ChatGPT Mastery: Boost Your Marketing Strategy And Your Business Growth With ChatGPT

1. Welcome to the Class: If you're watching this video, you probably are an entrepreneur, freelancer, marketer, or business owner, and you probably have heard of charge EBT or you maybe are already using it. The problem with charge but t is that you get great results only if you write great prompt. So basically you need to know how to ride a good prompt in order to have a good outputs. But how can you improve your prompt if you have no experience whatsoever? Well, that's why I'm here. Chow, My name is Ralph, and in the last 15 years, I've been helping companies and freelancers grow their businesses through data driven experimentation and creative thinking. I've had the opportunity to work with huge brands like Vodafone, HP, Microsoft, PNG, and many more. In this course, I'm going to show you how to use charge UP T for marketing strategy and business goals. This course is divided in four modules. In the first module, I will explore common mistakes made when using GPT and how to avoid them. In the second module, I will dive into advanced use cases of tragedy. In the third module, I will share with you some tips and tricks for optimizing your use of creativity based on my experience. Finally, in the fourth module, I will showcase some real life examples on how you're going to charity in your day-to-day work. By the end of this course, you will be able to use GPT-2, improve your marketing strategy and to reach your business goals. And if all these sounds good to you, I'll see you inside. I also want to say that you can download some extra resources related to this course. Just visit guides dot link slash GPT m. The link is here and in the description of this course. And you will have access to a checklist to improve your prompts and also a list with hundreds of free prompts that will blow your mind. 2. The Project: I'm super excited that you decided to wish this course. Let's talk for a moment about the project. As a student of this course, you will be part of a community of like-minded entrepreneurs, freelancers and marketers and business owners who want to learn more about IGBT and improve the way they use it every day. I would love to hear about your prompts, the one that you are currently using, and also the ones that are working for you. Sharing your experience with the others will not only help you, but also give you a different perspective of how to improve the way you see end-use GBD every day for this class project. Share with me and with the other students your best prompts and explain the purpose of each prompt. Tell us how you came up with idea of that prompt and how you use that prompt to optimize your work. I know it can be hard if this is your first experience with charge a pity. Don't worry. If your first prompt look too simple. No pressure at all. Just something when you are ready and if you're not, no worries. Just take a look at the other projects and get some inspiration for you. Now, let's get to work. 3. 1.1 Common Mistakes in using ChatGPT: Welcome to the first module of this course. This is where I will explore some common mistakes made when using charge BD and also how to avoid them when using GPT. It's really important to remember that it's an AI language model. And as such, it can only provide responses based on the information that's been trained on. And one of the most common mistakes people make when using THE beauty is not providing clear instructions. E.g. if you ask something like, what is the meaning of life without providing any additional contexts, the response you will receive may not be what you're expecting. This is because Jupiter, he needs some specific information in order to provide a relevant response for you. So if you want to get the most out of Egypt, it's really important to provide clear instructions and also as much context as possible. Another common mistake people make when using GPT is Assuming that is always right. Well, that's not true. While GDP these incredible advanced and can provide also some accurate responses. In many cases, it's important to remember that it's not perfect. So if you receive very sparse then doesn't seem quite right. Don't be afraid to double-check the information with another source. So by providing clear instructions and being mindful of Jupiter limitations, you can get the most out of these incredible tool. In this module, I've collected some of the most common mistakes for n for each one I will share with you all. So some examples of the improved prompts, I'll see you in the next video. 4. 1.2 Poor Instructions: One of the most common mistakes when using Church BT is to give poor instructions. If you don't give, if you don't provide clear instructions to diabetes can lead to a variety of negative consequences like poor performance is incorrect. Outputs. Of course, wasted time and resources. Why? Because without clear instructions, check GBT may not be able to understand the context or the goal of the prompt. And this can lead to outputs that are most of the time irrelevant or incorrect. So let's see an example now. Let's say that we have our blog about dogs and we want some help to generate an article. So the worst way to do that is with something like this. Let me copy and paste the prompt here. So to tell GPT generate an article about dogs, as you can see these like super generic. There is no, no context at all. We are not saying one kind of article. We need what we want inside the article, and so on, so forth. So I'm going to stop this here. Just, I just wanted to give you an idea how we can improve this. So let's see an improved version. How all the prompts ready here. So I'm just going to copy and paste these. So an improved version could be something like generate an informative article that describes difference, different breeds of dogs and their unique characteristics with a target audience of first-time dog owners. Use at least three sources to support your writing and aim for a length of 500, 800 words. So let's see the difference now is, as you can see, there is a target audience in mind here. This is really important because we are writing for people that aren't having a dog for the first time. We are telling the length of the articles or 500, 800 words. And also we want to focus on the different breeds of dog and their unique characteristics. And as you can see from the results, I'm going to stop this here. And I want to show you the whole article. I think it's going to be quite long, but you can see, you can already see the difference between the first one and the second one. I'm going to click just stopped generating here. So first mistake, please be careful. Try to always provide clear instructions to change everything. 5. 1.3 Incorrect Input Formats: Another common mistake that users make when providing input to charge if D is incorrect input formats, what does it mean? Some example are e.g. providing incomplete sentences or using an inconsistent language, or sometimes also including irrelevant information into the prompt, e.g. if the input to change apathy is just a list of keywords rather than a sentence. Most of the time that's useless because the model may struggle to generate relevant output. And also if we include some inconsistent languages, e.g. with different terminology charge APT may not be able to understand the meaning and the contents of the prompts. So to avoid these mistakes, It's important to always use a clear and consistent language to provide complete sentences and also to ensure that the input is relevant to the task that we will not happen. So once again, let's do the example of the dog breeds. The wrong way to ask Judge apathy these information is to just write dog breeds into the, into the prompt. So as you can see, this is just a list of dog breeds and nothing else. So let's stop here. Improved version of this brand could be something like this. Let me copy and paste. So generate a list of the list, five popular dog breeds, along with a brief description of each breeds, temperament, and characteristics. So let's see the output here. So now we told that we would just want five dog breeds, but we also want some extra information. So the descriptions are brief descriptions and also some characteristics and temperament of the breed. And as you can see, the output is completely different from here. So please always be careful and try to use clear, a consistent language when you write your prompt. 6. 1.4 Not Fine-tuning for Specific Use Cases: The next mistake is not fine tuning charging for specific use cases. And it's quite common when users try to use some. Fine tuning is the process of training the model on a specific set of data or task to improve the performances and to improve also the output as well. So if we do not find during the model for a specific use case, the result can be an outward that is completely irrelevant for us. So let's see, once again an example. So a bad parent could be something like this. Write a short story. This prompt is too general, doesn't provide any specific contexts. Contexts, and not even some requirements for the soy. Which activity could generate any type of story which may not be relevant for us, not appropriate for our specific needs. So I'm going to stop this year. Let's see the improved version. So the improved version could be something like that. Let me copy and paste these here. So generate a science-fiction short story that takes place on a decent planet and involves time travel. The story should be between 500, 1,000 words and include at least two distinct characters with unique traits and motivation. So as you can see in this revised prompt, we have pride some clear and specific requirements for the output. So we told Judge apathy that it should be a science fiction story. On a distant planet with time travel, we want two different characters. So as usual, I'm going to stop this here. But I just wanted to show you difference once again. And let's see something more work-related example. So a bet prompt could be something like this. So could be asking, how can I improve my sales? This prompt is too vague. Once again, doesn't provide enough explicit specific information for GPT-2 generate our relevant output. So it's unclear what kind of sales we're referring to and also what kind, what specific aspect of the sales we are interested in improving. So I'm going to stop this here. Let's see an improved version of this prompt. It could be something like that. Generate a list of at least five effective strategies for increasing, increasing online sales of a software product. Including techniques for optimizing the website, creating product description, compelling product description, and using social media to drive traffic. So in this version, we have provided some clear and specific requirements for the output. So we told the GBD that is a software product. And also we gave Egypt some ideas of what kind of aspects we are interested in improving to e.g. using social media, optimizing the website and creating compelling product descriptions. 7. 1.5 Not Iterating: The next mistake is not to iterate on the output. And I think this is probably the most common one. What do I mean by that? Most of the time people just ask tragedy for something and they stopped the whole process there. When the first output that I get. So let's see an example. Let's suppose that we are writing a blog post about email marketing and we want some ideas for the title. So generate five titles for a blog post about email marketing. So this is just the first step. When you ask charging money for something. Always suppose that this is the first time you have to iterate to get a better output. So e.g. let's say that we don't like these five titles here, so I don't like them. Give me five more. So one of the first idea could be to ask for more output. So let's say that the first five ideas for the blog posts, blog post, we don't like them. So we asked you to give us five more ideas, or e.g. we can say, now, make them for the B to B market. Because maybe we work for just for B2B companies. Our blog is just for B2B companies. So we want to write a blog post that is for B2B markets. So something that we could do is to tell GPD this and he can, it can improve the output actually, or something that we can do. We can say, I don't know, I don't like the third one. So account-based marketing, I don't like number three. Give me an alternative. We just, we're just going to change number three. And let's say that we like this one beyond the inbox, innovative B2B email marketing tactics for customer acquisition and retention. But let's say that it's too long, so make it shorter. This is something that they do all the time making shorter, make it longer. So we went from beyond the inbox, innovative B2B email marketing tactics for customer acquisition and retention to innovative B2B email marketing tactics for acquisition and retention. So yeah, this is some ideas of what you can do, but always, always, always iterate on the output. Don't stop on the first one because you can get a better version. Always. 8. 1.6 Not Providing Examples: A very common mistake is not providing IGBT with some examples and some custom elements. I know that charge if it is great for brainstorming ideas, I use it a lot. But instead of starting from scratch, I will suggest to always add some ideas, your custom elements into the prompt. Let's see an example. So let's say that we want some help to generate an email to reach a prospect after showing the demo of our automation marketing software. So I really bad prompt is something like that. So let's see the output here. So of course, aegypti is going to write the subject of the e-mail and they're the the copy in the body of the email. So yeah, this is quite general. I hope this demonstration was informative, blah, blah. Our software is designed to help business like yours, blah, blah, blah. I love to schedule a follow-up, discuss the software, blah, blah. Thank you, Mr. regards. But let's say that we have some personal information about the meeting, the demo. So let's say an improved version. So we can write a prompt like this, generate an email to reach a prospect after showing the demo of our admission marketing software. And this part is the same. But then we say, Here are some elements to add. The demo was on Friday. So we want to say in our email that the demo was on Friday. We had some very good Italian food with our prospect and also the name is Mario. So let's see the output now. This is going to change it a little. Following up after a great demo and lunch. The subject is already a different one, DR. mario. So now it's custom-made. Of course, this is just an example, but imagine if you have a lot of information that you want to put inside the e-mail. So I wanted to talk to you for taking the time to meet with us and font Drag for lunch. I'm glad you enjoyed the Italian food. We believe in building relationships with our clients and sharing good food is just one way we like to do it. So we started from the idea of having some Italian food and we went to building relationships. That's why it's powerful in my opinion. Blah, blah, Follow-up. This part is completely the same. An example, just a very simple one, but imagine that you have, I don't know, some numbers you want to include inside your output, e.g. during the demo, you mentioned, I don't know, our price for the for the software or you mentioned I don't know, like a trial a couple of months of trial, or you mentioned to schedule a follow-up appointment on a different date. You can put all this information into the input and you can get a better outcome. So always provide GPT with examples and some custom elements is going to make a huge difference. 9. 1.7 Not Include Requirements: I see a lot of people writing prompts without including any requirements and constraints inside the pronged, actually including the requirements is one of the most important best practice for ensuring that can generate some output that are specific for your needs. Let's see an example. I'm prompt, without any requirement could be something like debt. So generate our recipe for lasagna. So again, super broad, super generic, we're just asking for a lasagna. But actually we can improve this prompt and we can add some requirements. I'm going to show you how I usually do that. Let me copy and paste a prompt here. I'm going to stop the output just to show you the difference. So let's say that we actually want to vegan lasagna. So generally the recipe for a lasagna, this part is completed the same and then I add an extra part of the prompt with requirements. Let's put this here so it's easier to read requirements. So the recipe has to be vegan, should not include any meat, dairy, or other animal products in also choose, it should serve at least four people and take no more than 2 h to prepare and cook. So those are some requirements and also constraints. So let's see the difference in the output. Now. Of course, the most important difference between this one and the first one is that this one is vegan. So the ingredients are going to be different. And also, we asked to do something for four people. So the amount of each ingredient is gonna be completely different. And we already can see this from here, from the first version here, in the second version here. So I'm going to, once again, I'm going to stop the D output, but we can already see vegan ricotta cheese, vegan mozzarella cheese, and some different ingredients in this version here. Of course, this is just an example, but once you know that you cannot requirements and constraints, keep this in mind, especially if you have special requirements, e.g. in your industry, or some special constraints that you want in your output. So do that and you will always get a better output. 10. 1.8 Not Providing Feedback on the Output: The last lesson of this module is about feedback. It is very important that you provide feedback on the output because charge OPT is always learning and you will see a huge difference if you keep providing feedback on the output. Let's see an example. Let's suppose that we want to write an article about influencer marketing. So let's start with the classic one. Generate five titles for a blog post about influencer marketing. So the simple version of the prompt, this, this, this lesson here is quite similar to the one where we spoke about, you know, keep iterating with charging P t. But now the way we iterate is very important. So you can give feedback using those two little buttons here. So thumbs up, thumbs down. But I always like to use the prompt to give feedback. So e.g. let's say that those five ideas, they are too spammy. They sound too spammy to charge. It. Gets the information and gives us five different ideas. So I'm sorry to hear that the title seem spammy. Here are some alternative options that muscle more informative. Way better, especially the first one I loved the first one. But now, let's say that we still don't like these five ideas, so they sound too formal. So now tragedy is going to make the titles less formal in a more engaging and conversational tone. And as you can see now is from TikTok and Instagram navigate the word influencer marketing, The Good, the Bad, and the Ugly. I love this one. Number three is, is quite cool. I love this one as well. Behind the filter, I can look at the reality of influencer marketing. So we started with is keeping this information here. So we don't want something that is too spammy, like those clickbait title. But also we don't want something that is two formats, so we want something that is informative but also conversational. And you can keep doing this until you get the results that you want. So I always like to use also the two buttons here to say that I like or I dislike this one. But if you write down in the same chart, tragedy remembers what you have done before, what we're seeing before. And it keeps, as I said before, keeps learning. So you'll see a huge difference and the output is gonna be so much better if you give judges always feedback on the output. 11. 2.1 Advanced Use: Welcome to the second module of this course. This is where I dove into some advanced use cases of Jupiter. This module is gonna be very useful even if you have already some experience with charge apathy. I will show you how to give judges pithy more specific information to work with, which can help provide more accurate and relevant responses. And that will also share with you some of my favorite prompts, tactics and also tricks to improve the experience which IGBT at the end of this module, you will be able to fine-tune your prompts to get better results. And you will also be able to provide the context to use it on your own content, to set a tone of voice, to use placeholders and much more VCs. My favorite module is full of hidden gems, I promise I'll see you in the next video. 12. 2.2 The Prompt "Act as": Example of this module is the prompt act. As the prompt Act does is used to Instructure GPT-2 generate texts as if it were a particular person or character or even an entity. And by providing a specific in-person to the GPT, it means that you are providing a set of characteristics, beliefs, and also behaviors. And if you use the actus prompt, charging party can be trained to generate text that is very tailored to your specific needs and also interests. So let's see an example. So let's start with a prompt without the actus. Let's say that we are writing a book about a novel about time travel. So generate five ideas for titles for a book about time travel. And this is the simple version. Now, I want to use the style of Stephen King. Let's say Stephen King, I love Stephen King's act as Stephen King. And j array five titles for a book about time travel. So let's say your difference. The clock makers curse the timekeeper lemons, temporal terror yesterday, nightmare. You see the difference now? Yeah, I think you get it. Of course, this is just a simple example, but imagine that you can use a real person or also a character from a book, from a movie, whatever. But also we can use this approach. So the actus prompt also in a more professional way. Let's say that we want to write some tweets as CEO from corporate company and write five tweets about leadership. So as I said before, once you say to GPT, once you are providing a specific persona, in this case, a CEO from a big company. It means that you are trying to emulate the tone of voice, those set of characteristics, beliefs, and behaviors. Because of course, you know, our CEO from a big company probably has a different approach and also tonal voice from a CEO of our really, There's more company or a startup or from a freelance, whatever. And then also going to share with you a very useful resource. This one is completely free. I'm going to put the link in the description of the course. This is on GitHub and it's a list of prompts. They all use the actus approach as I showed you before. I suppose this is probably 100 or probably even more. It's a very, very long list. So every time you want to use this approach, you want to use the actors prompt. Go to this list and there are always nice ideas. And most, some of them are also very funny. So you have stuff like life coach, personal trainer, What else? Investment manager. I don't know. Adventure game, title generator. And there is one that is super funny. I'm going to show it to you. Let me, let me find it. Was here. Yes. This one, I love the Archie artists. So you can say to judge a duty to act as an artist. The artist is someone who use just characters, draw something. So e.g. here, we want to draw a cat. We have telling GPT-2 drug cut only using ascii characters. You can see the results. It's super funny. Of course, it was just a funny example, but the list is quite huge and it's full of interesting examples. So keep this in mind and always use the act as before. I'm prompt if you went to an output that is like super tailored for your needs, for your interest. 13. 2.3 Providing the Context: One of the most important aspect to keep in mind while using charges, but t is contexts. We saw a bit of that in some of the previous videos, but I want to go deeper now, if you provide a context, your output is going to be so much better. But because providing context to try jeopardy. Providing context switch IGBT is important because it helps the model understand the specific tasks and also the information that is being requested. Without context, GBD may not be able to interpret and respond to the prompt and the contents can include details such as the Toby, specific questions, background information, and any relevant details that will help GPT provide a useful and accurate response. So let's see an example. Let's say that we want, we want, we want an idea for a video, YouTube video about charge between. A bet prompt can be something like debt, generate our tidal for a YouTube video about GBT. How can we provide context today? So I have an example here. We are going to use also the actus prompt that we saw before. So act as a YouTuber. You have a YouTube channel about the marketing and business. This part of the contest, marketing and business are two important keywords here. Your audience is mostly made by small business owners. So now charging, but he knows that we are talking to small business owners. So mostly entrepreneurs, but small business. And also generate tidal for a YouTube video about using chatty BT in a small business. So it's very narrow, very specific. Let's see difference now revolutionize your small business, which IGBT the ultimate tool for streamlining, streamlining customer service and boosting sales. And you can see the difference with the first one. So always provide context. You can use the actus as a starting point, but I always suggest you to keep in mind the audience and also to give some information about the output that you want. So in this case, the output was a video about using GPT in a small business. So let's see a different example here. Let's say that we want this for freelancers, e.g. so your audience is mostly made by freelancers and also generate a title for YouTube, I usually charge EBT. So now the audience is different and the title is gonna be different. How freelancers can use GPT to automate client communication and increase productivity. So here, the focus was on customer service and mostly boosting sales. Now here is about increasing productivity. Why? Because if you are a freelance, that productivity is an important part of your day-to-day life. So yeah, always give charging party, always provide the contents and you're gonna see the difference. 14. 2.4 Use it on your Content: Not a lot of people know that you can use charging party with your own content. What do I mean by that? When you are writing prompt, you can add into the prompt some of your content or someone else's content. Like, I don't know, e.g. a. Competitor. So let's see an example here. So I open this blog, post this from HubSpot blog. This is just an example, but you can use it on your own blog post. And this is like marketing without a budget. Use these tactics. So let's copy and paste one of these. Let's start with the first one, e.g. so I'm going to copy this part here. The first, I'm cheap. It's about developing our content marketing strategy. I'm going to say to check GPT-2, create a summary. So create a summary. And here I'm going to pass the content. Let me remove the URL from here. Of course, there are some limitation about the amount of content that you can put in a prompt. And we're going to see that in a couple of videos. But for now, let's focus on the example here. So this is quite useful. Of course, if you have very long article or blog posts and you want a summary, maybe for you to study it, or if you want to repurpose part of your content, maybe for a blog post, to a newsletter or to a LinkedIn post or something else. And also something that I also like to do is to create bullet points. So create a bullet list from the contents. So as I said before, this is a great way if you need to report from some content, e.g. you want to start from a blog post and then you want to create a list of, let's say, tweets, e.g. so a small short texts for your Twitter account or I don't know for captions on Instagram or I don't know, post on LinkedIn. And also, you can do the opposite. So we can start from a sentence. So let's copy and paste something interesting from here. Let's see me something that is nice. Blah, blah, blah, I'm giveaways. User-generated content was the webinar. Webinars or a grid and cheap way to educate your customer across various stage of the funnel. So let's copy and paste here and we can ask Joe GPT to expand. Oops, let me remove this from here. So we start from a sentence and then we ask creativity to create something bigger, longer, like a blog post, newsletter, copy for a YouTube video, whatever. So yeah, I mean, this is just an example and I use someone else's blog, in this case, HubSpot, but you can do it on your own blog. If you knew a lot of content and you want to repurpose your content in time. But you can also use your competitor to get ideas. If you want to. Just have a summary to study something that is a very long content. Whatever. I'm gonna stop this one here. But the takeaway here is always remember that when you're writing up a prompt for OPT, you can use your own content inside the prompt. And this is quite useful. And not a lot of people know about this. 15. 2.5 Tone of Voice: One of the best part of working with charge a beauty is using the tone of voice. Not a lot of people know about that, but you can actually analyse tone of voice, which IGBT. You can recreate the tone of voice. You can do a lot of cool stuff with it. So let's start from the first example. So I'm going to show you how you can analyze a tone of voice. And this can be your own tone of voice are also someone else's tone of voice. So I have the same blog post that I used before here. So this one, let's copy and paste this paragraph about content marketing strategy. So let me copy here and say to analyze this tone of voice, I'm going to copy and paste this year. I'm going to remove URL from the end. So this is quite useful because you can do it on your room tone of voice. As I said before, if you have your own content, I don't know, a blog post and newsletter or something else, or you can do it on someone else's content, e.g. your competitor. And you can ask to GPT to recreate the same tone of voice in order to create some new contents. So in this case, e.g. the tone of voice in this passage is informative and helpful. The speaker seems knowledgeable and confident in the information they're providing and is eager to share it with the reader. They use phrases like according to solve it in 20 2023, marketing trend reports to establish credibility and to provide evidence for the claims. Is optimistic. The speaker highlights the benefit of content marketing. I love this, I love this. I mean, it's great. There is a sense of enthusiasm. And in Seidman, now, we can say that we won't use the same tone of voice to generate some new contents. So let's say use the same tone of voice and generate paragraph about this one was about content marketing, if I remember right. So, yeah, conduct marketing strategy. So let's say we want something similar about email marketing. Here we are. So now charge if D is going to use that tonal voice. So informative, helpful, confident, knowledgeable, optimistic, blah, blah, blah, everything else. So as you can see here now, we have some new, fresh content using the same tone of voice but on a different topic, email marketing. And this is great because you can use this to replicate your own tone of voice when you are generating contents. So first, lecture, GBT, analyze some of your previous content in order to learn all the aspects of your tone of voice. And then asked IGBT to use the same tone of voice to generate content. But there is something else that we can do is that we can also change the tone of voice. So e.g. let's, let's go with a simpler example. So let's say that we won't generate five titles for the, for this content. Okay, So we just wrote a content about email marketing and now he's doing about this one. Okay, yeah, it works. So now we have five titles and now I'm going to tell Joe GPT-2 use a different tone of voice. So generate five more Titles with our formal tone of voice. Because I don't know if for some reason or we want something that is more of a formal approach, maybe I don't know because it's our style, because of the audience we talk to. So as you can see here now, the tone of voice is completely different. But we can do, we can do whatever we want here. We can say, generate five more titles with funny tone of voice. Tone of voice. So now everything is gonna be less formal. More funny approach. I like this one. Why you don't need to flow the Internet with crappy content. I like it. How to video content marketing strategy after midnight. I love it. From boring to binge-watch it actually, they are quite good actually. And also, I mean, you can do whatever you want. Let me show you a more extreme one. So generate five more titles. Weeds, Let's say direct and spammy tone of voice. I wouldn't do that. I mean, of course, but it's just to show you that. Oops, it says that it cannot do it because it's against ethical guidelines. Sorry for that. Let's keep it. Just direct. You are going to receive a lot of these messages when you use GPT. So with that direct, now, let's say that we want to generate five more titles with a direct tone of voice. Okay, I removed the spammy into the prompt. So now everything is more powerful strategies, rise to the top winning content marketing strategy. I mean, if you like this style, if you like this approach, if you like this tone of voice. Why not? Yeah, this is it. I mean, now that you know it, you can use a lot of different tone of voice. And also, what I like to do is get this asked LGBT for a list of tunnel voices. So give me a list of 20 different tone of voice that I can use in my content. So now you have some ideas and you can tell charge a PT to recreate some of that. So if you just need some objectives, is always a good idea to ask Djibouti for the adjectives and then use the same adjectives in your prompt. So maybe you want something, I don't know, educational, but also inspirational. A great combination. Or you want something that is direct and the persuasive. So give me five more titles that have inspirational, inspirational, and let's say also casual tone of voice. We are mixing now the objectives, small steps, big impact. I love this from good to great is a great book. Actually. The future is now, blah-blah-blah. Yeah. I mean, you get the dream, right? So once you have the list of objective, you can mix them to create the perfect tonal voice for your content. 16. 2.6 Prompt Size: Something really important that you should know about charging beauty is that there is a limit of 5,000 tokens that you can use uncharged GPT. And be careful because these limits includes both the prompt that you are writing to charge a, B, D, and the answer that you get from charge UP team, how can we calculate exactly the tokens? Fortunately, there is official tool from OpenAI is called tokenizer. And in this tool, you can copy and paste your text here. And this tool will tell you how many tokens you are actually using. So as we can read here on the bottom, one token generally corresponds to four characters of texts. For common English texts, this means that more or less, if you use 100 tokens, means that you are using 75 words. Let's make an example. So if we copy and paste this, the Wikipedia page of GPT, Let's copy and paste the whole page. And here, copy and paste. Here. We can see that the whole page is actually 5,400 tokens. It means that if we use this content here in charge of beauty, we are gonna get an error. I'm going to show you this. So let's say that we want to write make summary, and we paste this whole text here. Jupiter is going to tell us that the message you submit it was too long. Please reload the compensation is something shorter. So if you use something that is too long, charged with is going to give you an error. But even if you use something that is shorter, Let's say that we can remove something from here. So let's get 3,900 targets. Remember, 4,000 is the total between the texts that you write in the prompt and the answer that I chair GPT gives you. So let's copy this one here. And let's try again. Make bright summary. This is too long, so you can't actually do it. So what's the solution here? You can use this tool called GPT splitter. And what it does is that it splits the counter. You are giving an input to charge, you put D in different chunks, and then you can actually load the single chunks inside charge OPT with a different prompt. So let's see an example once again. Let's copy the whole text here. And let's say process. This tool here just created four different chunks, 123.4. So now you can copy the single chunk in there and it's going to work. So let's add something here. Let's try with. So now let's copy and paste these into GPT. So the first one, let's open a new chart. In the meanwhile. Tragic got the first one. It's copied the second one. Now, let's copy the third one. Now, let's copy the last one. Now, we asked you to write a summary of the texts that we just provided. Now we have the summary. So that's how you can first check if your prompt is too big. So if the content that you are writing that you are loading into charge, if it is too long. Also, you can split the content in order to have different chunks, different prompt that you can copy and paste inside GPT. The very nice thing about this tool is called GBD splitter. I will leave the link in the course description is that you can paste a text here, but you can also upload a document here. So if you have a doc file, make with Word or PDF or something else. So let's go back here. We got the summary. 17. 2.7 Split the Input: In the previous lesson, we saw that we can actually give information to judge if the t using chunks of texts. So instead of giving everything in one prompt, we can actually split the information in several prompt. But we can use the same approach. Not only if our prompt is too big, if the text is too long and if we have the problem of the 4,000 tokens, but we can use the same approach if we e.g. if we don't have all the information at once, we can kind of use a Q&A approach with charging for t. So we tell Jim GPT to wait for us until we have all the information before giving us the result, before giving us a response. So let's see an example here. I prepared a prompt here, so this is a problem. Then we can use, I'm going to write five words one-by-one. Every time I write a word, you reply with word number or K word. Don't write anything else. So it means don't wait for me, wait for the next word. When the word number is five, write a blog post about GPT using the five words. I just passed the you say I'm ready when you read it. So what does it mean? It means that with this approach, we can, as I said before, split the input and we can, instead of giving charge of t, the five words all in once, maybe we can add a no. Think about it. I don't know. We are working with a team, we are brainstorming the information or for whatever reason you may have. So now I'm going to wait for the IGBT. Okay. Judges with you just say that it's ready so I'm reading, please give me the five words. So let's start with the first one. So let's say future. Okay, So we told charging party to give us number of words in this case is one. Okay. And then the word in this case is future. So let's go with the second 120-203-2020. Got it. Let's say that is a blog post about business somehow. Then we want to talk about money. So the impact that tragedy is going to have the monocyte. And then let's also talk about time. Okay, so now judge it, but it has all the five words. So now he's ready to write the blog posts. So bright, the blog post here at Blockbuster using the word future, 2023, business, money and time. So let's see 2023 seer businesses here. What else? What do we have? Money, time and future? Papa, papa, yeah, somewhere in here it should be also monitoring time, money and probably future. I don't see future now, but it should be here somewhere. Future is here, okay, here we go. So this is a nice trick. Not a lot of people know that they can actually split the input while using tragedy, but it's quite powerful. Of course, this was just a simple example. But imagine how many situation, you know, in how many situation you can actually exploit. Exploit this, this approach. 18. 2.8 Using Placeholders: So in the previous lesson that a pass some information to charge if d using brackets, whether you can actually use brackets and braces to create placeholders. Placeholders in Georgia, PTR, like text strings that are used as a temporary substitute for our specific piece of information that you are going to provide. You know, or another source is going to provide this placeholder. It can be used in various contexts, e.g. questions, request statements, and they serve as a flexible way to communicate. Communicate which IGBT and placeholders are powerful because they enabled users communicating with judged in a more efficient and personalized way. By using placeholders, you can ask specific questions. You can request a specific actions, or you can share personal messages without having to provide all the details upfront. And this makes the composition more dynamic and interactive and also helps charged with you to better understand and respond to your needs. So I'm gonna show you some example. The first one. So now I'm going to use this one. It's a mix of using the placeholders and also providing splitting information I provide to tragedy. So I'm going to write some content in the format name, placeholder, company. And you will reply with hello, name, I love company. It's a great company until I write stop. So let's try this one. So now we can say Mark. Microsoft knows that. Instead of name, now we won't mark and instead of company now we want Microsoft and Amazon. Hello Hannah, I love Amazon. It's a great company. Jane's Open AI. Hello James, I love OpenAI is a great company. And of course this is just an example, but you can use these to create very powerful prompts in order to give jeopardy a lot of information, input information. Another use case of this one is let me copy and paste this one here. What is the capital city of country? So we don't say the country to Deputy, we just say country using the brackets. And this is a placeholder, which now is going to tell us which country you are curious about. So Italy, row, great, France, Paris, the US, Washington DC, and so on, so forth. You get it. Another example I want to show you is that we can do something like that. So even more powerful, we can say to charge a pretty, write me an email. I want to show them of my marketing automation software to the prospect. The email has to be 50 words. Stop, use name as a placeholder for the project name and company, for the prospect company. Don't use other placeholders. So now charge if it is going to write is called e-mail. And we have company here and name here. And this can be copy and paste inside your e-mail marketing software or inside your CRM marketing automation software. Or we can tell to GPT to write us the whole text again, but with some information that we are going to give any input. So e.g. rewrite the same email five times for people called Mark James, Janet, Anna, and Tom. And the company is always Microsoft. So let's see the result. So charge if it says here are five variation of the email. Hi Mark. I notice that Microsoft blah, blah, blah, blah, blah. Hi James. I notice the Microsoft bla, bla, bla hygienic. I, Hannah, Hi Tom. So again, once again, this is just an example, but once you know that you can use braces and brackets to create placeholders. This is a powerful trick and you can use it in a lot of different situations. 19. 2.9 Criticise mode: Another advanced use of charge if the T that I want to show you is the criticised mode. Sometimes when I use charged with t, When I asked to repeat it for some content, I also ask it to tell me what can be improved and why. That result is not really great. It's kind of a nice trick if you want to, if you want some suggestions of how to improve a response. So I'm going to show you a couple of examples. So the first one, the first one is this one, act as a marketing expert and give me five ideas for a YouTube video about charge EBT in small companies. And then tell me, what do you not like about those ideas? In the same prompt, I first asked charging for some ideas, but they also ask GBT to have more information of why those ideas are not great. So let's see the result. How to Djibouti is revolution in customer service, productivity, marketing strategy, real life example, baba blah, customer support. What I don't like about this idea is that they all focus on benefit of, of charge between small businesses, but they don't explore any potential drawbacks or challenges. So if you are actually recording a YouTube video about, I don't know, customer service, we could talk about that. So we can say that it's quite greatest charity in our customer service, but there are also some drawbacks and challenges and we could talk about those or e.g. it's important to acknowledge any limitation or concerns that a small business owner may have about implementing charge EBT. And also some ideas may be too broad and not specific enough to be engaging for viewers. So e.g. we can start with this one. But instead of being generic for small businesses, we can say small businesses in B2B markets, e.g. or I don't know, small business in the fashion market or I don't know the food market, whatever. Another way of using the same approach is this one. So I first give a simple problem, like give me a title for a blog post about GPT in small companies. And once I get the result, I do a second step where I asked charity to act as an expert, in this case, a marketing expert, and tell me why the title is not so great. Let's see. The title. The original one was five ways tragedy can help small companies improve customer service and boost productivity. So typically starts saying that the a solid tidal clear consists, blah, blah, blah. I will suggest to make the title more attention grabbing or unique. So here we have some examples, e.g. like how ceviche if D is helping small companies overcome customer service challenges and streamlined operation, or e.g. using a question. So it's charging the customer service solution your company needs. So we have the first idea, and the idea remains the same. So about small companies, customer service and creativity. But we also have two or three extra ideas in order to improve the content. And this is great. You can use the criticize more. Every, every time. I mean, every time you ask for something, I also strongly suggest you to use the criticize mode and have some more information in order to improve the output that you got. 20. 2.10 Let ChatGPT Improve your Prompt: The last lesson of this module is my favorite one. It's quite easy actually, but it's super powerful and very few people use this approach. You can actually ask GPD to improve your own prompt. Yeah, you had the right. Let me show you an example. So let's say that I have very bad prompt lot, something like write me a blog post. Now, you know that this is not the way you should write a prompt, but let's say that we want to ask the GPT, how can I improve this prompt? So how can I improve this prompt? We asked LGBT to have some suggestions in order to improve the prompt. So let's say what does suggestions are. So e.g. give me the topic. Provide a specific topic or theme, the audience. So give me also the audience. Are those beginners, experts or particular demographic group. The tone, the length, the style guide, blah, blah, blah. This is super harpoons, super-helpful bar. We want something more. So let's ask the GPT to actually generate five examples of the prompt in the improved version. So now we know that we should add some topic and audience and tone and lengthen also style. But let's see what we get. So the first one, those are great. So topic, audience, and tone, length. Now we have all the information we should provide the GPT about this prompt so the problems can be read me a blog post and then we add the topic with the topic, the audience with the audience, the tone with its own length. When we copy and paste this one. And that's it. And you can do these before any prompt. So before actually giving the prompt, which LGBT just ask it to improve it a little. Then also ask to IGBT to see the improved version so you can learn while using it. And this is great. 21. 3.1 Extra Tips and Tricks: Welcome to the third module of this course. This is where I share with you some tips and tricks on how to use GPT based on my personal experience, GPT is a powerful tool for generating texts, but there are many additional ways you can use it to enhance your productivity and creativity, e.g. you can stretch it pity for generating hashtags and emojis for social media posts by including relevant keywords and phrases in your inputs. Another technique the dialogue is to use judge pithy with brainstorming frameworks like scamper or the sixth thinking nuts, to generate new ideas and explore different possibilities for formatting the output of GBD is also very important, especially if you are using degenerate the text for a specific purpose. You can use formatting techniques like, uh, Bolding, underlining it Alix, to highlight the key information and makes the text easier to read. And these are just a few examples of the tips and tricks that I'm going to share with you in this module. I'll see you in the next video. 22. 3.2 Hashtags: I love using charging for T, for hashtags. I'm going to say, I'm not very good with hashtags to pick those. So I always ask GBD to give me some ideas, some suggestions for some hashtags. You can do that in several ways. So the easy one, it's just something they did generate five harsh our stocks. I don't know for a YouTube video about, I don't know, holidays in Brazil. And we have those hashtags. I also usually ask distributed to write the hashtag specific language, e.g. if I need the content to be in Italian, I want to use Italian hashtags, of course. But what we can do is also, we can say judge APT, what kind of content we are going to write. So e.g. I'm writing linkedin post about let's say leadership. Please. Give me five harsh dark. For D. Corp. Rate. Word, corporate, B2B word. Let's be more specific here. So now we have leadership development, corporate leadership, between leadership, business leadership, leadership skills. I love being very specific when I asked her GPD for some hashtag because I get better our stock. We can also as GPT-2 give RSpec for Twitter, for Instagram, e.g. would are the best hashtags to use on Instagram. If I'm posting a picture. We'd motivational quotes, Motivational Monday, inspiration, motivational quotes, quote of the day, daily motivation. And we can also do something like this. So we can say, write me five motivational tweets and odd. Some harsh dark. So when you are actually creating, when tragedy is actually creating the content, you can ask Judge be T to what the hashtag in real time. So this is for some motivational tweet, but if we are writing a post on LinkedIn, we can do the same. So write a motivational poster for linkedin. Has to be, I don't know, 50 words top or some harsh tag at the end. Yeah, of course, you know, the instructions were not very, very well here. Now that the prompt should be more precise with some more information. But this was just to show you how to add some hashtag while you are creating the content. And if you are, if you want to know some ash data about, let's say an image, e.g. a. Picture that you are posting on Instagram. You can say something like this. I am posting on Instagram picture of a vegan lasagna that I just made. Give me three harsh dark. Alright, can use the California plant-based Meatless Monday. So if you are posting this picture on Monday, this last one is going to be good. 23. 3.3 Emojis: In a similar way, I also asked her GPT-2 help me picking the right emojis. I don't know about you. I'm not great with emojis as well. We are stuck in emojis, are always need some inspiration. I don't want to use always the same emojis. That's why sometimes I asked IGBT to suggest me some ideas for the emojis and you can do that in a very easy way. So give me ideas for emojis to use in a motivational post on LinkedIn. So I told her GBT, that is a motivational poster that is on LinkedIn. And I wouldn't think emojis add-ins important. Here, there are two important information, motivational posts and LinkedIn. So as you can see here, there is a chart, an increasing charts. So we see progress and growth. Target the Moshe. This is about goals and also a rocket, so ambitions and reaching for the stars. So LinkedIn and motivational here where two important information. But also what I do sometimes is me. I don't know, five motivational tweets. We saw this example in the previous lesson. Now we're going to add our stocks and also emojis. So ReadMe file motivational tweets are some hashtags and emoji. So don't be afraid to take risks, blah, blah. There is the flexing the arm and also the star and then the rocket and the sun, the light bulb, the chart, the fire, the target, everything. And when I have something in my mind, I also check GPT-2, give me some alternatives. E.g. what are some good alter motifs to two D emoji? We the targets. So I don't want to use that emoji. Now, we actually got the same one, but the others are quite good actually. So the trophy is a good one. If you want to show, I don't know, some achievement. Wind or something. The triangular flag also for marketing or milestone in our journey. The pushpin, yeah, I love this one. So this is how I use IGBT also to improve the way I peak the emojis. 24. 3.4 Using Brainstorming Frameworks: A very powerful way to use charge EBIT is to leverage some famous and known frameworks. E.g. I. Do that when I have to do some brainstorming and I need some help with brainstorming. So let's say that I want to use the scamper framework. I don't know if you are familiar with the scamper framework is probably one of the most important and also most famous framework for creativity, lateral thinking for brainstorming ideas. So I usually, what I do, I first asked charge APT. If it knows the the framework I'm talking about just to be sure that we're on the same page. And it's going to use the framework that I want to use. So yeah, this is the one. So basically scamper. It's, yeah, it's a framework in our creative idea. There are seven steps, so each letter here is Substitute, Combine, Adapt, modify, put another, use, eliminate, rearrange or reverse. So now that I know for sure that GPT knows what discomfort framework is, I'm going to ask GPT to help me with a brainstorming. So helped me with the brainstorming using the scamper framework. So the one that I just told you, but let's use the seven steps one-by-one. I'm brainstorming ideas to sell more online courses. So now for each one of these, substitute, combine, adapt, modify, put to another, use, eliminate and rearrange to IGBT. It's gonna give me some ideas of how I can actually sell more online courses. So let's go with substitute. And we have three ideas combined, and we have three ideas, adapt and we have three ideas, and so on, so forth. So now I just want to read a few ideas with you to give you the value why this is important. So the first one, e.g. swap out the course instructor, instructor for a celebrity or an expert in the field. So using the word substitute or e.g. combined, combined multiple courses together, create unbundled deal for a discounted price. Or let's go here. Eliminate, eliminate any outdated or either relevant course content to streamline the course and make it more focused and so on, so forth. So if you use any framework, if you know any framework, there are tons of frameworks out there. For creativity, lateral thinking, brainstorming. You can actually use that as a template to tell Joe GPT, now, use this template, this approach and helped me with my task. And this is kind of powerful. 25. 3.5 Using Storytelling Frameworks: If you liked the idea inside the previous lesson, I'm going to show you another one that is related by even more powerful. So we started the previous lesson with the idea of using the scamper framework. But that's of course, because I know what is comp or framework is. I always use it. But what if you didn't know any frameworks? I mean, you know that I can follow and use frameworks as a template. But what if you don't know any frameworks and you want to use them. So what I do is I first asked you to give me some frameworks and then I use one of those. So let's go from, from brainstorming to something different. So storytelling, copywriting. So I'm going to ask to GPT-2, telling me five well-known frameworks for storytelling and copywriting. So if you didn't know any frameworks, IDA is a framework, a famous one. Pass. The Hebrew journeys, e.g. now tragedy is telling us five of those frameworks. This is great because we didn't knew those frameworks. Now, let's see that. Oh, we want to use the first one, so aida model. So now use the first one item and write a story about an Italian guy in the space. Let's try. So. Now the aida model. Attention either means attention, interest, desire, and action. So now the story is splitted into four pieces. So attention, interest, desire, and action, but we can use the same framework. I mean, this is a made-up story, but we can use the same framework also to generate some content for our newsletter, blog posts, for our adeno, LinkedIn or YouTube videos or whatever. So use the aida framework and write a blog post. Know, let's say YouTube video on how to use GPT to grow your business. So now tachypnea is going to use the same framework. So aida, again, attention, interest, desire, induction. To write the YouTube video about how to use char GPT-2, grow our business. And this is super powerful. You can use the Hebrew journey, I love it. The path formula, whatever. If you don't know a framework in your own area, just as TBT, to give you some frameworks, then you read one you like, and you ask distributed to use that framework to create the story or the content for you. 26. 3.6 Formatting the Output: An extra trick that you can use is austerity to format your output. Not a lot of people know that, but you can actually ask GPT-2, use it Alec, bolt, lists and so on, so forth. So this is quite useful if you have like big outputs and you want to improve the way the output is shown. Let's see an example. Now, I prepared a prompt. I'm going to copy and paste it here. Generate an idea for a YouTube channel about vegan food. The easy part, I need a title description are stocks and the idea for the cover image, and this is where it tells LGBT one I want. And now the formatting part format the title as a header format the hashtags as a bullet list and format the cover image in metallic. Let's see the output. So yeah, we got the title as a header. As we said, the arch dogs are bullet lists and the cover image description is in metallic. And of course you can use this trick to format the output the way you need it. 27. 3.7 Creating Tables: The last trick I want to show you is how you can use GPT-2 create tables. So this is another way to format our output. So let's say that we won't give me five tools for email marketing. Okay, so let's start with the easy one. We have five tools, Mailchimp costs and content. Hubspot, Campaign Monitor are sending blue. Now, generate a table. We, the name of the tool and the website. Now we have the table Mailchimp, mailchimp.com, constant conduct that because I cannot.com and so on, so forth. And if you need more information, you can add more information to the table. Now, odd table, the head quarters. Now we have the updated version of the table. One with one column with the tool, one column with the website, and one column, a column with the headquarters. Of course, this is just an example as usual. So instead of tools, email marketing tools, you can put in here whatever you want. But it's really powerful to have tables, especially if you have very big and long output. This is a better way to format it into, use it to show it to your team, to copy and paste it and document or in a presentation. 28. 4.1 Real Life Examples: Welcome to the last module of this course. This is where I will showcase some real life examples of how you can use GPT in your day-to-day work. Of course, we have seen a lot of real life examples in the previous lessons, but I want to recap some of them and also want to show you a few more. All the following examples are based on my personal experience. We charge a pity and I hope they will help you to improve your work, your creativity, and also your productivity. I'll see you in the next video. 29. 4.2 Brainstorming Ideas: So the first real life example that I want to show you is how to use GHB T to brainstorm ideas. These probably the most common use of IGBT. And we saw a little bit of that in several lessons before this one. So I usually start with a simple prompt like this one. Let me copy and paste it here. Generate a list of five ideas for blog posts, but vegan recipes. So if you create content, if you have a YouTube channel, I don't know a podcast, a blog, even I don't know our LinkedIn page and Instagram page. Whatever. You can use GPT to brainstorm idea by yourself or with your theme. It's a great help even if you have a team. So we start with a simple prompt and we get the five ideas for vegan recipes. As we saw in this course, we can actually improve the prompt. We can tweak it a little to get better results. So the first iteration that I will suggest is this one. So let me copy and paste it here. So I always try to be more specific with my inputs. So I tried to give him, to give to charity buddy some keywords. So e.g. generate a list of five ideas for blog posts about vegan recipes, including this one recipe about pasta and one about ice cream. Maybe for some reason I want to talk about pasta and ice cream. So I'm gonna be more specific on that. So let's see the result with this prompt. So the first one is a creamy vegan pasta recipes that are easy and delicious. The second idea is five vegan breakfast recipes to start your day, right? The third one is how to make vegan ice cream at home. So as you can see, now, we got ice cream was one of our keywords and also bus stop, that was the other keyword that we use. So let's improve it a little bit more. And I would do something like that. So generate a list of five ideas for blog posts about vegan recipes for sports people. So now I'm giving the target as well as we saw in one of the previous lesson. This is quite important. Include at least one recipe about Boston in one about ice cream. So now, with the information of sports people as a target, let's see the result. As you can see, the result is already different and I love this one. So high-protein vegan meals for athletes, vegan pasta recipes for endurance athletes. How to make vegan ice cream for athletes. Vegan pre-work house next to fuel your workouts and Recovery meals for vegan athletes? I loved those when a usually brainstorm with church PD. What I do, I mean, in when I'm here in these steps with some ideas, I always asked for some titles. So e.g. now we can say, okay, now give me five titles for the five ideas. So we also generate the titles. So, yeah, here we are. We have the five titles. Now let's suppose that the first one is actually going to be a YouTube video. So something that can be very, very useful. It's something like this. The first one is going to be a YouTube video. Please generate the structure of the video as a bullet list. And the structure is always very useful even if you are writing a blog post or recording a podcast or any kind of content you are doing. This of course, doesn't mean that charge if it is going to tell you what to do and what to say inside your content. This just the structure. This is just an idea. Now, you have to put the content, your experience inside the, inside the contents. So now we have the structures or at least we have an idea. If we don't like, I don't know the first one, we can ask distributed to change the first one so I don't like the first recipe. Please change it with something else. So instead of the middle number one, Lenton and quinoa salad, now we get vegan, chick pea and sweet potato curry. You can always do that iterate with the, with the result. And now that we have our final idea, we can ask to GPD something like this. Now, give me an idea for the Tammany in five dash tags for the video. So now you have all the information you need to do to create your content. 30. 4.3 Content Repurposing: In the previous lesson, we saw that we can use charge be T to brainstorm ideas, to generate ideas for our content or for whatever we need. But we can actually also use charging P T to repurpose content that we already have. And this is great. So if you have blog posts, if you have videos, if you have tweets, if you have a newsletter or whatever, you can use that content and create new contents starting from that. So a good example is to start from a long-form content, a blog post, e.g. a. January, some tweets. So let's, I'm going to use once again HubSpot just as an example, of course, you can use your own content here. So let's say that I want to use this piece of content here. So let's copy this part here. Let's copy. And we can ask GPT-2 generate three tweets from this content. Now, we give Josh have a TDS prompt here, and we can have the tweets. So we start from a long form content and we get some tweets, but you can also do the opposite. So e.g. let's say that you have this tweet and this is yours, maybe. Like an old tweets and you can create a blog post from there. So let's say that you have a tweet about quality over quantity is key when it comes to content marketing creates a blog post starting from the tweet. Number three. Let's see the result. Now, Chess Jeopardy is going to do the opposite. So the opposite, so now we start from a small piece of content and tweet in our example. And we can generate a long-form content as a blog post. But it could be also, I don't know, a newsletter or it can be a LinkedIn post or whatever you need. Of course, we can iterate this. We can do whatever we want. So let's say that e.g. you need a video from that we saw something similar in the previous lesson. So let's say that you want to start from the, from the second one, show short-form videos are the hottest marketing trend of 2023. Create a video script from the second tweet, as we saw before. Now, charging party is going to create a script. This is not the structure, this discrete, so this is actually what you can say inside your video. It doesn't mean that you have to follow the script. But of course this is, as I said before, this is just a starting point. So you can have a structure, you can have ideas, you have some inputs, and then you start from that and you add your own experience. What you've learned, or your successes and failures, and everything that you think is important inside that content. So this is great for videos, but let's say that we went, we won't create something different, e.g. a. Carousel for Instagram. So we can say turgidity. Now, generate our Instagram Carousel. Tell me what I have to write in each slide. So of course, an Instagram carousel is a series of slides. So now, for each slide is telling us what we can write in this slide. And it's going telling us also some useful ideas for the image or the video that we can use inside that slide. So, of course, cannot generate images or infographics or videos, but you can use it to generate the idea behind that video or that infographic or that image. And once you have that, you can create the actual content. And this is quite good. 31. 4.4 Translations: Gbd is also quite good for translations. So e.g. in my case, I'm Italian, English is not my first language. I live in the UK, spend most of my day speaking both Italian and English. Jupiter can be very useful in my case, but it can also be useful if you speak English. But you have clients all over the world and you have to write emails or content or whatever in other languages that is not yours. So we can start with something like this. E.g. I. Can ask GBT to translate this sentence from Italian to English. So translate this sentence in English Channel Marco. Man-eating bulimia, chronic preventable Jacobian will discuss. So this is gonna be high market. Tomorrow. I'll send you the e-mail with a quote we talked about. And you could say, well, I do that with Google Translate. And it's easier. Yes, that's true. But something that Google Translate cannot do is this. So you can ask your GP t to have more alternatives. So e.g. I. Don't want to use this version. I want to know different ways, several ways of how I can say the same sentence so I can answer should be D. Now give me five different alternatives. So hey, Marco, I'll be sending you an email with an estimate we discuss tomorrow or I don't know. Hi, Marco. I'll be sending you the email tomorrow with a quote we discussed. And if you remember, we also spoke about the tone of voice. So I can use these in translations as well. So I cannot change it be t. Now, give me two ideas, very, very formal tone and two ideas with a more funny tone. So let's see what happens here. So now it's telling us, okay, now you've got the two ideas with a very formal tone. Good day, Marco. Please be advised that that will be dispatching the email containing the cost estimate we discuss tomorrow. I will never do that. More for neutron. Yeah. Oh Marco, the estimate evil is coming at you like a boomerang tomorrow, so watch out for it. It was just an example, but you can use, you can always use that when you do translations. And as I showed you in one of the previous lesson, if you don't know what kind of tone you want to use, you can ask GPT-2, give you some ideas for some tones. So e.g. give me ten ideas for a tone of voice is I can use in my emails. So now you have ten different ideas. So now e.g. you know that you want your translation to be more adeno, humorous or casual or inspiration out. So now we can copy and paste this one. Let's copy the first one. Translate descendants in English, but using inspirational tone of voice, using an inspirational tone of voice. So let's see the output. Hello market. Tomorrow, I will send you an e-mail that will set you on the path towards achieving your goals with the code we discussed. I love it. I would never do that, but I loved the the idea behind it. So yeah, Jeopardy is dirt. If you had to do a lot of translations, if you have to translate some content, if you were to send e-mails, if you have clients internationally all over the world, or if you were to write something that is not your own language. 32. 4.5 Grammar and Errors: Since we are speaking about translations, something that I love to do with charge a bit t is also check the grammar. So I always asked attributed to help me with an English sentence because English is not my first language. So if you're writing content in a language that is not yours, it's always a good idea to use GPT to check the grandma or find errors, typos, and also to improve the sentence, I'm going to show you some examples here. So I usually start with something simple like this. So let's copy and paste. This sentence is okay is the one we got before. So check the grandma. Hi Marco. Tomorrow I'll send you the email with the quote we talked about. Now is telling us the grammar in the given sentence is correct. So now let's try with a sentence that is actually there is something wrong in it, so check the grandma. Hi Marco. Tomorrow is not correct. I'll set you instead of sending e-mails, instead of e-mail with the quote we told about. So let's see what Judge Betty fixed. So the correct sentence should be high. Marco, tomorrow I'll send you the email we decode we talked about. Here are the corrections. So what activity did was, you know, tomorrow should be tomorrow set your emails should be send you the e-mail and talk about Should we talked about and this is great. I always do this. Also. I charge it with T Y. It decided to change something and fix it. So what was the reason? I like to know also the reason, once again, because English is not my first language. So maybe it will be great for me to understand the difference between set the e-mails and send the emails or talk about and talked about. So now we got the whole the whole thing. So firstly, tomorrow was corrected to tomorrow as the lateral is the correct spelling of the word set you emails was corrected to send you the image to make the sentence grammatically correct, set is not the correct verb in this context and also talk about was corrected in talked about. This is because the current form of the webinar content is the past tense, talked as the discussion has already taken place in the past. I love it. I mean, of course if you are an English speaker, you don't need this. But if you are writing content in a different language, this could be baby useful. And also what I do sometimes, instead of the problem, check the grandma, fix this sentence or whatever I asked to improve a sentence so I can write something they do improve this sentence. So let's say that I'm writing too. I don't know. A client. You're Marco. I'll send you the email in a minute. So maybe this is correct in a very direct and more funny way. If I'm talking with a friend or with someone that I know. But it could be not a good idea if you're talking with your boss or with a client or whatever. So I always ask turgidity to improve the sentence a little to see if there is a better version of it. So let's see what we got here. So they improve, to improve the given sentence, you could make it more professional and use proper grammar. Here is one possible revision. Hello Marco. I will send you the e-mail shortly. The sentence is clear, more sea ice and uses a polite tone appropriate for the professional communication. So once again, if that's not your first language, this can be very useful in a day-to-day situation where you are sending e-mails or writing content. And also imagine that when you have like a huge piece of content, like a blog post, you want to find errors and type in something that is quite big. This can be very, very useful. 33. 4.6 Finding information: Charge EBT is also quite good for finding information. We saw that in some previous lesson, so let's see how we can use that in our advantage. So e.g. let's try this prompt here. So find me five tools for automation marketing. So we saw something like that in one of the previous videos. So now GBD is finding us information about five tools for automation marketing. This is quite useful, e.g. if you are searching for competitors. So we can say, I'm building a landing page tool. Tell me ten competitors. So now we have a list of ten competitors that do something similar to what we are doing. So Unbounce, lead pages, click funnels, Insta, page, weeks, WordPress, GetResponse, blah-blah-blah. All of them. What I like to do. And I showed you that when we spoke about tables. I can ask, Judge a pity if you are doing our competitive analysis, competitive research, put them in a table and add the website and the LinkedIn page. Maybe because I want to, I want to check them on LinkedIn. I want to add them, follow them on LinkedIn to see the content strategy. I want to double-check the website to see how they are creating content or the UX UI of the tool, whatever. This is, quite good, but we can also use this in a different way. Let's see e.g. that you have a podcast and you are looking for some ideas of guests that you can have in your podcast. You could do something like this. So as a podcaster, so the first thing I learned, if you remember, we spoke about the powerful act as prompts. So act as a podcaster. You have a very famous podcast about artificial intelligence. Where you discuss, Sorry, where you discuss with your guests the importance of AI in our society. Now, you need to invite five new guess for the next episodes where you want to talk about AI in marketing, generate a list of five guests and tell me why you would invite them. So this is great if you want to find guess. So GPT now is going to tell us who those persons, who those people are, and also why. We should invite them in our podcasts, e.g. Shelley grammar. Shelley is a futurist speaker and an analyst. Christopher Penn is an expert in marketing data science and AI. Paul Rutter pore is the founder of the marketing artificial intelligence Institute. Katie King is a keynote speaker, author, and consultant. Jason. Jason is the VP of data science at zeta global. I love this if you need some ideas. And we can also ask for more. So we can say January, the structure of the first episode with a list of ten questions for the guest and tell me what's the reason behind each question. So GPT going to write for us the first one was Shelly grammar, now is writing the structure of the podcasts. So this is the structure more or less. Of course you don't have to follow the whole structure, but you can have ideas, e.g. you can generate questions. So why does this question is important? So the first one, let's see an example. How would you define AI and its role in marketing? This question sets the foundation for the conversation by establishing a common understanding key concepts. I like this. I mean, it's quite useful if you are, if you have a podcast or a YouTube channel, I don't know. Guess boss with someone from some guests. So this is quite useful as well. 34. 4.7 Inspiration for Images: The last example I want to show you is how to use charging for finding inspiration or inspirations for your images. So e.g. if you need to create a thumbnail for YouTube, a cover image for your blog post, or an Instagram carousel or whatever. So as I said before, touching beauty cannot generate images, but it can give you some ideas and you can go and find the right image. So e.g. let's start with a simple prompt like this. I'm writing a blog post about the best cities to visit in Europe. Generate five ideas for a cover image. I can use my blog. I always asked for more than one ideal. So I can actually think a little bit about it. Pick the right one, or sometimes also mix more than one idea. So e.g. a panoramic view of the FL towel in Paris, France, a street view of the colorful buildings in gambler Stan, Stockholm, Sweden. A view of the canals and gondolas in Venice, Italy as sunset view of the Santorini, Ireland, Greece. And our shot of the Gothic architecture of the Old Town Square in Prague, Czech Republic. But I can also do something like this. So if you have a very specific needs, I need to post an inspirational quote on LinkedIn by Albert Einstein about creativity. So LinkedIn, Albert Einstein, creativity, those are three important keywords in this prompt. And we saw that keywords are very crucial for creativity. What image can I use? Give me five examples, five ideas. So let's see. The first one is a photo of Albert Einstein with a simple background. Okay? Yeah, that's quite classic. Black and white. The second one, an image of a painter or a sculpture, sculpture working on a piece of art. A third one, a photo of a beautiful and inspiring landscape or cityscape. The fourth is a photo of a group of people brainstorming. And the last one is an image of a light bulb. Light bulb or a spark of inspiration, which represent the aha moments that Einstein quote speaks to. I love the last one. We can also use it for Instagram Carousels. This is quite good. Actually. Generate an Islam Instagram carousel with five Italian recipes. For each slide, I need the name of the recipe and the idea for the image. So this is very good because we are asking to give, gave us, to give us the, also the image that we should use in debt slide of the carousel. So the first one, as you can see, the structure here is slide number, name or a recipe and image idea. So pizza margherita, a close up of a freshly big margarita pizza. Perfect. Second one. Spaghetti carbonara. A photo of a bowl of spaghetti carbonara were crispy back on black pepper and grid Parmesan cheese. Please don't do that. Bagel environments are not the original recipe, but anyway, bruschetta, blah, blah, tiramisu, result of communism. So this is perfect. So let's use your imagination here. If you have a need of using a picture, an image somewhere in your content, in, I don't know, in your, in your ads, even you can use GPT. You can ask to Djibouti to help you find the right image. 35. 4.8 Conclusions and Final Tips: Well done. You've reached the end of this course. I hope you have enjoyed the course and learned some valuable techniques for optimizing your use of this powerful tool. In the first module, I spoke about some common mistakes when using GPT and how to avoid them. E.g. we saw why I iterating is important, how to provide examples and how to include requirements in your prompt. And those are going to make a huge difference. In the second module, I showed you some advanced use cases of IGBT. We saw the powerful act as prompt, but also how to analyse and recreate tone of voice and how to check the prompt sides and split it. In the third module, I shared with you some tips and tricks for optimizing charge apathy. Based on my experience, we saw the power of using frameworks. And this is probably one of the less no techniques about GPT and it's very powerful. And finally, in the fourth module, I showcases some real-life examples on how you can use in your day-to-day work. Content repurposing is my favorite one, but it will also great results for translation brainstorming, and so on, so forth. And before saying by, let me stress the fact that using GPT as a sole source of information or decision-making can lead to errors. So please use charged with D as a complimentary tools rather than a replacement for human intelligence, use it in a safe and ethical way. Be mindful of the potential biases and limitations of the model. And always, always consider the impact of degenerated texts on your audience and be sure that it aligns with your ethical principles and values. Please remember that you can download some extra resources related to discourse. Just visit guides dot link slash GPT m. The link is in the description as well, and you will have access to a checklist on how to improve your prompts and a list with hundreds of prompts that will blow your mind. Well, thank you for joining me in this journey. I hope you have enjoyed this course and most importantly, you have learned something new. I'm very confident than the tactics, tips and tricks that you have learned in this course will help you discovered the full potential of therapeutic. Remember to keep experimenting and learning and always be open to new ideas. And if you have enjoyed this course, please, please, please leave a review. It's very important for me and it only takes a couple of minutes for you. I also invite you to submit your class project, and it's been a pleasure. Ciao.