this is just an example of a chat that is tailored towards writing specifically. Now let's go back to New chat. The chat that I just exited is actually right over here. So while this is doing its thinking, I can go ahead and do another chat. I could do a learn one. So maybe I'm a teacher. I want reflection exercises. I could make some with Claude, right over here. We can also do some coding if you want, and then Claude's choice. So these are like, some different things that you could do randomly with Claude. It's pretty fun. So that's for attachments. These are for different types of chats. Here is the model button where you get to choose your model and basically tailor the response type to your need. We also have a voice feature, so you get to talk directly into the microphone and then get a result based on that. And then we can do a voice mode, which is just no typing at all. So if you're on your phone and you just want to talk to Cloud, you can use the voice mode. That's the chat. This is the project. This is where you make your new project, edit them, delete them, search for them. These are the artifacts. So these are products that you get to build based on an idea. This is, again, all the editing, building, deleting. And then we have a customized tab. So this is where you get to integrate different tools that you have outside of Cloud, bring it into Cloud and get kind of like an agent, in a sense. You can add plugins, connect apps, and create new skills. Skills is something I'll talk about later. But this all goes to show how much more professional and handy Cloud can be. So these are all the chats here. This is the design that I told you. You get to build visuals with Claude. These are some examples. This is your account, billing and all that stuff. This is chat. If you go to cow, it's going to remove that chat option and you get to work with all the other tools. You can make automated task, you can make agents, projects, and even scheduled task. So maybe if today's Tuesday, you can make a task so that it sends out an email every Friday to X YNZ. So that's an example. And then we have the coding section. This is again for the artifacts. If you want to make products out of an idea, you have sessions, you have messages, and there's basically a chat bot down there where you get to see a task that you want coded or just go back and forth with Claude to brainstorm. And you can see that we have the models down here as well. We can do attachments, do either local or connect your Github, if you want, and do a bunch of other things. Now, as I said, this whole coding section is an entirely different story, so I'm going to steer clear of this for this course and just show you guys the chat and the cow. On the top right, we go back to chat. This is the incogNohat, so it's not going to save anything you say here. And be careful with this because if you want clot to remember certain things for memory, you do not want to be in that page. And that's pretty much the interface. So right now it's just a bunch of tools that you may or may not know about. We're going to dive deeper into them. But right now, at least you know where everything is and where you can access certain parts of Clot. 6. Understanding Tokens, Context Windows & Projects: So you may have come across tokens or credits when you used clot. And if you're on a free plan where there is less usage for you to reach the bar, you may have seen a red error saying that your credits are going to reset at a certain time or that you can no longer chat in that session. So when we're looking at the app right now, the only clue about how much usage we have is when you choose a certain model. So right now, I'm on SontFive. You can see that I have two times more usage until this date. But other than that, there's no progress bar over here, and the only way for you to see how much more you have left your plan is to go down to your name settings and then go into usage. So this is settings right now. If you go right below billing, you'll see this. Click on it once, and you can see what your current session is looking like. Now, I'm on the P plan as we established. But if I use maybe the more advanced models such as Fable which is the newest model or the Opus one, this bar is going to fill up a lot more quickly. So what is the purpose of this tab? It basically tells you how much more you can use Cloud with the most precise capability. And the more complex model you use, for example, Opus or the newly released fable, the quicker this bar is going to fill up. But if you use IQ for really quick stuff, it will take a while for this to fill up. Now what happens when you've reached the end? There are certain things that can happen one you will just not be able to chat in that session, or if you are able to, it's very limited. So maybe you will be limited to, like, more basic models rather than the advanced one so that's one downside. Another one is that you're just not able to use it completely. And if you are on the free plan, you fill up your usage a lot more quick. So that's something to bear in mind, but we're on P, and we're not going to have that problem. So current session, whatever you're currently on, starts when a message is sent, and whenever you send something that is larger than a message, for example, a document, a photo, it's going to take up more of this bar. Now you have the models here. As we saw, there's different models. They all reset around this time. So you can kind of come back here and see how much usage you have just so that you are sure you're getting the best out of Cloud. Now, fable, as I mentioned, is something newly released. Currently, they're doing promotions where if you have the P plan or something high you get to use it within your plan, and it uses 50% of your daily usage. So that's why there's two separate ones. This is all the other models, and this is Fable alone. Now, fable is above Opus, so it does take a while for it to process your prompt and you really want to use it for something more complex. Another thing you could do is to turn this button on, which means that whatever you're using that's above your limit will be charged separately. So you're no longer being charged for a monthly thing, but also for these additional credits. That's what this is for. Can turn it on if you need it, but if you're doing something basic within Cloud, you do not need to turn this on. Now, we're seeing a few keywords here. We're seeing credits. You may have heard of tokens, and I'm just going to go over what those exactly mean and why it's relevant to you as a Cloud user. So tokens are small chunks of texts that Cloud is going to process when you're putting in your so depending on how big your prompt is, your token size will increase as well. And just for reference, 100 tokens is about 75 English words. This is not something you need to memorize, but if you ever heard tokens, that is what it refers to. It's just about how Cloud reads your prompt, your English prompt, or any other language prompt and then processes that for that action. Now, there's also something very important in Cloud and all the other chatbots, and that is Context window. So if you were to put something very generic within a tool like clot, it will give you something generic. If you were to give it more context in your prompt, it will actually look at those and include them into that solution. Now we're going to learn more about prompting and giving context constraints in further lessons, but just know that context windows are very important because it does relate to how much processing time Cloud will need to give you that response to your this is also important if you use one session more than necessary. So say you had a session, so just for reference, this is a session right now. If I were to go in this session and use it for years, at some point, the stuff that I've said earlier in the beginning of this session is going to fall out of context, meaning that Cloud's memory per session is not infinite, and this also changes depending on the plan that you have. So you want to make sure that you're storing only the most important and necessary information in memory so that Claude can use the context in your session to respond. Don't want to remember every single detail as you're using that session. If you want to work on different things that do need that memory, that context, for example, if you are having a session about your educational path, there's going to be a lot of details personal to you, like the degree you want to follow, the end goal that you want to have, the university you've applied to all these information, there's a lot of information cannot kind of have that box of information mixed up with your work box. So for your work, you may be working on different projects about certain campaigns, and there's a lot of information that you may want Claude to remember. Good thing about this specific tool, Claude is that it has something called project, meaning that instead of having one session named to do list and have every single thing in your life in it, you can separate those into projects. One project for your education, one for your work, one for your daily life. Now, projects are right over here, and this way, you're able to have one dedicated space for one purpose, and you're able to have that context window constraint. To what is needed and it does not mix with anything else that you're using clot for. Here are some examples, study project, career project, research project. In the end, you as the person, you have certain demographics, you are using it for all these different purposes. Clot is going to know you as much as you let and it's going to use as much as information needed to focus on these different areas. And it will not mix the following. Wess if I had all three topics in one session, the overlap can happen a lot quicker. Now, we're going to see this in action in further lessons. So if it sounds confusing right now, don't worry about it. So just to wrap this lesson up tokens, you can think of them as your units, contacts windows as the capacity for that session, and projects as a way to have different rooms with their own capacities for certain things, avoiding overlap between memory and the details that you want to keep separate. Now that we know a little bit about, you know, how this whole AI thing works, what our tokens, where our money is going, actually, we can now move on to the next lesson where we learn a little bit more about prompting and how you can best save those tokens by writing prompts in the smarter way. Mm 7. Writing Better Prompts: So you have probably heard about prompts, especially in this day and age. So prompts, you can think of them as an instruction that you give an AI tool to either do for you, give you tips on, or anything else that you need AI for. So a very simple instruction could be, Hey, could you write me a short story about rabbits? That is a that is an instruction. Now, an instruction could be very specific, could be very generic, but that does fully change the way you get that outcome. So if I ask for a short story about rabbits, I'm going to get I could get any rabbit. It could be anywhere. It's something out of my control because I did not talk about it. I did not say a blue rabbit. I didn't say an elderly rabbit. So just like how you would be careful with certain instructions, you would do the same with prompts. Now, we are here on the homepage, and I just clicked on a new task. See that there's cow and then there's chat. I'm going to switch to chat right now because we're just going to see how prompting works in general. So you want to give an instruction, as I said before. Let's do a very simple one. Write me a caption. So it's not even a really long sentence. I'm going to click Inter and let's see what we get. So this is the general format of a prompt. Do this for me. Now, you can see that immediately Claude is going to ask for more context. This is automatic. I did not really say anything prompt, but a little quiz pops up, and I need to now use Claude's little test to make my prompt more detailed. So let's say it's for food and it's going to be funny. So we get the Q&A. Now without me having to say that in my prompt, it is writing me some funny food related captions. So we get seven options, and then it's asking me if I want to refine any of these options. So I could say, refine number six. Again, another instruction that I'm giving Claude, and I'm just following up whatever it's asking me. So it took number six, and it made them more refined. And I could ask for a certain tone. So that's pretty much what a prompt is, but of course, you can make it a lot more detailed in order to get that final result quicker. So if I were to jump from here to here, I would need to write something a little different. So write me a caption. And then we said we want it to be funny slash witty. So we put that right here, caption about what's the caption for food, and we can add the refine element with a refined pike. So I kind of combined all of Claude's responses, and now I'm going to click Sent. I'm using the default model here, Sonnet five, but you can switch to anything else. So there we go. Now, instead of Claude having to ask me for clarification, I just went from one prompt to, let's say, the best caption for my post. You can be even more detailed here. So we're just going to take this. You can use the copy button up here or edit this to change the original prompt. If you click on the Retry button, it's going to look at that prompt again and give you another set of response. So I'm just going to copy this here, paste it in the bar, and we're going to maybe say something more about the food, maybe about a certain location and see what Claude is going to do with that. So write me a funny caption about pizza. With a refined take, period, let's say, look into the current best pizza spots in New York. So that's something that Claude needs to do research on. I did not really give it much information. I'm just asking for something very detailed. So immediately, you can see that it's doing a little research. From those information online, it found, you know, some information it could use. So here are the restaurants, the pizza spots, and it's listing them down. And then it's asking me if I want to take one of these to further refine my caption. So we pizza caption, a couple more options that we could use. So let's, you know, make this even more detailed. Let's edit our top prompt. So look into instead of this, we're just going to put that spot that we need. It's safe, and as you can see, got rid of that previous response, and it's going to give us something else. Searching the web thinking at the same time because we are using Sonnet, and here are some new captions. You can go even further than this, maybe write this caption for an audience of mostly teenagers or mostly kids. You can keep adding onto this instruction. And I want you guys to think about this as a little brief that you're giving Claude. Now, a brief may come with some images as well. So you may want Claude to write you a caption based on a photo that you took while you were walking in New York, and that's another thing that it can do. So if you wanted to add a picture, you click on the Plus, add files or photos if there's, like, a that you already have and you want it to be more refined, more detailed, you can add it right here as well. But right now, I just want you guys to know what a prompt is and how you could add on to a prompt to get quick and more relevant responses. So we went from this a very generic thing. It could be about anything into this prompt where I'm asking about pizza in Brooklyn, and I'm asking for something that is funny with a refined take. So go ahead and practice with prompting, start with something very basic and see how you could build up to a more detailed prompt that give you exactly what you need. The next lesson, we're going to continue with the prompting basics and learn about giving context. 8. Giving Claude Context: Mm hmm When you're talking to a friend about an important project or an important incident, you give a lot of context. That way, your friend can better understand what you went through and better help you or offer you some advice. And that's the exact same logic you need to have with Claude. We talked about giving generic prompts and how the result is going to be generic, as well. But in this lesson and the following after, we're going to focus on how we can build stronger prompts and what your goal should be every time. So I'm back in that same session from the previous lesson, and I was basically talking about a caption for a pizza place in Brooklyn. So what we're going to do is take a simple sentence like this and turn it into a multi part prompt. You don't have to do this for every single prompt in cloud, but if you're actually looking for a result that you want to use for your work or something more detailed, you may want to consider this technique. So let's set a goal for our lesson. I want to do, let's say, a LinkedIn post about my experience and my knowledge about local pizza shops in Brooklyn. Instead of being like, what are some good stores or pizza places in Brooklyn, I could turn this into a multi part prompt. But first, I'm going to type this in so you can compare the result we get from this versus the result we would get later on. I'm just going to hit sent, and this is where it's going to do a little research, give you some results based on the location. And then afterwards, once it finishes searching, we can do write me a LinkedIn post about what makes a good pizza spot in Brooklyn. So here it just gave me a visual display because I didn't say how lot should output this. Did I want a bullet list? Did I want a paragraph, one sentence, one word. But essentially, I just asked this question, and it did give me some highly rated pizza spots in Brooklyn. And when you scroll down, there's a few bullet points with some keyterms and it's asking me a follow up. But now I'm going to ask for an output, but I'm not going to say what tone, what sort of length, and let's just put that in, see what we get. Here the sentence that I put in the prompt, it is lacking what's called context. So any sort of background information that I would need to give Claude, give me exactly the result I need. Here's a LinkedIn post. It's not bad. It's very clearly AI, if you know what I mean, I just has bullet points, things that people already know. Maybe important tips that are available on other Google search results. But there is no context in there. I didn't say, Hey, I'm a food influencer, maybe I'm a food critic or my audience, what sort of tone clot needs to use. That's what we're going to do. So I'm going to have this as our first point of reference. You know, it's not a bad LinkedIn post, but it could obviously be a lot better. So we're going to do something like this. So let's do, like, task. This is what Claude is meant to do. On text. This is Things About Me. Audience. Who is this post for? Output, and I'll do one more here for research. So these are five things that I'm going to include in my prompt for this specific task. Now, if you're doing something else like building a landing page or being like an app, that's going to be different. First thing we're going to start with is task. First, we're going to assign a task to Claude, and I want you guys to turn away from the generic write me or me very basic command and instead do something like a world famous food critic in Brooklyn. Write weekly LinkedIn, write weekly LinkedIn posts about best pizza spots in town, can the time include the local experience and highlights. Simple stores, something like that. So now it's no longer clawed but a food critic. One that is based in Brooklyn. The task is you're going to write Din posts. The task. So it's not going to write blog it's not going to do captions. I'm asking for one type of text. For context. Let's say I am a food influencer, and I like to write a lot. A new local spots I find in Brooklyn need content that is interesting. Fun to read, but also very informative. I need to post or I pose twice a week on Linktn. N a lot of engagements. My audience consists of other food influencers. Calls for actually, let's say, tourists visiting Brooklyn for the first time. All right. So that's what it is. If you have an age range, gender, anything that's relevant to what you need, you can put it here. I'm going to put that because tourists could be kids, adults, but mostly people who would even have access to LinkedIn. Research, pull data from this month Data from this month, look at Jing spots on Google Maps. You could also do Yelp or wherever you are. By that, the places still open. And offer at least some vegetarian options, but this needs to be say a 500 word post with points for an easy read. Mojis to grab attention. A fun but pure tone and have a call to action at the end of the post for Dance to follow me for more content. And we're gonna do one post. We're gonna do I three. We're going to write three Linked in posts to choose from. They include a different angle. Or age of the posts. So this already is very different from where we started from. I'm now going to submit this. I'm going to use Sonnet five. You could use a different model if you prefer. You can just use Sonnet so that you're having some tweet, not something too quick, not something too out. And we're just gonna hit Send. It's going to think, it's going to research. It's going to apply our command to the final product. We can see that it's reason things right now, and we're going to dig deeper into researching in a further lesson. But we're just going to see what we would get from something that has a lot more contacts. Finish the three different type of articles. I have them right here for, like, a quick of them, and then the file itself is right here on it. If you can't see this, just click on it once and it will open it right here. See option one, hidden gem angle. So that's one angle I could take Craft. So if I want to look into the historical context of the area and then for trending or big name places, listing me the different pizza spots, what's unique about them and how it relates to this angle. Click over here, we have our options written out. So this is my title. Mojis I did use bullet points, and it included the vegetarian that I had. So that's option one. Um, option two is right here. Again, same thing. Vegetarian, emojis, bullet points. And that is consistent across all three particles. It's just a starting point. Just by putting in a little bit more time for your prompts, you can get results like this. Now, we are going to talk more about refining. You know, you get this, but maybe it has something you don't need or something you want changed. That's going to be something we will focus on in a further lesson. But Jaire we started from, you know, apps visual, this very plain post, and something a lot more unique. Li the next lesson and see how we can add constraints to our prompts. 9. Using Constraints & Refinement : Alright, so we stopped here in the previous lesson. We got a bunch of LinkedIn posts, and we just want to refine this. So let's say I read through this and I want to leave out some details. And that's something you could do within the same session and even edit the same document. So before I go out from here, when you have a document, you can view it like this. You can see the code. Say you want to put this on a website or just take a look at it like this. But you can save this as a Microsoft document, printed as a PDF or publish the artifact through one of your connectors. So say you connected LinkedIn to Cloud, it's going to upload it for you automatically, but that is something we'll get into later. You can refresh this. So say I made some changes, but I'm not seeing it here. That's what this is for. You can expand this to a full screen and just close it like this. Now, this file will be saved in your clot, so don't worry if you don't see it on the side, you just click on it once and it comes back. So let's talk about some constraints. Traints are things you want Claude to steer clear from. So that could be certain neighborhoods in Brooklyn that I want to leave out. I don't want Claude to focus on that at all. It could be certain menu items that I don't want Claude to talk about. It could be not including vegan pizza spots, those sort of things that I don't want are going to be my constraints. The same thing as a negative prompt, basically things you do not want Claude to even consider. What we're going to do is work with the same article, and I could see that it's looking at certain neighborhoods and it's giving me different dates. Let us think of some constraints here for our case. I could have Claude not mention a certain neighborhood. As this one. And given that it gave me a bunch of dates, I could have it leave out restaurants that were opened recently, so we could do that. So it's even asking me a follow up question, but I could instead do just add constraints. So work on the same document, but implement the following. In the updated version. So you could do bullet points here just to shift eight for the little asterix and it becomes a bullet point. Do not include let me see, implement the following constraints, actually. So do not include spots in Bushwick. Do not include spots that were recently. So maybe nothing past 2010. Do not include the word. Let's even try a word here. I'm going to see what's being mentioned a lot. So let's say I changed my mind about the vegetarian thing, do not include the vegetarian options. So three constraints, same document. Let's send that and see what happens. So, again, it's looking at the different pizza spots at the same time. So it has, like, ten results that it's looking at, and now it's basically running the command, and it's working on that document. Alright, so now it gave me a different article. So we have three different spots, and I can see the dates. They are indeed before 2010, and no vegetarian mentions. Comparing it with the stuff above, we can see or actually, that's the different pizza spots. So now we have different dates. It's giving, like, a historical context. The address is not Bushwick, and I don't see that vegetarian section in the post. Still have the emojis. We still have the three options that I asked for, and it's looking pretty good. Now, it did give us a different link. So when I press on this, this is the document we have, but it's the same thing as the document above. So it did edit the original. You could leave that out, so you have two types of documents to look at, but I just wanted it to edit the original. And there are 500 words, and now we're just going to add more constraints. So let's do another one. Edit the same document, but remove all Imogs and reduce the length to, let's say, 350 words, only keep option three. Now we're going to take that bigger document, shrink it down into one option, one post. So it's going to finalize it. I didn't ask it to do more research, so hence not seeing the results like we did earlier. So we got the same document, and there is no emojis. There is no option one, two, three, and it's about 3:20 words. Even though I asked for 350, it's around 320. You can see Claude kind of has its own reasoning over here. Say, this is my maximum, I could mention it there, but 320 works for me as well. Has the hashtags. You can ask it to add hashtags, remove them completely, research what hashtags actually work in this case, and replace it in there. But I'm just going to keep what's here already. So that's just one example. Instead of asking Claude to write you a LinkedIn post, you can have it do research, have a different tone, have a different format, an output, and even a different role. So if I did not tell Claude to pretend to be a food critic, we may have gotten a different result. Maybe the tone and the description of the pizzas were not that engaging. The way a food influencer talks or reads is very different from a regular person. You know, maybe they care a lot about ingredients, about the historical context of the restaurant and other things. So this formula that we had earlier, you could use this for your next prompt and just see what sort of different results you're going to get. So practice with this, you know, you can target Instagram captions next, try something you're really passionate about and you know a lot about and see how Claude can help become your assistant. 10. Prompting Best Practices & Common Mistakes: Now that we have a better idea about how to prompt Claude and how the general application works, let's talk about some tips to improve your experience and some mistakes that you might make and how you can avoid them. So first, let's talk about some beginner mistakes and how you can avoid them as you move forward. The first mistake that a lot of beginners use is treating Claude like a Google search. So find me the best restaurants. How do I become famous? How do I become? The questions will give you a response, but you're not really using Cloud for its actual capabilities. So instead of just asking very generic questions or giving very generic task, try to add a lot of contexts and break down your prompt into something like this. It may seem like a lot and may be useless to some of you, but this actually saves memory and also provides you more space in that session. So you're not wasting a lot of the capacity in this session with random prompts that didn't even give you what you were looking and you can just skip all that that we did in the beginning, get started with a template like this and save yourself memory tokens and just time. So that's the first mistake. The second mistake is not verifying what Claude gives you. Now, a lot of people that use AI assume that because it's trained on the Internet and all the resources out there, then it must have a good understanding of what is factual. And what is true. But you as the user need to verify every single information that comes out of Cloud, especially when it has to do with the research aspect, when it has to do with political events, you know, stock prices, that sort of stuff, because Claude, like any AI out there can make mistakes. So in this example, you know, I asked about good pizza spots in Brooklyn. But what I can do with this document is ask Claude to verify it again. Now, this may seem redundant, but it actually helps Clause to go back to what it gathered and see if it made any mistakes. So can you verify the information in the document? You know, I don't want to have to copy each of these guys. Double check myself that if it's open, if it serves the pizza that I'm looking for, I can just have Klatt do it for me. Now, a lot of times, it really depends on what you asked in your prompt. Klatt does a good job of, you know, processing the information in a better way and making sure that it's not missing out any information or presenting you with false information. But since we're dealing with pizza here, I did ask it to verify this information, and it's actually telling me where it got the information from. Now, if I'm the user and I say that Wikipedia is not to be trusted, I can now tell Claude to remove whatever you grabbed from Wikipedia outside of this document. So confirmed accurate, and it's even flagging some things that it doesn't have enough information in. So now we can adjust our document or just keep going with what we had here. But again, this is about pizza and food, but if you're doing something with, you know, market research, if you're talking about, like, a different culture, you want to ask a very simple follow up. Can you verify this information? Times you may see Claude being like, oops, I missed out this information. Thanks for catching that. And it will redo that research for you. So mistake number two, assuming Claude is perfect, it can know everything and give you the precise information. That is not correct. The third mistake is trusting Claude too much with your personal information. Now, Claude does do a lot about information security, but just as a good practice with any AI tool, Cloud included, do not provide any sort of personal information that can be used against you. Now, it's not like Cloud is going to, you know, write you an email and blackmail you, but say there is a data breach at some point, hopefully not, at least you were safe and you didn't provide them with your home address, your financial information, personal information about your family, and those stuff. Now, I'm not saying don't say anything about yourself because the whole point of giving context is to provide some information about you, but you have to know the difference between safe information and unsafe ones. So safe information could be your name. Job you work at currently, the area that you're in, your academic or career ambitions, what you like, your hobbies, these sort of stuff, that can be considered public information. So make sure you are aware of that. If you're uploading a document that includes your personal information, make sure you're blocking those out or taking a screenshot that does not include those vulnerable information. Sake number four, which is something we're going to get into in a further lesson is not taking advantage of the memory that Claude has. Now, a lot of beginners will assume that every session sort of resets, go put in all the context again, and they keep repeating it over and over. So say my context is that, I'm an artist and I like to work with oil pastels. I want to work in New York, and I keep copy pasting that in every session because I'm assuming Claude is not going to memorize that. But if you put it in once, you can lock that in Claude's memory. And then if you make another session, Claude is going to remember those details about. There is a better way of utilizing memory and basically allocating a space within Cloud for one task, and that is the project section. We're going to get into this later, but a lot of beginners leave out the project capability and they don't realize that Claude is saving everything in memory, and that is going to be universal within your account. Next time you want to talk about your art aspirations, you don't need to reintroduce yourself, your dreams, your goals because Claude has already saved that for you. Now, let's jump into some tips that you can practice moving forward after this course, during the course so that you have a better experience. While we are on the subject of memory, if you want Claude to remove something from its memory after you're done using it, you can mention it directly. So let's do a little exercise together. I have my session open. I'm now going to say something for it to save in memory. Save this in memory. I like oil pastels. So it's going to do a little bit of processing here. So you can see that I got this error. So this is actually something you need to turn on before you proceed. And this is something a lot of beginners miss out on. So simply go to your name down here settings or just do memory in the search bar. And simply turn this on. Allow Clot to search for relevant details in the past chats. So you just turn this on and that's it. Now, another cool thing is that if you were used to using HGPT for the entire time and you just discovered clot, you can kind of export that information from HGPT import it into clot. You don't need to redo all your prompting and all your projects in this new language model. So that's this option right here. If you do start Import, you're going to copy this prompt into whatever tool you were using beforehand. So hat GBT Gemini, copy it, and then you paste it in that model. It's going to basically give out everything it knows about you, and you're just going to copy the result from there, put it here, and that's going to be added to memory. So you don't have to start over every time you are moving from one tool to cloud. So now, whatever I say will be remembered across chat. Similarly, if you said something that you don't want claw to remember anymore, you can have it remove that detail from memory. So all you have to say is, say I said something, it would be like, remove that detail from memory or maybe put the detail in quotation, and it's going to remove it from the chats memory. So that's a tip that you can use if by accident you said something or something you don't want Clate to remember. The next tip is to actually have Claude help you figure out what it is that you need. So say I'm an art student and I want to write a little personal blog, but I'm not sure what sort of content I need in order to go viral or in order to get a lot of audience. So say something like this, I am an art student who wants to start a new personal blog and wants to go viral. That's something I want to do. My goal is that ask me questions that would help you generate the right content for me. So I'm having Claude do a little pop quiz with me so that it can better help me. So there's a little pop quiz. What kind of art do I want to make? I'll say traditional. I'll do essays, and maybe polished. So I answered these according to what I know about myself, and now it's going to apply those into its strategy. So now I have something a little bit more polished. I don't have to think about writing a prompt because this combines with this, and it basically sets clot in the right direction. Tip number three is having clot two multiple outputs or receive multiple inputs. Now, you may have seen that you can easily add files or photos. But if you don't know how to do a very good prompt or basically articulate what you're looking for, you can upload multiple pictures, your reference pictures or multiple reference files, and have Cloud look at those. For example, if I want to design a new menu for my restaurant, but I don't exactly know how to articulate the vision that I have, I could upload some of my favorite restaurants menus and have Claude pull out the keywords that will really help me or, like, how I should like word my menu items, something like. Or it could be something within your sphere. Say you were walking around the neighborhood. You took some pictures of certain architecture you like. You upload those into Cloud, have it pull out some keywords that you guys can further discuss so that you can do a better job on your future project. So basically the whole idea of using the inputs as much as you we did this already in our previous lesson, but you can also multiply your outputs so that you're not limited to one result and one file. You can have Claude basically process something and output something for Instagram, LinkedIn or Tik Tok, and maybe within those different tools, you can get different variations. The variations for Instagram, three for Tik Tok, three for LinkedIn. That will basically save you time. You don't have to go back and forth with Cloud. I will do it once and you can spend more time thinking about which one you like. The final tip that you can utilize is using the voice feature right here. Think about it as being on the phone with a friend. Sometimes you just want to bounce some ideas. You don't want to type forever. Kind of turn this on and it's perfect on the mobile because it becomes like a telephone call, you're just talking to Claude, like, Hey, I got this idea about an ice cream store. I don't know where to begin. Claude is going to talk to you, like a little friend, ask you questions, what are your inspirations? And you would basically save time from typing, pulling in links, and just basically use what's in your mind and try to articulate it via a phone call. The good thing is that even though you're using the voice mode, it's going to be transcribed within the session. So you can always come back to it and see what you guys talked about, build upon it the next day with a new idea that you have or something you want to further refine. So those were some tips and mistakes that you could keep in mind moving forward with the course, and I hope you guys were able to identify how those are helpful for you at this stage. Moving forward, we're going to dive deeper into some of the tools that Claude has, and you're going to see more of what I just said in practice. Let's actually move into one of the really cool capabilities that Claude has, and that is research. 11. Research Faster with Claude: Now, let's dive into the research feature that Claude has. Now, research is going to be a little bit more advanced than what we've been doing so far because it's going to look more carefully at the sources it's using and give you a lot more details compared to what we've been doing in the chat. So to enable research in the app, you just have to go to the plus and go all the way down to research. So you can see that we were on web search this whole time. Research is going to be slightly different. We're going to see this little bar. It's a little icon next to our chat, and that's a good indicator that you're in the right place. I'm going to leave the model as it is, feel free to change it. But we're basically going to have Claude research good venue for a birthday party that I'm planning for a 12-year-old. So we're going to do a very simple prompt. I'm going to say, give it some context. I have a birthday coming up for, let's say, my 12-year-old. And I'm going to turn this on, actually, just to make sure it's not picking up anything about me, but sort of starting from scratch. So with the Incognito mode, it's not going to look at the other sessions that we did in the previous lessons. It's only going to use the context that I'm giving it right now. So I have a birthday coming up for my 12-year-old. I'll say Son and I need to find a venue that is kid friendly and offers catering. So that's something I need. So this is not going to be enough because we didn't really give context about what Claude needs to research. So we need to give it, like, a certain area, a certain platform to search for maybe finding out what 12 year olds like, you know, that sort of thing. So, oops, let me actually turn on research here. So I'm going to give it some rules first. The formatting isn't really as important, but when you make your prompt clean like this, it's easier for it to be understood, and I'm also doing it like this so you guys can easily follow along. Rules should be outdoors should be suitable for large crowds. We could say food options need to be ins gluten free and offer lots of veggies. Just making up rules here. I think I'll leave it like that. Look into Brooklyn. I'm just gonna use Brooklyn again and find me places online that I can call and make arrangements. My budget is, I guess I'll just put $10,000 just randomly. I want to host this birthday during the weekends. So research about the places, list them and tell me their price and their pros and cons. And then in the end, I could be, like, asking more questions that can better help me plan for this birthday. Say this is something I've never done before. Claude may think that I'm missing out on drinks or an element that I have not mentioned. Maybe all I'm thinking about is just affordability, catering, it being outdoors, but Claude may suggest that I also look into photographers, clowns, that sort of stuff. So you can have this line so that it helps you out with the planning. I'm just going to send this, make sure research is turned on. Once I'm done, I'm just going to hit Send and let Claude do its thing. You can see that it's going to get more details from me. I invited that, but sometimes if you don't give enough context in your prompt, Claude will automatically ask that from you. So how many guests am I expecting? Let's say, around 50 outdoor venue, we could do a garden how far ahead, maybe within two weeks. So more details, it's going to dive in and research to get us those venues. 28 sources and counting. It's also telling us how long how much time has elapsed since I've given it these answers. Now, I had to go in incognitomo just so that you guys can see Clock can pick up the context, but say you were doing this for an actual event, I recommend that you do it in a session and not in this form of chat just because you can always go back to it, build upon your prompts. Maybe you discovered something about one of the guests. Say someone said that an you can add that detail to that chat, get updated response, a different catering option, and that sort of thing. So just bear in mind what this does to your sessions. It's only about it remembering these details. So when I exit here, it's not really going to remember this birthday party planning because I asked it to. When you click on this bar right here, we can see that it has been 2 minutes so far, 162 sources and counting. But when you click on this, you get into this side panel where you get more details about the process. So it created a research plan, and now it's gathering the sources. It's getting multiple sources from different websites, and now it's looking more into the 41 sources that are probably more relevant to my ask. We're just going to let it do its thing. If you ask for a more simpler research, it will take less time, but I wanted my command to be very specific, just so we can see how detailed clot can be together. So it finished. It took about 6 minutes and 17 seconds. Again, if you have a simple research, it will take less, but for something more complex, it will take more. So I'm just going to close this and show you the process it did. It looked over 239 sources. It took this many minutes, and it's just telling me that, you know, a report of what it did. And then it gives me a document. Once you click on the document, you get a nice header, and then it's telling me the different places it found, how much they need, pros and cons, key findings, different details on each of those spots. I have two, three, four, five, six, six places to choose from. Recommended caterers for gluten free and vegetable forward. We've got four of those and what's unique about them, so zero waste. Then it's giving me an action so today I should call this park and ask about certain things. Next day, I could make some inquiries about picnic house and a bunch of other stuff. There's also some do not do like this guy, and some things I should bear in mind for my birthday. So pricing and then a little risk about the timing of this. I have this document. Again, I could go here and be like, fix this, look into this place, maybe add another allergy in there. So maybe half of my guests are allergic to strawberries, making sure that this goes in my plan, and when I'm looking for caterers, I inform them that there's also that allergy. So you can copy this, or you could download it as a PDF, as a markdown or publish it. Web. So if I want to share this with other people who want to help with the planning of this birthday party, I can. And if we have a team within Cloud, say that we're all on one plan, if I click on this option, they will all get a copy of this document. Now, the research feature can be used for a lot of different things. This was just one application, but say you want something for business related works. For example, your team is thinking of coming up with a new drink. You can use Clot to research what's popular right now, what is everyone's favorite flavor? What is the market like? And because you're on the research mode, you can see that it does take a lot more time. It is more detailed, and it does give you a better result compared to the web search. So these are straight out of credible sources. That's something that the research will make sure happens. So if you click on this, you can see all the different sources it went to. And as you can see, they are quite relevant to birthday parties in New York. So I could see all of these guys, different results. I could click on each one if I wanted to read it myself. So I'll just click on one real quick, and I could learn more about this place that I so now if I click on X, you can see that my conversation was not saved in the different sessions because I went on incognita mode. But if you want to do it here, all you do is press on this research, ask the same question, but it's going to be saved here, and it may use the context that you've given in these other places. That's how you can research with Claude. It's really cool, and there's a lot more that you could do with it. So go ahead and research about something you want to do. It could be like the stock prices. If you're into sports, what is the next event? What are the prices like and that sort of stuff? So, give it a try, see what you can find out and how relevant the sources were to your question. Mm 12. Summarizing Documents: Sometimes you are given a big document that has so much text in there and you don't have enough time to read it. With Claude, you can easily summarize documents, websites, and it will make you a nice polished summary of it using bullet points, bulleted text, and any other details that could help you either better understand a document or take actions based off that. So for this lesson, we're going to be using a sample document that is available for you guys to download. And we're just going to add it to our chat bar right here. So here I have a PDF. It's about ten pages long, lots of text, and it's basically a meeting note and different agendas within the document. So I don't want to spend the time reading all of this. I'm going to use clot to summarize it for me. So you could either stay in the chat or do the co work if you wanted to do certain things after the summarization. Example of this application is saying, asking Claude to summarize this document and then maybe email it to certain people right afterwards. So we're going to give some permissions for this, and then it will do it on its own, sort of like an agent. But I'm going to stay within the chat just because we're getting started with this whole capability. And we're going to see how it works. So we're going to give this command, summarize this document for me, and bold the important parts, the agendas using bullet points. Then list me actions I should take for the meeting or goals. Point out some follow up questions I could ask from my team. And let's pretend that this was a meeting note sent to me, say I was sick that day and I wasn't there. So I'm going to give it that context. I was not at the meeting, summarize this document. We're going to hit sent and let's see what we get. So this was a sample document, as I said, and you can see how quickly Claude picks that up. So it's a sample placeholder document, no unique content. So that's something we already knew, but it did still do what we asked. So it did a summary, logistics, attendee experience, and then it gave us some action items that I could do if the document had actual owners, so it doesn't specify it here or deadlines. So if we had actual people, it will tell you, reach out to Peter about X, follow up with Jenny on Y, that sort of stuff. But it also did the follow up questions that I asked. So these are things that I need to ask my teammates just to catch up since I couldn't attend the meeting. Now, you could do documents. You could do multiple documents or even do websites. So I'm just going to grab a Wikipedia page and paste it right here so I could have Cloud summarize it for me instead of me reading the whole thing. But we're going to go over this Wikipedia page, sugary drink tax. Say our team is looking into designing something that has to do with this topic, and our boot has given us this document, this website to look at and form it into a document. So I'm just going to copy the URL, go back to Cloud, and then give it the instruction. Read this website and summarize the important details inside a PDF file, highlights latest trends, affected countries, important numbers and action items for my team. So here's the website. I'll put it in parentheses, and then I'll give it some more context. So my team and I are looking to develop a sugar free drink. Help me make a document to present at our next meeting. So this is our prompt, very simple, very straightforward. You can do multiple links, a link, and a document, and then have Claude do it for you. So this will be like your summary session. So here's my PDF. It's going to build it. And if you want, you could have the same prompt done with the research. So say I only have access to one website that I found, but I do want other information to be incorporated into this PDF. So I could tell Claude that this is the website I have and for it to research on the Internet and find other websites, extract the important information, and put it in the same document. That way, I don't have to look for other websites, copy their URL as well, and give it a longer prompt. So we can see that it's building the document as we speak, and now it's ready. So I did some summary of the document itself, and here's something I could present in our next meeting. Make it a full screen here. So we got a nice little abstract, latest trends, affected countries. We do have some formatting issues, but I could have it to redo it. But we're getting all the countries, important numbers, action items and other stuff. So I could be like maybe separate the tables into an Excel file and keep the other text in the PDF. It's going to look at skills. We're going to get into skills later. But basically, those are special instructions that I would have given Claude as the user to do Excels a certain way. So if I have a specific format, if my company has one header that needs to be in every single, you know, Excel sheet, Claude is going to just check that real quick. But I didn't do anything with the skill, so don't worry about that. So now it's done the split. We can see that we have a PDF and an Excel file. So the PDF, only the text, which is perfectly fine, but we just remove those tables and put them in an Excel file, which should be a lot better for formatting and the columns in rows. So now I could present both to my team and have a good meeting using Claude as my assistant. So you could use this for summarizing, building documents, combining documents, splitting them, and doing so many other stuff. Try this with a document that I have for you guys. I'm going to leave these two in the resource pack as well. So you guys can take it, try to combine them, separate them, maybe take this one Excel file, break it into two or three, have the research feature, add more information to it, and just experiment with what you could do with the cloud chat and the Cloud research. Now let's move on to our next lesson where we're going to talk about brainstorming. So when we have nothing in our hand and we're just tossing around some ideas. 13. Brainstorming Ideas & Solving Problems : In the previous lesson, we looked at how to extract information from a website and then output that into two different formats. So we're going to continue working on this specific file, sugary drink tax briefing. And what we're going to do is use Cloud information now that it has context about our needs to just brainstorm some ideas for the new project. So as we said in the previous lesson, let's say, the company is looking to design a new a sugary drink, and this was part of the research. So now that I have the research component done, I could look into, you know what flavor the drink should be, what are the trends and all that stuff. And essentially, I will be brainstorming with Claude. So I'm going to be within the same chat. Ideally, you would want to do this in a project, but we're going to get into projects in a further lesson. So we're just going to keep this in the chat. So this is where we left off. But now I need to give a little bit more context about the project itself and basically tell Claude that I want to brainstorm with it. Maybe I could have it interview me again, ask me some questions, and then that could help me get closer to that drink idea that me and my team want to execute. So let's start with a very simple prompt. My team and I are looking to develop a new sugary drink, and we want it to be trendy, fruity and fun in terms of design. So ask me some questions that can help me or help us brainstorm ideas about the new drink. Very simple. If you want, you could mention other stuff. I will mention a budget. We have a budget of around, I'll do ten K again. We want to do a form of AB testing. With two different flavors and figure out the best selling one. Again, Son at five, I did not change the model. Let's hit Inter and see how Claude is going to help us brainstorm. So I have five questions that I need to answer. The first one, what flavor direction feels most trendy, food you fun for your target drinkers right now? So I didn't mention my target drinkers, but maybe I'm aiming for teenagers. I'm aiming for every age. I'm aiming for children. That's going to greatly affect, you know, that flavor of the drink. Ocation is also important. So let's see if Claude is going to ask us about that. But for now, I'm going to go with tropical, and you can choose more than one answer here or even type something that's not over here. I'm going to go to the next one. There we go. It's asking us about the age. So I'll say maybe health conscious, the base. You could say carbonated soda because we chose sugary drinks. The important thing to measure for the AB test. I think I'll just say taste preference and the packaging appeal. Last question, how do we want to run this test? So we could do a pop up just for fun. And now that it has my answers, it's going to help tailor my ideas, put them down in one session. And later on, we're going to see how we could build visuals with Claude. It can give you flow charts, SVGs, landing pages, and so many other stuff. So this is only the beginning. Alright, so we have an answer. We got two concepts. It even gave us a name. On food Guava. And then it's giving me some packaging directions because I said the taste and the packaging needs to appeal to the audience. So it took that information, and it's giving me some ideas around here. And then here are some budget breakdowns. I didn't really ask for it, but Claude is going to pick up the fact that this is a team project and I need to bear these in mind. And it's also telling me how to run DAB test. Lastly, we have some additional questions that I could answer. I'm going to ignore the questions, but just come up with a random question that I will have during this brainstorming session. So what should I call the drink? I want it to be fun and easy to remember, but not pure English. So we get some ideas here, and I'm just brainstorming. Giving different languages, invented blended stuff. This is more of what I was going for. And then I could build up on this idea. So talk to me more about blended names and how I could incorporate this into my packaging. It's telling me a little bit about how the names work to begin with. It's giving me more names that I could think about, maybe share with my teammates and how we could carry this into the packaging. So these are a lot of ideas that I could now explore. Say I want to share all of this with my teammates, but I don't want to have to relay the message myself or write everything up, as well. So summarize these ideas into a visual presentation that I can share with my teammates or with my team. Let's give it one lead. I'll choose one of these guys to begin with. We can go for this guy. I like the best. So notice how in this lesson and the previous lesson, it's talking about a certain skill. We're going to get into skills later, but as I said, those are really handy when you're working with certain templates that you have to adhere to. Say everything that you publish for your company needs to have a certain letterhead, a certain tone. Those are things that you get to save in clot skills so that every time you ask for a presentation or an Excel file, it applies those rules to every single file. So you don't have to make edits later. But again, we're going to dive into that later on. So now it's building the JavaScript file and you can see this happening live, and now it's going to convert it into that visual presentation. Again, you could make this a simple written PDF file or an Excel sheet, whatever format you prefer. As long as you mention what it is, Claude will execute it for you. So we got three slides. I didn't say how many slides, but you could also specify that. Alright, so now we got six slides. It's going to open up a side panel, and it's a PowerPoint presentation. So there we go. Got a nice color. Going on where the name comes from, meeting the flavors and the test plan and open questions for the team. So this is one way that you could take all your brainstorming session and produce something out of it so that you could share it with yourself or anyone else working on the same project. And notice how this thing just popped up. This can be connected to external platforms or your local drive. So Cloote could make this save it in a folder within your computer or send it automatically via email, upload it automatically to your Google Drive, and that is done via connectors. Another thing that we're going to learn more about in a further lesson. So now we have our brainstorming session. It's complete. I have an idea. Let's go ahead and see how we could plan with Claude so that we could actually take action from this idea. 14. Planning Projects and Tasks: So this is where we left off. We made this presentation out of a very simple brainstorming session that we had with Claude. Now I'm just going to close this off and give a little bit of basically more context about the timeline I have for this project, some restraints that I have, and have Claude give me, like, a checklist that I need to do in order to be able to execute this new drink launch. So I have three months, let's say, to finalize the packaging and flavors for this new drink. Ask me more questions. That will help us build an action plan so that I have the drink ready. Very simple prompt. Let's see what sort of questions it's going to ask just so that it could build upon its own context. So clarifying questions. We got three more. Where am I in the flavor formulation? I think I'll say number three, who will handle the production? Let's say we already have someone and then what happens after the AB testing? I guess we'll do pitch to investor. And then this will be my goal, but we need a checklist as to what I should be doing. Say, I'm missing out some important details, and then we got three more questions. So this just prompted on its own. I didn't say asking more questions, so I'll say dedicated team. Where do we stand on this? So it handled. For the pitch, I think we'll do like three to six months. And it's just going to process what we give it, ask more questions if it needs, and actually give us a good plan. So month one, week one to two, we got these things that we need to do second month, third month, and it's even telling us a few points that we need to consider moving forward. And it's going to ask me if we want to output this. So I'll just have it do both. Here are the two files. We got the document and then the Excel file. And as you can see, it's formatted really well. We get checkboxes, too. The same information about the action items per week and per month is written out and also the points that it wanted us to watch out for. And then we can see the launch timeline. Click on that. And this is what it looks like. I'm going to extend it here. And it's looking pretty good. So it's color coded, and it's nicely formatted. I could download this, bring it into Excel, or have Cloud do additional changes right inside this chat. You just hit X right here to collapse that. And you can do multiple outputs. This could also be a landing page if you want to advertise the drink before it's out. Cloud can do that for you. It uses the code capability. You can also get that done with the live artifacts feature, which we're going to look at in a further lesson. So this way, you can come up with a project, make a good plan for it, get some research involved, and finally get these actionable items that you could share with the team or focus on by yourself. So now you can try this with anything else that you want. It's pretty easy, pretty straightforward. And now you have the tools to start something from one idea, one concept, build upon it with the research feature and the following output. Let's go ahead and dive deeper into these tools, which are the core features of Cloud. 15. Working with Projects : In this lesson and the ones following this, we're going to look at the core features of cloud in a more basic way just to get you guys started. So as you can see on the sidebar, we have a bunch of things that are not as common in the other large language models. May have wondered what they are. They look very different when you click on them. And we're just going to do a quick dive into each one because as I said, there's so much more you could do with Claude, but it helps to get a very basic understanding first, do a few experiments, get a hang of it, and then go onto the more advanced things. So first, let's get started with projects. If you click on it, it's right underneath it's going to look very simple, but essentially projects allow you to have a dedicated room for one specific thing. So you can think of projects as something more elevated than the chats. Now, in the chats, they're going to Claude is going to remember the things that you tell it. Context is still in there. It collects the information you provide, it gives you outputs, and all of those things. But projects allow you to have more focus and more details involved in that need. And they're really helpful for long term things. For example, you can see a few things down here, study projects. You just started a semester. You feel like you're going to struggle with physics, for example, you can start a study project where it's going to be with you the entire semester. It's going to learn your learning style, what you struggle with. You can always come back to it every single day, ask it questions, have it generate you quizzes and stuff like that, and that is a long term thing. So it's always going to be available for you. The same thing could apply for your career, say you're looking for a new job and you want Claude to be an assistant that will help you tailor your resume, write your cover letters, and maybe write emails, prepare for interviews. This again is a long term thing. So you want to leave zero rooms for hallucinations. You can also do the same thing for research. If you go on this example project, there's a chat bot just like the chats here, but there's a different box for context. This is where I said it's going to basically use a lot more detail for the outputs because it's first going to study whatever you put in here. Say you made a study project. Going to put in your textbook, your instruction notes, maybe the syllabus, stuff like that, and Claude is always going to refer to this box before it gives you an answer. So this way, it's not going to randomly pull information from Wikipedia when it's trying to help you prepare for your next physics exam. So it has this box over here, and there's also a cowork feature. Now, cowork is going to be something more actionable. Clot is not going to just give you an answer, like the chat, it's going to do things. For example, this, of course, depends on the permissions you give clot. For example, with cowork, it could read your slack, find one message, print it out for you here. It can send emails. It could look into your Google Drive, extract all the pictures. I could look into your downloads folder, stored everything from, let's say, the amount of time you've used it, and so on forth. So it it has the ability to interact with the things outside of clot as long as you give it permission. So whether it's one task or multiple tasks, you can decide on which one you want to use. But we're just going to focus on the project here. So let's go ahead and make a new project. We're going to give this a name. So let's give this a project. I'm going to do something for studying, just so it's versatile. We'll do, like, Irish literature, too. Let's say, that's the name of my course, and then some goals. What do I want to achieve with this long term project? We could say, get better at understanding historical contexts, unique terms. Character analyzing and other stuff. So we gave it the goals that we're trying to achieve with this project. And then there is an ad folder feature. This is something that you get to use if you want to upload a folder from your computer onto clot. Say if you're doing a project about basically sorting out meeting notes and making agendas, if you have a folder full of meeting notes, this is where you could upload. I'm going to leave this out and basically upload two separate files. So create the project, and now we get that chatbot instructions and the contacts. So for the contacts, I have two PDFs. You guys can download it as well. But these are things that are related to my course, and say, my instructor gave these two files to. Just hit the plus, upload them here. So it could be from your device. You can add text content. So if there's any important detail that your professor mentions in the class at the beginning, you can put it right here. If it's a file, you get to upload it, and you just have both options. There's also GitHub as well. So I uploaded those two for my context. Again, you could do textbook as well, but I have one about James Joyce and one about the troubles. And these are just Wikipedia page. Are free to use for everyone. Next, we have instructions. So add instructions to tailor the responses. This is if you wanted to output in a certain way. For example, if you want, you could say that whenever you're trying to analyze a character, make sure you use one or two quotes from the stories or make sure that you're using quotations from the sources, that sort of stuff. But something unique to your needs, you can put it right over here. So maybe you can do this since we're doing literature. Analyze the short story and provide two quotes, direct quotes. Every time you are asked about a character. So that's one instruction. You could do multiple. Let's save this. And this is unique. I could always go back to this, by the way. And you can always add more context. Now we can either use a chat feature or the cowork. If you go to cowork, you have two options, and these are the permissions that I was talking about earlier. So for example, if you enabled Claude to build something with coding and save it onto your Github, it's going to ask you for permission. So build this file question mark, and then you could approve it or you could have it not ask you at all. I recommend having this on because sometimes it may want to do things that you didn't think about and you want to make sure that you're aware of that. Posting things on your slack. You want to know what's going to be on there if it's going to be from you. So make sure this is on this. We again have the same models. You have the voice mode, and of course, you could add things to the chatbot. So we have the skills. This is something I'll tell you later. The connectors are also for later, but we're just going to focus on this for now. So I'm just going to grab a short Irish story to bring into here, say that is my homework. And we're going to have Claude help us navigate through this difficult story. Given that it has context about James Joyce, the author and the troubles, just so that it knows a little bit about Irish history, it's going to be able to better help us. So I'm just going to grab that story. So I just downloaded Ulysses. This is also available for you guys. So that's my story. We could add it to the context as well. If this is going to be a long term file. Say you're going to spend the entire semester for this, you can put it in here. But if it's one homework, you could put it in the chat. So let's say summarize each of the characters within the story. Very simple prompt. Let's send. And while this is happening, we have some new things on the side. So we have progress. So if there's a task that is going to basically happen in the background, it's going to show up here. If you're saving anything onto a local drive or onto a cloud location, you will see them here as well. And these are just the contacts. So we have the project right here. So now, because I asked it to analyze all the characters, we have a folder that contains this text file. So this was a PDF. But it was like pictures of the pages because it's an old story. So first, it extracted the text. And because I gave it the instruction earlier to pull quotes every time I ask about characters, you can see that it's exactly doing that. So now it has the solid quotes. It's basically running through everything it's doing for you, and it's creating the files. You can see the amount of characters, and I could do follow up stuff here, make edits. But notice how on the left side, we've been doing chats this whole time. This guy is different, so it's a different type of session. The file is complete. Here is the character summaries, and there's the two quotes that I asked for. These were my instructions. It's giving me a little bit about each one. I get to scroll down and maybe use this for my study sessions or for my essays. So again, you can make changes here. The files are saved. I could be like turn this into a pop quiz to prepare me for an exam. So turn this file into a pop quiz for me to study with. So it's going to basically use this space to make everything needed for me to get better with my study. Here are the questions. So we have part one, part two, and there's, like, the answer keys underneath. So that's an additional file that's saved. If I close this sky, we now have three files in my working folder. So that's pretty much a general idea of the project. This was one application, but you can imagine that you could use this for any other long term goal that you have. It's really helpful. It's saved for you right here. You can always come back to it and build multiple projects. Now, something that goes hand in hand with projects are the artifacts. Is right underneath here, and we're going to take a look at that in the next lesson. 16. Creating Interactive Artifacts : The artifacts are products that you can build with clot, and they go hand in hand with projects. So this could be an interactive quiz, it could be a landing page. It could be a simple calculator, things that you get to make in a few seconds with clot and save it for whatever reason that you have. So it's right underneath project, and when you first open it, it should be empty if you haven't used artifacts before. On the right side, we get to filter a type of artifact. So we have chat cow which is what we looked at in a previous lesson and code that I'm going to leave out, as I mentioned, because there's so much you could do here. You want to build a new artifact, you just go over here. There's two ways. You have the chat one for one task and the cowork for multiple task and multiple outputs. I'm going to start with the chat one. We kind of looked at a cowork example in the previous lesson. So now, immediately, it's going to ask for a category. So what is the product that I'm going to build? So I'm going to do a productivity tool. Let's click on that. And it's going to ask me some questions so that it could better build this product. So kind of tool, let's say, time management. Main problem help me with everyday health, I guess. I must have feature, positive and motivational quotes like that, and I'll do, like, bullet points so that it's nice and organized. Use Shift and Inter to basically format this because if you hit Enter, it just sends the prompt. So that's two. I'm doing Shift eight for the asterix, and now we get three bullet points. So right now it gave us a little brief. This is the name of my tool. These are the sections, the check ins. And then we have the quotes here, rotating positive quotations and a progress snapshot. Now, it's going to ask me to double check this if I want to make any changes, I could do it here or else it could build it for me. So say this is perfect, I'm just going to say build it. So now, because we are in artifacts, it's going to build something for us that's available in the clot application, we're on the desktop again. But you get to maybe upload this onto Google Drive, share it with your slack members, and basically decide where this product goes, but I'm going to keep it within the application for now because the connectors is going to be something for a further lesson. So here's my product. It's done. You can see that I took a while, but we have today's focus. We got a timer feature, and I'm just going to make this full screen. We get a nice color scheme, and we got our motivational quotes that I get to refresh for something new. That's an example of an artifact. This is an HDML. You can download this, host this somewhere else if you prefer, but you can see the preview or the code and do with that as you will. You can download it or publish it, as I said, and make changes if you want via chat. So you can be changed the color scheme, use these two colors that I like or something of that sort. Very straightforward. Artifacts are great when you combine it with other features within Claude. But as a general idea, this is how you work with it. On its own. It's really fun to explore with. And as you saw, there's different categories. But in the next lesson, we're going to look at something that's more visual and used for more visual applications. And that's the design feature down here. It's fairly new so make sure you're updated to the latest version so that you could follow along. 17. Introduction to Claude Design: Claude recently added their design feature, which allows you to turn ideas into visuals, whether that's a poster, a logo, or any other sort of format, and it's done right within the application. So say you brainstormed an idea, you know, you came up with a logo idea or, like, a new business to start, you now don't have to leave Cloud for Visual aid, but do it all within this tab right here. You click on it once, you're going to be brought into this window, and it's pretty similar to a lot of the other platforms. They've now added a chat bot that outputs visuals. If you've used Canva before, this may seem familiar, but there's a big difference with Cloud and those platforms. With Cloud, you can do prototypes, slides, documents, wire framing, and even animation. And you could do reference files as always, get different design systems, can use templates as we saw, and you can even program something like within a code. For example, you use clot code to build a landing page for your new business. You can now turn that code into an actual prototype inside. Let's get started. We're going to do a very general briefing of this because as I said, every feature within Cloud has so much capabilities for you, but it's good to just get a basic understanding first before you experiment further. As you can see, it's Beta, so by the time you guys see this course or a couple of months later, this may completely change. But essentially, we have this chat bot where we get to speak to Claude to build whatever we're trying to build. You can see some suggestions. We have wireframes for landing pages, we got briefing documents and all that stuff. First thing is the plus icon, which allows you to attach reference files. You can attach folders. This could be from your local drive or something else you downloaded from your work computer and then reference another project. We went over projects already, but say I wanted to turn my Irish literature project into something live, I get to connect it right here. And then we have code. If you are using Cloud for code, Github, local code, which is within your computer, you have these two options. We have designs. If you've used FIGMA before, then this should be pretty useful for you. If you haven't used FIGMa, it's basically a design platform where you get to build wireframes, mobile user interfaces, web, and so many other stuff. Whatever you design there, the full package, you get to download it from FIGMaimport it right here so that you have basically more control over how your prototype comes out. Design systems, so you can set up the system so that it maintains your branding. For example, if you're designing for your company with a design system that you set up here, it will keep the logo, the colors of your company, typography, tone, and so many other stuff. And then we have skills, which we'll get into in the next lesson and then connectors, which is also the lesson after. These are references, different types of references. Design System which is looked at it, if you did end up making one, so if I just click on it, it will show up over there. You can either create one from scratch or you could have Cloud code do it for you. If I click on that, I get to put in my company name. If you're doing this for yourself, you can put your own name. You can put a link from Github, code from computer, Figma file, and add all the design assets that we talked about. You can also put in some notes regarding the tone, the audience, and then you can put in like b to further specify what you plan to use this design for. And then whatever you save is going to show up here, a drop down menu, very handy. Templates. These are things that you make yourself. So if you just go down here, when you make one from the projects, they're going to show up here. You get to reuse templates, but right now we don't have anything here. There's a search bar as well. Your design systems are also going to be lined up here and your projects as well. Once you have that, you can add in some code as well. Again, I'm leaving the code part out. The models are right over here, and then you have the Send button. So here are some templates that you get to use to get started. We have prototype, slides, document, wireframe, and animation. Let's get started with the slides because we did use Claude to make some study guide for us for that literature class. So now you can see my template has slide it's going to maintain Power Point formatting, the rules that are associated with that in terms of the number of characters per slide and other stuff. You can add in a project, and I could let's actually do prototype just because we did slides already. So let's say design a one page landing page for a new skincare product. Very simple. Of course, the more details you put in here, the more it's going to be tailored. You can mention typography. You can mention the style, the audience, and many other stuff. So now we have this engine running, and you can see that I'm getting all these questions that I now need to answer to better help Claude continue with this design. So it's not going to just take something from you and give you a bunch of random designs. It's making sure that it's relevant to your actual need. So here's a chat bot. You can see it got named after what I asked. And we could hide this sidebar. You can go back to the chats and now Claude has some questions. So whatever follow up you have, you would put it down here. Let's actually look at what sort of questions they are asking us. So product name, online description. Let's call it I don't know, beauty Smart, Beauty smart, the best moisturizer in the world or something. So Beauty Smart, the best moisturizer in the world with, let's say, special serum. Overall aesthetic wipe. We want maybe warm natural. I'll do just that one. You can also do decide for me. So Claude is going to look at what you gave it here and make a decision there. I'll do like this guy, any colors. I'll leave this empty. We'll do elegance. What sections do we want? This one, ingredient breakdown. Something to buy, imagery approach. I'll do decide for me. How many design directions would you like to see? So how many versions of this design? I'll do one for now. So I didn't give it a lot of information. This is great if you're just experimenting. But obviously, if you have an actual project, you will have a lot to say in these areas. Continue once you're done. We now fed it these answers. It's thinking, designing, and we get to see this happen live for us. There's our logo. Got a button. Got the fonts that we asked for, the color is exactly what we chose. And now it's generating, I believe, a picture or it's a placeholder. We'll find out. Got some random ingredients. I didn't say what the ingredients are, but they seem pretty relevant for a moisturizer. We even got a random pricing. There is the ritual, the frequently asked questions that I asked for, and we even have a fotter. So right now, it's doing some final touches. And there we go. So this is the final product, which is looking pretty good. We have actual questions and answers here. Pretty cool, working buttons. So ingredients, pricing, FAQ, shop now. They all work. That's our logo, and I could just go in there and be like, change all the green to yellow, you know, make all the adjustments that I need. What you get to do with this design right now is share it. So you get to share it to your workspace, keep it private for yourself, copy the link, implement it with code so that you get to host it on whatever server you want. Make it a PDF. If you're proposing a design to your team, you get to do a PDF here, make it into an HTML file. Power Point, and there's even more formats down here. You can also present this. So new window in this window full screen. Make some edits down here. So if you go to simple, you can add a text if you want, like I could just do so fun. And I have all these options. We could add in some code as well if needed. And make some tweaks that we don't know how to implement, but we have an idea of. So I don't know, make this more kid friendly. That would be a tweak. Claude is going to read your prompt and make that adjustment. You can annotate this. So if I was the next person getting this design, say, my co worker did this, I could now edit this annotate this and leave my comments. So let's say too small, that's a comment. I could make the change. As you saw, there is let me make another one send to Claude. So Claude is going to make it bigger because I said too small, and it's all going to be listed here. And then there's zoom levels you could do. There's also drawing, which is pretty cool, so I could do something like that. Like, no need. You know, that's just an example. Leave a comet and do a bunch of other things. So, look, I could send this and now it's going to Claude is going to look at my critique and redesign the section. So, as you can see, it's getting rid of all the circles because I said no need. So you can either leave a comment for whoever that's going to make the changes, or you could have Claude make the changes. That's pretty much it. So now you could refresh for another version. Make a new page. If you want multiple pages for this design, you can just make it all over here. Now, I could go back to the design, just click on that little logo, and now it's saved here. So I could always go back there, make changes, and go about my life. Let's try the wire framing as well because documents slides, animations, they're very straightforward. I'll just use one of the template here. So build the user process. For a skincare shop. It's actually user journey with wireframe. It's hit Send. I'm not referencing anything else. This is just pure basically starting from scratch. Again, we have a little pop quiz that we got to do. So what is this journey for? We'll say website. Where should we begin? Onboarding. So when they're making the account, what's the vibe of the prant? Why is this unique? I'll leave this empty. Familiar stuff. One direction, and I'll just do structure only. Actually decide for me. So this is another design output. It's going to think, start designing. We get to see this live on the screen. The options are still the same thing as the previous one, so it's only a matter of the format and template Claude is using to implement that prompt. So, there we go. So this is onboarding. There's a little sketch for us. We can see the journey that the user would have to take, and it's looking pretty good. There it is. Let's try to make this full screen, and I now have these different wireframes that I get to work with. So it was done in a matter of seconds. You can hit Escape to go back and the logo to go back to the main screen. So that's how Cloud design works. Again, this is a new feature, so it's only going to get better, but feel free to experiment with the different types of templates and see what you can make either using real projects or some random ideas that you want to explore. Now, let's go ahead and move on to our next lessons where we're going to talk about skills and connectors. 18. Skills: Skills and connectors can really help boost your workflow because it allows Cloud to integrate and interact with other applications that you use daily. So this could be Gmail, drive, slack, or any other platform. Cloud can use the cowork feature to either help you generate emails, look for a certain file within your folders. Maybe highlight certain things you forgot to do from your slack and do a lot of other things. So in this lesson, we're going to set it up, and there's so much you could set up here, but I'm going to go with Gmail because I'm pretty sure everyone has access to that, but you're going to see very soon that there's so many other things you can connect to Claude and so many other skills you could tell Claude to save. To access that, we just have to go down to customize right over here, and there is Skills connectors, plugins. So plugins are straightforward. I'm sure you guys know how to use plugins. You could just browse some right here from Enthropic or partners. And, you know, for example, based on the design experience we just had, if I connect FIGMA, I can bring in my FIGMA files directly into Cloud without having to leave the application. So that's an example. But we're going to focus on skills and connectors in this lesson. So here are some enthropic templates that you can get started with, and they're very straightforward. You can see a lot of people rely on these. These are the amount of downloads that are currently available. I'm going to go with a very simple one, which is learn when hit plus right here. And now that is added to my skills. But what you could do is make your own skills as well. So I'm going to go x right here. Go to skills. There is AD, and there's a learn that we just added. So if you go to AD, you can create with Claude, write the instruction yourself or upload like a text file. First of all, what is a skill? A skill is a set of unique instruction that Cloud will execute every time there is a need for that. For example, if I give Cloud Excel skill, it's basically going to apply those instructions that I'm going to write here every time there is a need to make an Excel file. So we saw this in the previous lesson where I made a bunch of data tables out of nowhere, and we saw that it was searching for skill. So if I give a instruction right here that every data table has to be blue it's going to apply that every time I ask for a table. So it's very straightforward. We're going to actually experiment with it in this lesson. And the first thing you need to do is make a skill. We're going to look at the learned skill afterwards, but let's see how this works. So let's give this a name. We can say Excel format, space format. Description how the Excel files should look, and then the instruction. So every file needs to only use the color blue for the tables, and all the text needs to be folded. Say, those are the design branding for my company. Everything needs to be blue and bolded. You can also add, like, a hex code here if it's a specific type of color you want to implement it. So this is very simple. We're going to hit Create, and now it has been created. So we now have Learn and Excel format. So first, let's try the learn feature. That's something by enthropic. It's a something used a lot for educational purposes, for personal growth, for work, and basically any situation where you need to learn something effectively. So I'm going to go to this guy, the one with the Irish literature that we made the project, and I'm going to use the learn skills Learn. That's how I'm going to call this skill. Let's say, each character within Ulysses and the main conflict. So very simple prompt. We're going to see how it works here. So this is going to first clarify my question. So have I read some of the novel, and, like, basically, Claude is asking me how much I need to learn for this? So say I am starting from scratch. We now see that the skills integrated here are learned because I just called it C. Before I name characters, does modeled on the Odyssey mean anything to you already. I need to respond to this. I don't know anything, let's say, starting from scratch. But it's very interesting how Claude will make sure that it's doing the right thing for you. It will not just generate something random, which a lot of the other language models tend to do. But now I'm starting something from scratch. I'm using the learn feature, which integrates visual aid as well. So we can see the skeleton for the characters, and then we have like these characters, and this is going to help me learn it. I could ask more, dive deeper into this character, give me a summary of the conflict, that sort of thing. But that's how the learning works. If you go on it, click on it, you get to see the scale dot D. So it gets triggered when you have a explicit learning request. It does not get triggered for coding, writing, calculation, personal troubleshooting, and there's also a learning mode. And you can just read this if you want to know more about how this works. And because it's learning, it has a warm direct tone, which is really helpful if you are struggling with something big here. All right so that was the anthropic skill. Very straightforward. Again, there's tons out there. You could just add them to your application. But now I'm going to do the Excel example. So I think it was here. Okay. So slash Excel Format. We do the Excel file for the launch timeline document. So it gave it to me already, but I want it to be reformatted. And I just don't want to say every time, make sure it's blue, make sure it's bolded. So now it's checking the skill, and now it's checking the document, and it's going to do its you can see how helpful this is if your company has a header that always needs to be implemented. This is how you make sure that you're saving time with Claude since Claude would be doing, like, the writing and the research part. You don't need to go into every file and re edit it so that it has the logo. It has the colors and other things. So this is a really cool feature, and there is my document. It's now blue, no longer green, and everything is bolded. So if I could show you guys the original, I think it red it, replaced it, but you could ask for it to save the original as well. But as you can see, I didn't say blue or anything. I just called this skill that I made within a couple of seconds. So that's how skills work. It's really helpful. Feel free to experiment with it and add as many skills as you want. So if you go right here, you can add them this way, as well. So I could do two at a time, three at a time, or even none at a time. It really depends on what you need this for. Now, let's go ahead and see what connectors are because those are also very helpful. 19. Connectors and Tasks: So going back to customize, you can just go down here to Connectors, and these are the permissions that you're going to give for Clade to interact with the other applications. So I just kept in Gmail. So let's see how we can even do that. Just go over to Ad browse and then look for the application that you need. So I did Gmail, and it's already connected, but let me show you something that's not connected. You're going to click on the Connect button. It's going to bring you to a browser where you sign in, you give it permissions, and then it'll bring you back to this application. I can't show you that bar because you're just signing into your Google account or whatever email that you used for this application, giving it some permissions. So that's what I did with Gmail. And down here, it tells you what clot could be used with this app. Drafting replies, summarizing threats, and searching your inbox. Let's go ahead and close this since now I know that the status is active for Gmail. If you click on Gmail, you get to work with the permissions again. This could either be done through Cloud or through your Google account, but these are the stuff that I let Cloud do, things that need approval. So I highly recommend if you're using a very sensitive email to come in here and make sure that Cloud is not doing anything on its own, or if it wants to do something, it needs approval from you first. You are not going in here blindly, and then you wake up one day to see Claude has written like 50 emails to your coworkers. Alright, so let's hit X, and we're just going to go to one of these guys. I think this was the one with packaging. So this is the chat from a few lessons ago with the drink that we're going to work on. While I'm here, actually, notice how after we talked about the slide, the slide that we made for our new drink, it prompted clot design, so that's how you get to connect the two, but we're not going to go to that right now. So let's go to the productivity tools chat because GMAil does link into this field, and we're just going to go to the plus connectors and make sure that Gmail is turned on. So we could look through my emails and highlight any thread about productivity. So now you can see that it's looking through my Gmail because I gave it permission, and I'm just asking it to search my inbox. So I got an email from Gemini once, Slido Slido grammarly. So now I could have it maybe draft me a response to something or draft me a new email. So draft me a new email for my friend asking whether she wants to work project together. Show me the draft and send it to. So I'm just putting in a random email. We can put the actual email, and then you can hit SNT specify the project here, but I'm just doing something random. So it could be related to the stuff you were mentioning here, but you could really use Claude to either do responses, draft new emails, draft new initiatives, that sort of stuff. So subject want to work on this together. Hey, and all that stuff, let me know if you're up for it. And it did not send it it wants me to confirm it, which is good. So I could now copy this, send it myself, open it in the mail app, open it in a Gmail app, ask for some tweaks, and then have it send it on my behalf. So you can see how helpful this is going to be, especially with a cowork feature where we can have Claude do multiple things. So we could talk about a project and ask Claude to search for, like, a special email in our inbox regarding this project. Email seven people with their task, maybe draft a new presentation for us to show and do multiple things. The skills and connectors really come in handy here because you get to keep your standard or your company standard in whatever work that you're doing here. The last thing that I'm going to show you real quick is the schedule tab, and these are scheduled tasks that will happen with one routine. It could be daily briefing. It could be weekly reports. Maybe every week you could have Cloud run through your inbox and look for any unread emails. So that sort of stuff, you get to define the new task right here, set up manually, weekly checks. So search inbox for unread messages. And I'll just put this down here. You can save this to a project. I'll just leave it out like this. Make sure it's manual so that whatever cloud wants to do, it asks for your permission first. And you could have the frequencies set here, so weekly. And then you set a time, you hit Save. So it's very straightforward. Everything you make is going to be listed here. And another important thing is that you could have this happen only when you are awake, meaning that the application is turned on, and it's not going to do anything when you're not on your computer. So that's something you could do. There's next ran names. If you have multiple things lined up here, you know, it's pretty fun. If you're the team lead or the manager or the CEO, this can be very helpful for you. And connecting that with the skills and connectors, you can see just how much you can get out of Cloud. So now we know about all the core features within Cloud, what you can do, what it's good for. But now we're going to apply this to a project that we're going to do together from scratch. So that's going to be the focus for the next couple of lessons before the end of the course, just so you guys are able to apply what you've learned onto something and see what else you could do with this amazing tool. Oh 20. Planning the Study Guide: So let's work on our project where we're going to be solving a problem that occurs in our lives. This could be something small, such as finding restaurants every time our friends want to hang out, or it could be a bigger thing. So what we're gonna do in the first lesson, and it should always be the first thing you do is plan as to how Claude is going to help with that problem. This project, we're going to be building a workflow that we get to reuse as many times as needed, and that is going to solve a problem in our lives. So you can all think of something big or something small that you're dealing with right now and that Claude could help you with. It's usually something regarding productivity or communication, that sort of thing. And I want you guys to first think about that problem. For my case, it's going to be something regarding studying. So I'm going to be building a quiz workflow so that Claude could design a quiz for me every time I struggle with a new lesson. So this is a long term and if you have the mobile app on Claude, you can use the same workflow on your mobile as well. So the first thing you need to do is plan, and that's the focus for this lesson. Number one, what is the problem that you're going to solve? Number two, how is the solution going to be? Now, if you have no idea how to solve that problem, you can just use a general chat to tell Claude I have this problem at work, at school, wherever. And how can you help me? So I'm going to basically pretend that I don't know how to solve my studying problem so that we get to see it from scratch together. This is the planning phase for a workflow. It's very important. A lot of time it's overlooked. People think that Claude will figure it out for them. But the problem with that is that you're wasting your usage, and you're also wasting your time. So first, let's begin with a new chat right over here, and we're going to have a conversation with Claude. So I'm starting a new semester, and I'm worried about one of the classes. Say, it is it is linear algebra, and I want to build a workflow with Claude that will help me understand the equations and then be able to solve some questions through quizzes. Ask me questions that would help you build this for me and help me learn and improve long term. So right now, we're going to switch to cow because I'm asking for multiple things. First, I wanted to basically summarize my lesson then build me quizzes, tell me the correct answers, and maybe something else that clade things would help. So this is my prompt. I could make a project right here, so that's really helpful. But first, I want to start by getting an idea, so I'm not going to do anything else. Let's actually keep this on chat just from the scratch. Click Sents going to ask me the questions. So I've never taken linear algebra. Let's pretend. How do I learn best? You know, we could do one of these, or you could say, I have no idea. You can even skip. So say I like to see examples. How do I want to use this day to day, let's say, before each class. So these are my preferences. Claude is going to process them, and we're going to come up with a plan together. So now it is suggesting these things, preview of the topic. So if I put in the syllabus, it's going to know. Worked example, I try a few things mini quiz and a quiz log to see how well I'm doing or how bad I'm doing. And then we got a bunch of other questions. So do I have the syllabus or textbook? Let's say, I have a textbook. Track your progress, yes. And then it's going to ask me even more thing. I'll tell you, like, the title. And now, so far, we have what we need. This is the plan. Refer to your problem, see if this is going to help. Do you have any additions you want to make? And basically, we're going to bring this plan to life using all the features we have learned. So this is part one of the project. Make sure you have something you're happy with. If you want, you could write it down on a piece of paper, draw a diagram. Whatever helps with you. For example, if your problem is something work related and someone needs to input during the process, that's something you need to tell Claude, as well. For example, if you make a presentation, your manager needs to approve it first, and then you need to, you know, fine tune it, then present it to the investors, something like that. So that's something Claude needs to see in the equation. So make sure you have something like this, and we can move on with the actual research aspect or the setting up of this workflow. 21. Researching and Organizing Information: Once we have our plan, we know exactly what Cloud could help us with. We're going to make a new project. So this is as we saw a space dedicated for this problem, and we could integrate multiple things inside of here. So I'll go with another study project, and I'll call this linear algebra. Create the project. And this is where you get to add reference materials because I said I have a textbook, I could upload it right here. So I just uploaded my textbook. You can see that it tells us about the capacity. We have used 27% because there's a lot in a textbook. You can even see 73,000 lines. You can add anything else if there's a syllabus that your professor gave you or any other additional material that's going to help Claude, upload it there, and then continue. So now we have a context tab, the instruction, which it pre wrote, but could alter this as well. We got all these stuff, and then we could start with in this stage, if you're working with a problem where there needs to be a lot of research involved, you would do the research now. So what is the best way to learn linear algebra? You would have Claude look into that and provide you with some examples. So I'm going to go back to chat for now, where you're going to do research. What is the best way to learn linear algebra? Let's say, college. So this is the research portion. If you're doing work related things, this would be different for you. And now it's going to give us some things that it researched. Alright, so now we have all of these things. If you disagree or you want to alter something, you can do it right now. But say this is perfect for me, and now I just need to begin the study part. So we're going to go back to the project. Just click on linear algebra right here. And what I'm going to do is set up some skills, basically, instructions that I need Claude to do at certain points. Just going to head over to customize real quick, go to skills, and I'll do a new one for linear algebra or just linear. Use bold font for equations, and italic font for explanations. This is my unique Oops. I'll have to put this down here. So I'll call this equation formatting. This is an example of a skill. We would add it right here, and now it is saved. You can add multiple ones as well or browse for more. I'll add this one for later web artifact Builder because I did ask for quizzes. Hence why I want to do something with artifacts. You can add anything else that you want. Let's see what else? I think that's it. I made one and I chose one. Those are now saved in my skills. Alright. So whenever I want a skill, they are there for me to use. And now I could begin by learning a certain lesson. So what I'm going to do is just generate a fake document in this chat. Let's say that's my lesson. So build a PDF for Lesson one. And then we're going to download it, bring it into our project. But if you already have a document, you don't need to do this part. So you can see that because I asked for some problems, example problems, it also provided me the short quiz that I asked for, but this is not currently in the project, so this is something external. Here's our document for vectors. Looks pretty good. We have, nice tables, and we have a little quiz section. I'm just going to download this. I'm going to pretend like this is something that I get later just to show you, like, a long term application here. So now let's go back to our project. So whenever you see this, it indicates that you're in a project. I'm just going to go into my instruction and make sure that it's something for me. So there's a lot of things that I didn't really ask for, study flashcards and all that. So what I'm going to do is delete this completely so we could work on it together. So based on what I saw from the research portion, I found out that it's helpful to do some work problems, get explanations, maybe some quizzes, and a bunch of other things. We're going to do is tell Claude who to be for us to solve that problem. So pretend or act as maybe a tutor, helping the student with their understanding of course materials. You will do the following. So one, if document is provided, do a quick summary and present it to the student two always ask if they would want to do a practice problem. The include real world application for each document provided, since those will be like the lessons. And I guess that's it for now, and then we'll do one for tone. Then four, if student seems lost, present them with questions that would help you understand where they are in the learning process. These are the unique instructions for my problem. You could add in more or add in something less. You can also specify the way Claude interacts with you in this project. So keep things professional, maybe be very strict, you know, whatever works for that problem, and just save the instruction. Oh, and I forgot one thing. The quiz part, let's say, prompted. So generate an interactive quiz with a live artifact. So save the instructions. And now I could begin using this for my first lesson, maybe move throughout my semester, come back to this, ask you a question, and it will always relay and reflect on my textbook and anything else that I've provided here, as well as the instructions. So now we've researched and we've built this project that is relevant to the problem. So make sure you have an instruction. You have at least one document uploaded for your contacts, and everything is set up. So you know exactly what you need to do that's going to help solve that we're going to move on to actually testing this and seeing how we could use it in the long term. 22. Creating and Testing: So let's go to cowork and I'm going to upload my first document, say, I went to class, and I saw that, you know, I'm really lost with that lesson. So what I'm going to do is add that lesson that we made with one of the chats. So lesson one factors, and we're going to be like we had this lesson today. I am so confused. Just like that. And we're going to do slash. That's for the quiz. But let's do linear and just hit sent. So it's going to use the skill that I just applied. It's going to read the document. We can see the uploads here. Basically, all the equations are in bold and all the explanations are in italics. So that's the skill that we introduced here. It's being applied. So want to see a practice problem. That's something we kept in the instruction. I could say yes or no. I'll say for now. Let's see how it's going to work on this without any integration of design or artifact. So now I have to, like, solve this, and I could upload my answer right here. So this is like a little tutoring space that I have now built. So what I'm going to do now is if I want, I could add a connector here, let's say Simoab or a platform that's helpful for linear algebra. But instead, what I want to do is make me a pop quiz with live artifacts and cloud design. So now it's actually going to visualize it and not just give me these, but I made sure that this is only done when prompted. So it's not going to automatically use a lot of my credits to build these visual interactive features, but more so when I ask of it. So now five questions and check the answer. Let's say I have no idea what the answers are, and I got a zero out of five. So I could go back in here, be like change my questions, give me something else. So now I could use this feature to check for understanding. The only difference is that all these questions come straight out of my textbook. It uses that standard. Again, if you did syllabuses and that sort of stuff, it's going to be even more detailed, and now it has my first lesson. So I could come back every time I need a little tutoring, upload a lesson here, ask for an interactive quiz or no quiz at all. It's completely up to me because of the instruction that I provided. So here's another set of quizzes, questions that I get to do. Now, what you can do is do something that prompts clot design. So the little feature that we looked at. So we could be like summarize Lesson one and make a presentation. S, that's my unique learning style. It's going to run my skills if I had any. So I have one for PowerPoint. I think this is with Claude. It's going to add the task. We have different things in progress. You can see that. So now we have multiple folders running. We have three Python scripts, three progresses, and it's, like, a bunch of images that are being formed. I believe that's going to be like the diagrams used for learning. And of course, we have our context thing. These are all going to be saved, and as long as you sign into mobile with the same account, you're going to see this there as well. So this is going to take a little while, although I have song here, so it's not going to be that long. But another feature you could integrate here for this problem is the schedule tab. So let's have that run. You can see there's, like, a flickering light. We could build a new task set up manually, and I'll be like weekly quiz. So check my linear algebra. Understanding. So let's say use let's actually use the linear algebra tab right here, manual, frequency. I said weekly, so weekly. So use project context to generate a five question quiz based on student progress and weak spots. We could save this. You can choose a model. I think I'll keep SO at five and I'll have this be manual. Then we're going to do send the quiz via email. To the student. So hit safe, and this is going to happen every Monday at nine in the morning. We could even run it now. If you do in progress, closes. It started. Let's view. At project, there is our task. It's looking at all the things, and it found the lesson that we uploaded, doing a little script. And because this is weekly, it goes hand in hand with the project workflow that I'm building right now so that every week you could check in with me and see how well I'm doing. So, now this is doing its thing. We're going to go back to the previous chat, and now we have a little project presentation. There it is. You can see the visuals are really good and relevant. The design is clean, and I could take this a step further, open it in Cloud design or make some changes here. So, you know, we're getting all sorts of outputs with our tutor here. I think our task is done. So allow for all scheduled runs, create a new draft email. I'll allow it for now. I I go to my email, I just hit it over there. Let's do Allow connectors, the Gmail that we connected previously. Now it just made a draft. So I'm just going to check that real quick because I didn't say who to sender email was. Let's open up our email, and there it is draft, weekly linear algebra quiz. So here's the answer keys, here's all the questions, and this will be done weekly unless I stop it. So that's a really cool addition you get to do. It goes hand in hand with this guy. And I'll just change the name here as my like linear algebra tutor. There we go. I could come back to this every time I have a problem, test this three weeks later, four weeks later, and have this do everything in the background while I progress with my learning. So this was an example of a full project. We integrated the clot chat, the Cloud cowork, schedule task, skills, connectors, and even design. We use a bunch of the core features within clot, and we're able to build this very easy workflow where all I have to do is upload my next lesson. Have Claude look at it, build a presentation, do a little quiz, and help me better understand my lessons. So you just upload it here. If you have a screenshot of your own work, you know, anything you can upload it right here. And as always, you can change your instruction in the main project. So say your professor asks you to do another thing that you've not done before, you would specify it here, upload additional material for reference, and basically work with Claude as your assistant and have it build amazing workflows like this one. I hope you guys were able to either follow along with this example or build one for your own problem. And we're also able to test it with different things and alter the instructions if needed. As you saw, it's very straightforward. Claude does a lot of the reasoning on its own. We just need to give it some hints. And anytime you're stuck as to what to write for your prompt, you can have it ask you questions. So that's a really helpful feature that Claude does. It does like a little pop quiz for you guys, and it builds the prompt itself. So go ahead and try this for your next problems and see what sort of workflows you can build. 23. Class Project: Create Your AI Workflow: Now it's time to build your own AI powered workflow. For the class project, you're going to be using Claude to build an AI powered workflow that solves one problem in your life. This could be personal, educational, business or your work. Start by identifying the task that you need the workflow for. Then you're going to use what we have learned so far to build that workflow. Create a clot project so that you can organize this whole ordeal, then give clot some instructions via the skills, and if needed, connect external connectors so that your work becomes more efficient and it is tailored to what you need. Once your workflow is complete, you can go ahead and test it, refine it, and then see how well it executes. Most important thing here is that clot needs to be consistent for that task. So try it more than once and note of any inconsistencies, anything you left out, and what you can do to make it better. Once you're done, you can go ahead and take a screenshot of your project, the skills, the connectors you use, and a drawing of the workflow that you came up with. This could be a diagram, a chart, text, anything you feel comfortable with to explain what you did with Cloud. The engineer and designer of this workflow, you also need to state the problem you initially wanted to solve and how your workflow solves it. And the goal here isn't to just copy the workflow that we've done in the class. I want you guys to be unique as each of us has one need. We have different jobs, different tracks that we're on currently, and I want to see you guys fully explore this in your own way. So good luck, and I'm excited to see what you guys come up with with the project. 24. Congratulations! What’s Next?: Congratulations on completing the course. You now know how to organize skills with projects, build skills with Claude, and use the connectors feature. Remember that the best workflows evolve over time. So come back to that project, see what needs changing, and if your needs are now different. The good thing about project is that it's going to remember your instructions, and you can always come back and edit them if needed. If you haven't already, make sure to upload your class project screenshot and information to the gallery. I would love to see what you guys came up with and how you use the tools. Thank you for joining me for this class, and I hope to see you guys in our next.