How to Get Rich With GPTs in 2024 | Complete Beginner's Guide (OpenAI Custom GPTs)

open AI has just released the GPT store marking the App Store moment for AI looking back to 2008 when Apple Unleashed their App Store only those with app development skills got a slice of the millions the App Store generated the difference for us today is that the skills required to create for the GPT store and start making money can be learned in an hour or two rather than 3 years of software engineering at College that's why today I'm going to be teaching you absolutely everything that I know about building gpts and making money from them as a beginner so that you can take advantage of the Easy Money new platforms like this offer before it's too late however since there's going to be tens of thousands of people rushing to this new opportunity the challenge is not knowing how to create gpts but how to create actually valuable gpts that will stand out from the competition and generate you passive income for years to come teaching you the skills required to do so is the goal of this video now before you sit down and listen to another random guy on the internet teach you how to make money online if you want to be a millionaire next month this is the way to do it you need to know a little bit about the person giving you the advice if you're new to the channel and don't know who I am my name is Liam otley and I run my own AI agency morning side Ai and I also run the largest community of AI agency owners in the world where my students learn how to make money by selling AI solutions to businesses just like the gpts you're about to build I'm originally from New Zealand but now live in Dubai and I've been doing online business for the past 5 years over the past 12 months I've sold nearly half a million dollars worth of AI chat Bots to my clients and I've taught the people in my free and paid Community how to sell tens of thousands of dollars worth of their own the reason I'm making this video today is that within my own agency we are starting to see a massive shift in demand away from the normal AI chat Bots we're used to building towards AI agents like gpts that are equipped with tools and knowledge the change in demand has been so significant that we have pivoted almost all of our resources to going all in on this opportunity at I specialized in AI agent creation the majority of which is done through my AI agent platform agentive now I hate to ramble on about myself but when I was starting out an online business all those years ago I wasted my time listening to people who weren't qualified to give the advice they gave so I've been in your position before and I don't want you to go through what I did so that's a little bit about me I hope you think I'm qualified to be giving the information I'm about to share as you can probably tell by the length of this video I'm not holding any secrets back from you I've structured this video into three different chapters starting right from the basics of what is a GPT all the way up to the advanced deployment techniques that I've found to make money with GPT but feel free to skip around using the time stamps along the bottom so you can start wherever your skill level is chapter one is foundation in order to make money with gpts you need to properly understand gpts and the opportunities they present this chapter covers the basics for beginners right up to the dangers and pitfalls of becoming H gbt creator that no one's talking about right now chapter two is skills in this chapter I break down each of the five different skills you need to know to be able to crush as a no code GPT developer and to be able to make passive income from from the GPT store and some of the skills you need you definitely wouldn't expect and finally chapter three is creation in this Final Chapter I walk you through creating your first GPT that incorporates all of the available features on the GPT platform and I'll also show you how to jailbreak that same GPT and use it elsewhere on places like websites WhatsApp and More in order to get the most out of this video I want you to close out of all of your tabs save this video to your watch later and give yourself the next 2 to 3 hours to study the information in this video I want you to go and get a note book and a pen because there is far too much information in this video to soak it all in on your first run through and if you've done all that we are ready to get started Chapter One Foundation what are gpts for those complete beginners here we're going to be starting from the basics first things first we need to identify what is a GPT well gpts are best described as custom versions of chat GPT essentially they are customized versions of chat GPT that have been modified to be better at a specific task or more helpful to a certain type of person for example some of the most popular gpts to date have been canvas GPT which allows you to create designs and Graphics from natural language secondly at grimoire which is the number one Community made GPT at the moment which makes you a 100x engineer allowing you to build websites and more with prompt gramming this particular GPT is so high ranked because it has been cleverly marketed by the owner which we'll be going into dep on later then we have things like designer GPT which is a community made GPT which generates and hosts websites for you in just one prompt each of these and the other Popular Community made gpts which I'll leave a link to the leaderboard down below if you want to check it out but each of these have been created by modifying the functionality of the base chat GPT to be better at a specific thing by using what I like to call the three ingredients of a GPT these three ingredients are prompting so you can give your GPT a detailed set of instructions to determine its behavior for example the response length or the response style or the gpt's goal or purpose can all be communicated via prompting the second ingredient is knowledge the GPT Builder allows you to add your own files like PDFs spreadsheets or other documents to allow your GPT to draw information from these documents when responding to user queries so essentially a chat GPT on your data which was a hugely popular concept last year and thirdly actions so actions are special functions that your GPT can perform outside of just writing a text response to the user you can use either open ai's existing capabilities like image generation with doy browsing with Bing or the code interpreter or you can create your own custom capabilities that do things like send emails or book restaurants and more actions can also be called tools or functions as you'll likely see later in this video different combinations of these three ingredients of prompting data or knowledge and actions can create any kind of modified chat gbt that you can imagine but it is important to note that only one of these ingredients is actually required when creating a GPT and that is prompting with knowledge and actions being optional however knowing how to effectively find and add the right kinds of knowledge and create custom actions for your gpts is how real value is created later in this video I will be walking you through how to learn these valuable skills step by step so that you can immediately create high value gpts and start making money with your skills in 2024 it's important to note that GPT is just a fancy name for an agent an artificial intelligence an agent is a computer program or a system that is designed to perceive its environment and make decisions and take actions to achieve a specific goal or set of goals basically agents are chat bots on steroids the chat Bots we're used to typically just reply with a often pre-written response or some basic AI generated output agents can do far more than just send a text reply they can interact with apis and and anything across the web which allows you to do some pretty cool stuff with them you could theoretically have an agent that would fire a nuke for you as long as you told at the location and provided the launch code that's kind of the the triggering of of actions that you can do with agents if this sounds confusing at all don't worry because later in this course I'll be walking you through creating your own GPT or agent from complete scratch and everything will start to click into place gpts are really open ai's way of introducing the world to AI agents but this is just the first step in a much larger plan to introduce autonomous agents to our world more on this later the App Store opportunity at the time of filming this open AI has just released the GPT store people are calling this the App Store moment for AI which if this is true we have possibly the biggest opportunity to create wealth of the next decade in front of us as regular people this is the equivalent of being one of the first people to create an iPhone app in 2008 and List It On the App Store now I know that is a big statement to make so please stick with me while I back it up I know how pain V is hearing every second YouTuber talking about the big next opportunity but give me a second and I will back it up if you're a smooth brand gen ZZ kid like me we missed the App Store opportunity in 2008 because we were too young and even if you were old enough in 2008 creating apps was only really possible for the privileged few who had learned how to code and were really in the right place at the the right time to jump on the mobile apps opportunity the point here is that historically when new platforms arise it is stupidly easy to make money because there was so much attention so much demand and yet limited Supply the entire market and ecosystem around the platform has not really found equilibrium or found balance point yet so there are countless opportunities to catch a little slice of the wealth flowing into the space from investors and sponsorships and new startups with big budgets that they need to spend somewhere we've seen this with new platforms time and time again firstly of course we have the world Wide Web where early players caught the doom boom and were able to get millions and millions of dollars in funding with with just a name that says do com in it and ended up exiting for hundreds of millions or billions of dollars Mark Cuban sold his company broadcast.com which is radio put on the internet for a billion dollars and it's it's nothing now secondly of course we have the Apple App Store for the people who are early to the App Store and were early Builders the money came fast in the first 12 months following the app store's launch a guy named Steve de created a puzzle game called trism which turned over $2 million another guy called Bill rapos launched an app development company called wait for it great apps and was making $100,000 per month within a few months of launching an app called ieam that took 7 days to build a free application in the App Store two out of five stars hell some guy even made an app called I am rich that cost $1,000 and did nothing and sold a copy in 24 hours thirdly Facebook and Google ad platforms anyone who was early to these advertising platforms had ridiculously low advertising costs and basically printed money with super basic ad creatives and landing pages now years later the rest of the world has caught on and now companies struggle to profitably spend money on paid traffic due to the rising cost in competition finally number four is Tik Tok the growth that the early creators on Tik Tok saw was insane and many of these guys and girls have gone on to become some of the biggest stars in the world think of people like Charlie dellio and Bryce Hall and Addison Ray all these kind of creators now that kind of short form content is fiercely competitive and a lot harder to get that kind of growth I'm sure you get the point but what I'm trying to say is that early typically equals easy or at least early equals greatly increased chance of getting lucky and seeing disproportionate success if you tried to go onto Facebook ads and Get Ridiculous returns now the chance of you getting disproportionate success is much lower than it was when it first started that's what I'm trying to say here so it's clear that if the GPT store does take off and become the App Store of AI there is a huge opportunity for us but how can we actually make money from the GPT store open AI has announced that there's going to be Revenue sharing for GPT creators however Sam Alman has said that we're going to pay people who build the most valuable and most used gpts a portion of our revenue does that mean that there's a minimum usage requirement for you start to get paid how much will the revenue share be Etc there's a lot of questions around that I actually don't want to say too much on the revenue side of things just yet because at the time of filming this they haven't released a store yet but by the time you watch this they will have released a store yet hopefully more information will have been revealed by then so I'm going to make a video following up on this talking about Revenue sharing Etc so make sure you subscribe to the channel down below so you don't miss that but honestly I'm not expecting too much here in terms of Revenue sharing and I think it may have been more of a marketing tactic to get everyone hyped about building on the platform but I would love to be proved raw so if the payments from open AI are lackl what options do we have to make solid income from this gpts and agents book well I'm glad you asked the truth is the GPT store is only one part of a much larger picture sure creating valuable gpts and getting tons of usage that will earn you a few thousand a month is great but honestly I don't know if that will be possible for the majority of us the reality is that those with the audiences and the means to effectively Market their Creation in the gpts will have a massive Edge over your average GPT creator for example if I made a GPT for AI agency owners I could pump it up and its ranking on the store by making videos about it not all of you watching this can do the same all is not lost though as I said gpts are just a small part of a much larger picture AI agent building is the most valuable skill of 2024 so we've established that gpts are just a fancy word for AI agents right or AI assistants or whatever you want to call them basically AI chat Bots with a unique mix of prompting knowledge and actions the three ingredients that means that if you learn how to create GPT you're technically learning how to create AI agents no correct well the good news is that knowing how to create AI agents is going to be one of the most valuable skills and future proof skills that you can have in 2024 and Beyond not just for your personal use and productivity but for making money online as I'm about to explain the AI agent landscape is expanding rapidly and is set to be one of the most if not the most world changing fields of technology in the next decade a world where AI agents can speak to customers handle complex tasks and and join companies as part of their AI Workforce is really where all of this is heading but don't take it from me take it from all of the biggest Ai and tech companies in the world which have made it clear that AI assistance with custom actions AKA tooling is where they are placing their bets one of open ai's biggest competitors anthropic has recently announced that their model Claude will have API tool usage AKA actions and its 2.1 version Google has just released Gemini the multimodal GPT 4 killer which will undoubtedly support custom actions and agent Creation in the near future it looks like a bird to me even Amazon is expected to add agent creation and hosting services to its existing Cloud offering as well across all these different companies rushing to the same opportunity the same skill set will allow you to build agents I'm going to be explaining each of the five key skills needed to build agents and the steps you can take to master them in the next chapter of this course before taking you through creating your first GPT an AI assistant step by step at the end of this video what I want you to walk away from this video with is not only the perspective to see this opportunity for what it really is but also the knowledge you need to take your first steps into AI agent creation even as someone who doesn't know how to code but we're still stuck at the same question if we learn AI agent or GPT creation how can we make money from the skill if you devote the next 2 to 3 months to become a killer agent or GPT developer how can you monetize these abilities well who needs agents basically every business on the planet sooner or later this happens every time with new technology like the web and mobile apps when it first takes off there's a massive need for people who know how to build the technology for businesses so that they can adopt the technology and benefit from the Gams for example web development agencies popped up to make money by facilitating the adoption of websites by businesses which website seems quite crazy and stupid at first now everyone has one then mobile app development agencies popped up to make money by facilitating the adoption of mobile apps by businesses and now with this magical new AI technology is a massive need for agencies or Freelancers who know how to identify valuable use cases for agents and Implement them into businesses this is the opportunity that I realized last year and created the AI automation agency model which has since exploded into an online business opportunity and I've helped thousands to build their own AI agency like mine and both my free and paid communities but I didn't take this long just to give you a sales pitch I only bring this up because the successive both myself and of the hundreds of successful AI agency owners in my community it really proves that there is something here and a way to make money by selling these things to businesses if I do my job right with this video you should be able to get started and make money without ever paying me a dime at least that's the goal of the video this opportunity to sell AI agents to businesses or anyone who needs them really is the big picture of the GPT store release for example local businesses can use the AI agents that you can create in a few hours to automate their customer support or lead qualification digital businesses like agencies can use your AI agents to automate their researching and contacting of prospects solopreneurs can use your AI agents to supercharge the abilities with content automation or mass Outreach tools the list goes on for these kinds of use cases you need to break your GPT out of the chat gbt site and put them on websites and install them in existing systems this is easier than you think and I'll show you how to do this in the final chapter of this video my point here is that the AI agent Market is going to be massive and full of opportunity and personally I think it is one of the biggest opportuni of 2024 and Beyond mainly because it is so accessible to people to learn the skills needed to build this crazy powerful software people could only dream of a few years ago a lot of side hustles and make money online grifts are are borderline scamming and what I love about this is that you're actually building a hugely valuable skill set and implementing technology that will truly make a difference for your clients building AI agents presents you with an opportunity to charge deep into one of the most exciting fields in the world over the coming years and not only make money from anywhere but also open up doors to much bigger opportunities along the way market dynamics while I did say that the GPT store is only a small part of the picture I still want to give you the information you need to give in a good crack as I said if I teach you how to create killer gpts that same skill set can be transferred over to selling agents to businesses so it's a win-win with the GPT store launching I want to touch on the market dynamics that we'll likely see play out and being aware of this will ensure your strategy is better than most giving you a better chance of success whenever new platforms like the GPT store start there's usually a few different phases initially the market will be flooded with basic gpts that take off and a good for a specific use case but don't really have much to them maybe a couple PDFs and some decent prompting and these documents are likely publicly available and anyone can find them and recreate that same GPT for themselves but as these low barrier to entry GPT use cases flood the market they'll quickly become commoditized and lose all value if you make a GPT that 10,000 others can create in 20 minutes by by downloading the same ebook as you you're never going to get the usage numbers required to make any decent money on the platform because there might be a 100 different GPD creators who will given me the exact same thing and therefore all of the usage for that kind of use case is going to be split between all of the people but rather than you getting all of the usage it's going to be split between all of the other copies of the same GPT and you're not going to be making any decent money then the next phase will begin where Builders will need to get more creative and find other ways to create both unique and valuable gpts that will attract and retain users over a long period and creating this kind of value and long-term usage comes back to the three ingredients concept we discussed earlier we have prompting we have knowledge and we have actions to have any kind of long-term or consistent income from your gpts you'll need to be able to create unique value through one or more of these there are a number of different ways that you can create unique gpts so you might want to get your book out and and write these ones down firstly you can use expert or unique prompting which is fairly easy to replicate if someone else can come up with the same or similar prompt they can create the same kind of functionality you can but that is one way you can create some kind of point of difference with your gpts secondly you can use private or hardto find data as custom knowledge as you'll learn later in this video you can either kind of buy or scrape data off the internet and make your own custom knowledge base that very few can replicate because if you can get knowledge or information that people can't then people aren't going to be able to replicate what you've created with your GPT thirdly you can curate an interesting or unique mix of data in your GPT this is a pretty interesting one to me but the idea of you being a bit of an artist or a chef and and and putting the right types of data together whether it's publicly available or private or hard to find just baking an interesting mix that is hard to replicate and that provides a unique type of value that others can't provide and finally number four is of course creating unique or powerful actions and tools and functions that your GPT has that other people can't easily or or can never replicate for example if your uncle runs a roofing company and has dozens of high performing sales scripts that have been crafted over two decades to be really high performing you can take those and put those into a GPT as a custom knowledge base and call it a Roofing sales script expert by doing that you'd have created something that is difficult to copy and better yet you can throw in a roofing cost calculator tool that you created on relevance AI to increase its value even further in the next chapter of this video I'll be going into more depth on tooling relevance AI so stay tuned for that my point here is that it is much better to spend your time creating gpts that have some kind of unique value to them that is hard to replicate otherwise someone else is going to make the same thing and you'll have to split your Revenue if any with them be smart about where you spend your time and don't get caught up on dumb projects that will never make you a dime to make gpts that are actually unique and useful you need to master the three ingredients and become a GPT shf however there are two other skills you need to hit it big on the GPT store chapter two skills in this chapter of this free course we are going to be covering the five key skills that you need to succeed as a GPT Creator on the store and as an AI agent Creator working as a freelancer or an agency selling to businesses skill number one market research and use case identification it's pretty hard to make a GPT that thousands of people use every day and racks up some juicy Revenue sharing income for if you don't know what people want therefore market research and identifying a high value use case for your gpts is the first skill that you need to develop when doing research and coming up with your ideas that always needs to be through the lens of the three ingredients you need to be thinking how can I combine prompting knowledge and custom actions together in a way that creates something both useful to a specific user and unique and hard to replicate for example as I've done earlier on this channel you might make a gold investing assistant for the prompting ingredient you would tell the GPT that it's supposed to help with investing in gold using some of the knowledge based documents provided and the actions that you've given to it the knowledge ingredient of this GPT could be some PDFs on gold ADV Bing that you downloaded from the internet and jammed into the knowledge base and finally the action ingredient could be a gold price checker tool that you use an API to cetch the current price of gold and you make that available the reason I'm using this example is cuz I've made it in an early a video on my channel I'll link that L that up here if you want to watch that that's a very basic example and probably not the best one if you're looking to make money another example that my Twitter first heard about is a quotation assistant for framing businesses so I had a conversation with my auntie who owns a framing business in Oakland and she said that like the quotation process for us is an absolute nightmare um we have to go back and forth with them asking for like the size and the shape and the style that they want it's much easier if you able to create a GPT or an assistant that can collect the information from the user send it off to a calculator tool give them the price back send a payment link Etc so for the prompting ingredient of this GPT you'd say you are a framing business assistant who helps customers to get a quote for the knowledge component we don't need to use anything here and for the actions we' connect two tools firstly a generate quote which sends off to a quote calculator and Returns the price and then another tool that allows us to generate a unique payment link based of that quote price tender to the user and capture the payment so if you're looking to come up with a few ideas there's a few different ways you can do it firstly using your own experience in my opinion and my experience this is always the best option because if you know something about a specific industry say you've worked in as a content creator before or you've worked as a editor or something you know more about that issue and the kind of problems that they face and the kind of solutions they would need than anyone does so if possible think back through your past experience or what you even do right now think what kind of gpts could I create to solve the problems that I'm familiar with secondly you can hijack the experience of others so as I just mentioned with my auntie when I was having a conversation with her she told me hey this particular part about our business sucks and I was like okay that's an opportunity for creating some kind of AI system so if you can you have friends family anyone you can find who has specific information or understanding of some kind of area or Niche or industry they might be able to give you some insights on what kind of problems need to be solved and what can be solved with the GPT solution and finally you can do your own research on the internet you can use sites like Reddit and YouTube to research about a specific Niche or about a a specific type of user and try to figure out what their problems are and how you could solve those with a GPT solution one great platform for this that I have no affiliation with is called gummy research it has AI features that basically allow you to pick and mix different subreddits and then get AI insights about the kind of problems and and things people are complaining about so I'll leave a link to Gummy Search down below again no affiliation to it I just think it's a great tool for any of you who are doing market research and I recommend this to the guys in my accelerator and free community as well once you've got a bunch of ideas for gpts that you can build we then get to the next phase skill number two data sourcing preparation and curation in many cases but not all as you saw with the uh framing system that I mentioned before that doesn't need knowledge most of your gbds and agents will likely need custom knowledge element to make them more useful and relevant for your users this means that knowing how to find valuable data prepare it for use in your agents for best performance and potentially mix different data together and create a unique experience there's a set of skills that you need to have I find it's helpful to split data into three different categories public hidden and private so you might want to write these down public data is anything that you can find for free on the internet for example searching up gold investing PDF and adding any results find to my agent or to my GPT is an example of using public data of course since anyone with an internet connection can get to this data and replicate your gpts you likely face a lot of competitors using the exact same data who will end up siphoning off your earnings on the store hidden data is anything that requires a little bit of extra work to get your hands on examples of hidden data could include expensive or sort After books you have to buy or or find data scraped from the web and combined in a unique way or data collected yourself from things like surveys while all of this data is still accessible the hidden data like this requires some extra skills or resources or effort to create a valuable knowledge component for your GPT so a little bit of extra effort can go a long way to find this kind of data and finally private data is any proprietary data that is pretty much impossible for regular people to get access to this will often be data collected by companies on their customers sales or Trends Etc some of you may have this kind of data yourself or you may know someone who does you can offer to buy the data off them or to lease it off them and help them to monetize it by creating a GPT for them and share it with the world in that way this idea that people can monetize their private data through these AI agents in kind of a secure way is a very interesting concept to me identifying these hidden and interesting data sets will likely be the best option longterm to create value for most of you however to scrape and clean and sort large amounts of data can be difficult if you don't know how to code python is great for this and you can learn web scraping and data management in a couple of weeks if you're serious about things it's been hugely available for me in my own career over the past 12 to 24 months knowing at least the basics of python and JavaScript but it is a nice to have not a need to have to be successful skill number three prompt engineering the quality of your instructions to your GPT directly determines the quality of its outputs therefore knowing how to write good instructions or prompts is absolutely critical to your success the good news is is that it's not actually that hard there's a few key tricks that you need to learn that will put you in the top 1% of people prompting these models luckily for you I've already done a beginner's guide and an advanced guide on prompt engineering here on this channel that will teach you everything you need to know they'll both be linked below so make a note of it and you can come back and watch those later skill number four tool creation knowing how to create tools or custom actions that enable your GPT / agent to do things outside of the chat interface is key to creating something that is actually useful and valuable actions are one of the three ingredients of a GPT and as mentioned are often referred to as functions tools custom actions custom functions Etc they all mean the same thing if you're not too sharp on the devel V mment side of things they may be a little bit hard to wrap your head around but I'll try to break it down as simple as I can now in the most basic form tools are functions some functions expect an input others don't some return an output others don't but when triggered these functions do something this could be to interact with apis across the web or it could be performing calculations and data manipulations on some of the inputs that it received and more a simple example would be a capitalization function that expects some text as an input and returns the capitalized version of that text inside that function there would be logic to capitalize all the letters before returning that as an output when we create tools for our agents we are creating a function that do a specific thing which your agent can use whenever the conditions are met what makes gpts and agents so ridiculously powerful is how they can handle inputs and outputs of functions when the user triggers a tool e You by saying capitalize this text and then enter some text after the agent is able to pluck out only the text that they want to capitalize and use that as an input in into the capitalization function then once it has run the function and has capitalized the answer it can use that in writing a natural language response to the user and say here is your capitalized version and then input the capitalized version so it's not just going to take the input and then dump the output back into the chat it's actually going to answer the question and use the output as part of that answer in a way that's very very user friendly and natural this may not seem crazy to you but in the programming world this super intelligent and and conversational use of functions is a gigantic Leap Forward in order to be able to do this a decade ago would have been a complete nightmare now here are some more advanced examples of tools that you could create for your agents firstly you could have a stock price feature this tool would take and the stock ticker that the person is looking for EG what's the price of Apple right now and then the function that the tools connected to would uh use an API like the Yahoo finance API and would send off that ticket and would get the information back then it would respond to the user with the price of the Apple stock at that point in time or you could create a solar savings calcul calor which I've actually done in a previous video which I'll link below what this still does is takes in the user's address and their monthly power bill and it sends it off to a custom function that I created that uses the Google solar API to calculate how much they could potentially save in a 20-year time span tools like this are great for lead generation chat Bots for example you can put this one onto a solar company's website they can get their uh Power or the savings calculated then right after that I'd prompted this assistant to ask for their lead information directly after they got their quotation so they were able to capture that lead and then follow up with them after that and finally you could have a generate draft contract tool this would take in things like the contract type the value the recipient address and the name and then use a pre- prompted tool on relevance AI which we're going to learn more about in a second in order to generate a draft contract in seconds using the company's regular format for contracts the good news is that creating tools like these has never been easier for people who don't know how to code platforms like relevant AI that I'm a huge fan of allows you to build complex tools without coding and then a couple of clicks can connect those tools to your gpts and agents in order for your agent to be able to use your tool you need to create an API that wraps around it this is essentially a special URL that your agent can call which will firstly trigger the function that you've made secondly accept any inputs collected from the user in the chat interface and thirdly send back any data needed to write a response to the user therefore the easiest way to understand a tool is to see it as some custom function wrapped in an API that allows our gpts or agents to interact with it over the web cool right I know there are a few different ways you can create tools some are very beginner friendly and others are more advanced firstly you can find free apis and use them as tools you have sites like rapid API which have a bunch of free or cheap apis that are ready to use this means that someone has already created that custom functionality or the function itself and wrapped it all in the API for you to use so it's going to save you a lot of time you can go on to Rapid API pick a couple cool ones that you like and make tools out of them and then connect them to your GPT it's like a beginner friendly way to get started with tooling secondly you can create your own using no code or low code tool Builders as mentioned platforms like irrelevance AI I might go to for creating tools for agents when I'm not custom coding them myself I love the platform because it's extremely beginner friendly with tons of templates to get you started but their tool Builder also has a lot of depth which allows you to create some very Advanced tools as your skills improve if you've watched my videos before you know that I'm very big on this kind of depth concept when it comes to learning new platforms and tools if you're going to buy learning a whole new platform you want one that's not going to have a a ceiling in terms of what you can develop you want something that's going to allow you to make everything eventually as you get better and better and develop a more valuable skill set I actually already have a tutorial on relevance AI tooling for your gbts and I'll link that up here if you want to check that out it'll also be linked in the description and that video is a good intro to tooling with relevance AI I've actually been thinking of doing a complete guide to Tool creation for beginners in a similar style to this so if you want me to do that be sure to subscribe to the channel of course so you don't miss it but also head down to the comments and let me know if you want me to get that out sooner because whatever you guys want if there's enough demand I'll do it and of course this app here connects very well with GPT so if you want to connect your zaps to your agents and opening up really thousands of tool options with all of the zapier Integrations you can use zapier as well finally you can create your own custom functions and host them yourself of course you can just write functions with traditional code and host those apis yourself I actually have a bunch of tutorials sharing my custom tools for gpts like that solar calculator that I just mentioned so if you want to check those out I'll leave the those link down below as well but if you're a beginner I would start off with the rapid API and kind of pre-made API options and the low code noow code Builders like zapia and relevance AI for your tools there's one final component to creating tools which is creating an open API schema this is essentially instructions for your GPT or agent on how to correctly use your API so because your API might take in certain inputs it might be expecting certain inputs and and provide different outputs you need to give your agent a description of of how to interact with that API which is what the open a API not open AI open API schema is this is an absolutely crucial step and can seem a little bit tricky to beginners but I've got some super simple techniques you can use to quickly create your schemas which I'll be sharing with you once we get to the building phase of this video and as mentioned in the next chapter of this course I'll be walking you through creating your first tool and connecting it to your GPT we're almost done but there is just one more skill that you need to know before you can get started as a GPT store Builder skill number five GPT marketing the final piece to the GPT puzzle for anyone looking to make money on the GPT store will be marketing you can make it the most amazing GPT on Earth but if no one knows about it then you won't get any usage therefore you won't make any money while the GPT store will undoubtedly have some kind of discover feature that will Spotlight the top gpts or open AI picks but relying on this kind of exposure is not a very reliable method for marketing your hard work like I said people who have an audience or can use social media to get eyeballs to their gpts will gain more users and therefore rank higher on the list and therefore gain more organic users from being on this list Etc it's like an upward spiral of success for example the grimoire GPT which is a 100x coding assistant that I mentioned earlier in the video was cleverly marketed across platforms like Reddit tool sites like super tools and heavily on Twitter if you don't already have an audience then this GPT store moment might be a great opportunity for you to start focus on just one platform like Twitter and specialized in creating handy gpts or just one really good GPT for a specific type of user this guy has clearly focused on making a GPT that's handy for people who want to write software but aren't technically capable with the audience he's built up on Twitter he could now release new ones and new gpts that solve different problems for that same audience and likely see similar success one thing I haven't mentioned yet is that on the GPT store you will have a builder profile which is where all of the gpts you create will be attributed to with these Builder profiles in mind niching down to one industry or one type of user could be very handy so that you can become the go-to guy or girl for any gpts within that Niche so how can you Market your gpts the main ones I'd consider would be Twitter Youtube LinkedIn and short form content be it through reals or shorts or Tik Tok depending on your end user you will need to choose a platform that matches best with who those users are for example if you created a networking tool that would obviously be best marketed on LinkedIn with written content and video posts however if you had H gbt that helps students maybe that would be best marketed through short form content platforms where there's a lot of students and young people watching content you get the idea I highly recommend that you just focus on one platform at first as too many beginners try to be good on all platforms end up being bad at the all at once focusing on just one platform like Twitter or LinkedIn or YouTube and getting really good at creating content that generates new users for your gbts is the best way to go in my opinion I mean in my experience all I started with was YouTube and things have gone pretty good for me we only really just scratching the surface here in terms of GPT marketing ideas and I think we'll see some very interesting marketing methods over the next couple of months and years you could create full landing pages for your gpts to explain its value like the the data used the actions included and then even like run traffic there from Google or Facebook like so many different possibilities here you could even create a funnel of gpts for one Niche where each of them has been prompted to kind of link to the other GPT when a certain topic is discussed so kind of like interlinking your GPT so that they start using all of the ones you offer for that Niche if you get what I mean you can prompt it and provide links to them yeah anyway getting a bit distractive hey guys sorry to interrupt video video but before we get into the creation module I just want to ask you to head down below and leave a like on the video if you're enjoying I put a ton of work into this video as you can probably tell it's a lot more to come but if you also want to subscribe to the channel while you're down there a lot of you guys aren't subscribed to the channel or watch my videos anyway so if you want any subscribe so you don't miss out on more stuff and support me to doing to put this kind of free stuff out you can please subscribe down below like leave a comment if you are as well it's all for me we can get back to the video now' been brought up to speed on everything you need to know about making the most with the gpts and AI agent opportunity we can get into the final chapter of this course which is creation I'm going to be taking you step by step through creating your first GPT using all three ingredients and then I'll even go one step further and show you how you can Jailbreak it into an AI agent you can sell to businesses in multiple ways no other YouTube is going to be showing you that last part before we can create our GPT though we need a use case or a concept so that we can start to assemble the three ingredients for it for the tutorial we're going to be creating a Stock Investing assistant based on the wisdom of Charlie Monga Warren Buffett's right-and man at birkshire heatherway all of the resources including the prompts the documents and the tool templates will be available on my free resource Hub in the description below so you can create a free account on my resource Hub and get instant access and follow along if you're looking for the resources for this video it'll be on the left hand panel and down near the bottom somewhere for our planning we need to look through the lens of the three ingredients of course so firstly our promp thing is going to instruct the gbt to be an investing assistant that helps users to invest according to the principles of Charlie Munga using the documents providing we are not going to be authorizing it to make investment decisions on behalf of the user rather just provide some options so that the user can come to their own decision this is a bit of a legal gray area so it's important to not get yourself into any legal trouble by making financial decisions or or recommendations when you're not technically a financial adviser secondly our knowledge ingredient will consist of four documents that I've gathered the first two are PDFs on Charlie's investing principles the other two are transcripts from interviews I found with him on YouTube and finally our action or tooling ingredient will be a realtime stock data fetching tool that allows us to look up the price of stock in real time this is the most basic implementation I could think of for a Stock Investing assistant CU I think it's the easy opener for you as a beginner but if I was going to be making this GPT I'd probably use some kind of historical stock data tool that could get the past maybe 3 to 6 month of stock data which would allow the GPT to then kind of look over the stock data and analyze the past performance compared to other stocks past performance and then provide better recommendations based on this data so I think if you're looking to use AI as kind of a a data analysis and and reasoning engine that's probably how I'd attack this if I was going at it in my own time but for now for some publicity sake and to keep this easy for beginners we're just going to be doing a stock price fetching tool within this gbt so if you want to follow along head down to the resource Hub sign up and grab those documents then you can come back here and we can get started of course to get started we are on the chat GPT site here now in order to make gpts you do need chat GPT plus which is open now at the time of filming so make sure you sign up to a chat gbt Plus account I think it's $2 $30 a month or something then we can go to the left hand panel here if I go to explore gpts here is the new update of the GPT store at the time of filming the earli section of this video this hadn't released so here we can take a little look I guess while we're filming this um we have a featured section A trending section by chat gbt Etc doesn't seem to be anything crazy at the moment but it would be great if you could get on this featured or trending section all of you should take a look through here and see which ones are trending how they're being built and how they're being marketed as well but to get started on building our GPT we can click up to the right corner where it says create then we get this screen here so we can start to go through the process of creating it when you are creating gpts there is a create and then there's a configure section the create experience is a lot more conversational you can basically go through a conversation here with your gbt buildup and it's going to help you to construct it and convert your natural language into the ingredients here that we can actually put in ourselves directly so because we are smart GPT developers and GPT Builders we can jump straight to the configure page and start to build them quickly but also build them very specifically so we can tell it exactly how we want to do the prompting Etc when you do it through the create tab it's going to kind of make up the prompt along the way but we can be very specific with it here here we can give it a name we can say Charlie anger inting assistant the name of course is what's going to be shown on the store and some of the naming convention documentation that open AI is provided they say it's not recommended to put GPT in the name so in this case I've just called it Charlie M investing assistant um I can give it a description this is not going to affect the functionality of the assistant at all it's just going to be as you can see on the right hand side here this is the sort of description line of the of the GPT so I can say get help with your invest investing decisions using The Wiz Dum of Charlie manga and then we get to the instruction section which is the prompting ingredient of our GPT I already have one written here that's going to help us a lot um I'll go through this with you now you are an investment adviser that uses Charlie monga's wisdom to recommend investment strategies you are wise and have a long-term outlook on investing I'm going to add a little bit here for the legal side of things um you are not permitted to make decisions for the user only to provide options for them to choose between s then we reference the documents we've provided so a little bit of instruction on how to use the custom knowledge the knowledge ingredient that we're going to provide I've noticed it's very helpful to make sure it's using it at the right time so here I say you have been provided with documents containing transcripts from manga's interviews and also PDFs that are packed full of manga wisdom use this information to answer your investing related questions so that's telling it how to use the knowledge component which is something I recommend you all use when you're writing a prompt for your gpts if you're using a knowledge component and then this final line refers to the action or the tool that we're going to connect to it um you have one tool to help users get stock quote this function features the real stock data for a stock ticker provided by the user e Apple Microsoft Etc so that's telling it how it can use a tool which we're going to build in a second conversations are something cool you can add in for uh to enhance the user experience of your GPT I typically if I'm using it just a quiot prototype I'm not going to be writing them but if you are going to be putting these on the store and making it as user friendly as possible these conversation starters are an opportunity for you to really showcase the best of what your GPT can do in a little click button format so whatever I put here um I can do if I think about the most value that this gbt could provide would be uh probably related to the tool use um can you get this price of a stock for me so that's related to the tool I can say what are your five top investing principles for begin can you analyze this uh stock performance for me then then the user could paste in some recent stock performance data of the prices over the past 3 months or something um and then we can add on another one where it is how can I um evaluate a stock before investing these are all pretty helpful conversation starters I think think for someone who's looking to get assistance with investing getting a stock price analyzing the thoughts Performance top five investing principles for beginners evaluating a stock Etc these kind of things are going to pull from the knowledge base we use the tools or anything that we' provided as part of these ingredients to help them to do a certain thing related to investing then we get to the knowledge section which open AI has made very easy for us we can just upload the documents I will mention here that some types of information like like CSV and other files like this if it's a spreadsheet or any kind of numerical data the models are not that great at understanding that stuff right now um I assume that open a is going to release something that allows them to better understand or process these kind of numerical data sets at the moment I recommend anything text based so PDFs documents transcripts anything you can imagine that's text based that's going to be able to use that kind of knowledge much more effectively than your kind of numerical data at this point as I mentioned all the documents are going to be available for free on my resource Hub so if you want to follow along you can go down there and sign up and get access to these you can download them and you will get access to these four documents here so I've got two PDFs and I've got uh two documents the PDFs are from the internet and then the other two are transcripts I'm going to upload those there then we're able to enable some of the capabilities that open AI offers out of the box from their side to make opts more powerful at the moment this includes web browsing with Bing uh valy image generation and then code interpret as well so at the moment I don't think we need image generation web browsing could be handy essentially our stock data Checker tool um could be replaced with this web browsing one cuz it could search and find the price but for the sake of this tutorial I'm going to leave the web browsing capability on so I can pull some more data into the conversation using web browsing and say hey can you find me five stocks that have been in the news lately and pull those into the conversation then I can do sort of analysis and and pull data in ETC so we're going to leave web browsing on and then we get down to the action section which is where we can add in our custom actions which is what we're going to do now so if I click new action in order to create our action we need to find an API that we can use and then we need to create a schema for it which is what we're going to be doing now and as I mentioned one of the easiest ways for you to create your own tool and start creating custom actions is to find an existing API and then include that as one of your actions for your gbt so for a stock data price fetching tool I searched around for a free stock data API and I found this one here which is called fin hub.io no affiliation I'm just using it cuz it's free so you guys can get into it too if you're following along you can come to fin hub.io we can click on get free API key now I've already made a free account but you can do the same and you should be able to see the screen here which is when we get our uh free API key here which we can start to use in order to create our action or we need to know which particular API or endpoint we're going to be calling so thisf hub.io has a lot of different stock data related uh endpoints we can use this is something you need to get familiar with is looking at documentation and learning how to use these apis and I promise it's not as scary as you think so um if I go down here I think I saw one that I like which was stock price quote um so this is going to give us real time quote data for you with stocks constant polling is not recommended to get request and this is how we can use it here so essentially all we need to pass in as an input from the user end is going to be uh the stock ticker and it's going to give back the the current price the change the percentage change higher the day lower the day open the price of the day and the previous Clos price so we get this information back and all we need to do to call it is a little get request using this URL like this so if we decide to use this quote endpoint what we're going to need to do is collect on our gbt side A teer or a stock tecker say AAP for Apple from the user and the assistant or the GPT is going to take that include that in our API call send it off to finhub it's going to get the stock data it's going to bring it back and it's going to give us this payload we're not going to see that as a user but this is what the GPT is going to be seeing as the response so it'll give us back the information and then depending on what we asked say I ask what was the uh today's high the GPT is going to intelligently be able to pick out the high from this data and just send that back to us so that's what I'm talking about earlier in the video where I said it can very intelligently handle inputs and outputs and when is that we haven't really been able to do before which is what I I love about this stuff so much so now that we found an API that we want to use we now need to convert this into an open API schema as you see over here here's some examples if I go to the Json one we need to convert this API into a machine readable documentation or outline of how to interact with this API um this is what they look like open API title uh paths Etc so we need to convert house into something like this now this can be a very difficult process for beginners there's a lot of intricacies and and gotas that can really catch people off guard so I found which while for someone like myself or other people who know how to code it might be a lot easier to generate these but in preparing for this video I really struggled to find something that I was comfortable sharing with you all because there's so much complexity in these schemas I tried to create a gbt that would help you to create them very easily but again there's too much complexity so the easiest and most cost- effective way that I was able to find for you as beginners is to use a separate tool a separate platform that essentially put a wrapper around an API it allows you to easily input the information and give you a schema in a couple clicks and this is using a platform that I've talked about before on the channel I've done videos on them theyve been a guest Inside by accelerator so I'm very familiar with the platform but it's called relevance Ai and what I'm using relevance for in this video is to do exactly that which is to take an API like we found here and then put a wrapper around it and they automatically generate the schema for us which is great news for beginners and it's a really easy way for you to get your foot in the door and start building these tools this is not just a big sponsored video for 11 AI I literally took the whole night last night TR to find an easy way for you to do this and I came to the conclusion that relance was the best way and that's why I'm sharing it with you because I think it is the best thing for you as beginners so to convert our finhub API inpoint into a tool we can use in our GPT we need to hit to relevant ai.com and here you can sign up to a free account yes they do have a free tier which is another reason I like this method because I have a free tier which is going to allow you to get up to 10,000 runs uh for free so you can prototype and test out your own agents and start to use tools and custom Tooling in a very easy way go ahead and sign up for an account I'm just going to loog log into mine and then you'll see the home screen that looks a bit like this I'm not going to go through explaining the whole platform I'm just going to show you what you need to know to keep this short we can head to the tool section and up in the top right we can click create tool start from scratch and then we get to the relevance AI Builder I do have a whole video showing you how to create tools with relevant AI which I will link in the description but for now I'll show you the quickest way we can do it for the particular API that we're working with first thing we need to do is provide a name and a description so we need to be very descriptive about how we title this because anything that we put in here is as natural language is going to be taken over to our open API schema and you can see here we have a description and we have the title and descriptions this natural language included in your open API schema is what tells you GPT what this tool is and when it's supposed to be used so if you see here on this uh location tool the description for it is get temperature for a specific location and then further down we have the parameters which is the the name of the location we've got a very good description for it which is a city and state to retrieve the weather for with without giving accurate descriptions of the tool itself and of the different components of the tool your actions are not going to work as expected so it's very important when we come up here we give it a a accurate name so let's say uh fitch stock Pro quote data the quote data is all of that high low close Etc data not just the price I'm going to say fit stock quote data uh bu stock tier we give it a description um fitch stock quote data by ticker EG EG Apple I think that's good fet stock quote data by ticker we can ignore the knowledge component we can add a user input here I'm going to add a text input and this of course is what the uh GPT is going to be plucking out and using as an input into the function so um here we'll have a stock ticker uh ticker of the stock to Fitch data for EG again being very descriptive here because this is going to be carried over into the schema for us automatically now what I like to do is just change this this is the name of the variable so when this is entered into this automatically by the gbt um for our sake when we're building this it's nice to know what the variable is so this can be called ticker and then when we go down a step we can add an API request and now it's instead of having to convert all of this into a schema that's quite a complicated process we need to do a plug in the details for this API as we normally would and then it's going to automatically convert that for us handy way to do this is to set this to curl and then we can copy this URL head back a get request you can see here um method is get so yep we're good to go there paste in the URL and then what we want to do is automatically oh we have a little quotation at the end there automatically inserting my API key this token here represents our API key so if we go back to here I can copy my API key and replace this boom oh it was already done maybe it's already done it for yours as well um but it might be good to check and then all we need to do for this one is to replace the symbol here I can delete that and I can insert the ticker so that's going to dynamically insert this ticker variable that our gbt is passing it's going to take it out of here insert that variable into the symbol we're looking for and we've hardcoded our API key in there and it's going to return it as the output here so let's give this a little run if I go uh msft and I go run all and just play that we've been able to plug in an API we found put in the dynamic data and get the output from it that we were expecting I will mention now that this kind of API is quite a basic one um often times you'll have to have in header which is like or theorization and then you paste in your key it's a little bit outside of the scope of this video that is definitely something I'd like to address in future so if you want to learn more about making tools more depth as I said I do want to do that full sort of hour 2our long video on F creation and I'll likely use relevance AI for a lot of that make sure you subscribe to the channel and let me know Down Below in the comments that you want to see that because there is a lot more to it but I'm just trying to keep it as simple as possible for this tutorial so now that our tool is working we've got all the descriptions done correctly we can hit save and then to connect them to our GPT relevance AI has made this super simple for us if I go back then I can click this custom actions gpts button then I can click this button then I can select the top one here now the cool thing is that you can select multiple I'm holding the command key so you can build multiple tools and relevant Ai and combine them all into one schema and then pass that into your gbt as well but for this case I just want to take this top one and then here is the magic dip that saves a ton of headache for you as a beginner generate open API scheme and it's automatically generated on their end I'm not going to go into dip on on all the information here I'm just going to copy this hit over the chat gbt I can delete this I can paste this in I'm going to format it and there we go so that's all in there and as you can see our descriptions have successfully been passed in there we have the operation ID the summary the description Etc which has all been passed over well and here we have things like the description of the stock ticket so this is one of the inputs stock ticket of the stock to Fitch data for ETC so all of our descriptions and hard work there hard work I mean it was like a couple words but have been par over into this so we have the natural language to describe what we're trying to do final step let to go back to relevance AI we can click on generate API key here we need to set up some kind of authorization that allows lgbts to use our elevant AI so they can build us a t so we can generate that and click copy and there's a little tutorial here that you can walk through um I'm just going to show you it without watching it but if you get Stu you can click here it's got a nice little walk through um we're going to come back here click on this Cog for the authentication step we can go API key we can go custom we can type authorization with a z and paste in our API key save now that we've got everything set up we can test this action so down here there's a test button press that give it a second I'm going to confirm it and if you click the drop down here it's super helpful to see some of the information okay so lgpt is sent away AAPL as the ticker input as we can see here we have the ticker input which is correctly passing some information off to very helpful for you as you're debugging this if anything goes wrong is to click this drop down here you can also see some other stuff here like okay what response data are we receiving um how are we calling it okay this is what we're sending off as params Etc so this is the data that we're getting back as I mentioned it just sends it back in kind of a lump of of Json it is and now gbt is smart enough to take that response and convert it into an answer that actually makes sense so here we can see as of latest data apple is $185 59 which as we can see if I go here 18559 18 18559 so it's working which is great um and that's how we can get our first tool set up within the GPT Builder here one last thing that I like to do just to make sure everything is is died in properly as you can see up here it is using this uh called the XYZ API uh with the DAT operation so what I can do is copy this it's obviously understanding from the references and descriptions we put in the schema what operation we want to use I'm going to copy that and head back to our prompting here as you can see I've got get stock quote I'm just going to replace this with uh fetch stock quote data by stock ticket which is the exact name as you you can see in the in the actions here as you can see the operation ID is fit stop quote data by stock ticket so if you want to make sure it's using the right API and and calling things correctly it is helpful to take that and go back to your instructions and include it somewhere so there's only one thing left to do before we call this a day on this GPT build click here we can use doly to generate a picture see what it comes up with now to generate this picture I believe it's only going to be using the name here and the description um so if you want to change the prompt thing for your image generation you can change that uh using these two Fields here I might use the instructions as well I'm not entirely sure seems to be taking a while I've noticed this recently with the IM generation so let's uh let's improvise we get a cool little pick of Charlie here looking badass upload photo desktop there we go I think that's better than anything Dary could have cooked up for us at this point we're almost ready to put it onto the store if I come over to the save button here can publish to just use yourself anyone with the links you can share it around to your team here when you hover over everyone it will say your GPT will appear in the explore gpts page if you click on the little pencil here you can see that we can start to modify our Builder profile here I have my name as Liam C otley populated from my building details you can also connect the domain here so if you wanted to link this back to your personal brand or your your own personal GPT Builder site which is entirely different thing that we can get into in another video you could have your own like GPT Builder profile site that's a landing page that shows the links to all your different gpts that kind of depends on you being niched down to a certain type of GPT or type of user that you serve anyway getting a bit distracted there uh receive feedback emails to your email there I'm not going to check this that's your Builder profile and it's most basic form I assume it's going to become a lot more uh in-depth over time but this is all you need to worry about your name and your website if you want to connect it and your feedback emails if you go down to our actions you have a privacy policy URL what open AI needs for your privacy policy is a link to a web page that has your privacy policy on it so to make this short and tweak for the video I'm going to show you a little bit of a hacky way to do this what we can do is come to chat gbt first we need to generate our privacy policy so if I create a new chat write me a PR see for my GPT called Charlie manga investing assistant and what I'll do just to give it a bit of context but I'll paste in this here here's some info on it there we have a quick privacy policy that's been weped up by chat gbt for our assistant now I'm not sure sure how strict they're going to be on this whole thing but just for the sake of this tutorial and just getting these up and running I'm sure there's a whole video that needs to be made on privacy policies and stuff like that um but in this case let's get it up and running let's copy that then we can head over to something called design a GPT which is what I mentioned in the intro um but this is a GPT that's going to allow us to create a website very quickly and it's actually going to host it for us so if I say create me a web page for my privacy policy and include all the the following Texs on the web page so then we can paste in our privacy policy then I'll put in contact details let's just go Liam mail.com of course use your own one and then uh privacy policy we can R some of this last updated at what is it [Music] 12124 and then see what this can create us I'll leave a link to the designed GPT in the description below so you can do the same now can confirm and then hopefully when we click this we will have our privacy policy boom just like that so now we have our privacy policy posted on the URL I've got them buttons up here that we don't need to use but this is a very basic and quick way of doing things I think design gbt is great for this um cuz hosting a web page and getting it up can take a bit of time if you're not familiar with how to do a lot of the stuff so there you go so we have all of the information I'm going to copy that link head back to chat GPT go down to our actions and go to the privacy policy and paste this in I can go up to save I can click on everyone and it's now available to be published there have my Builder profile category yeah probably research and Analysis is correct and then I can confirm and have go to publish my GPT to the store and you're officially a GPT store developer now I can give a go can you get the price of stock for me let's give this thing a try and expected it's asking me for the ticker so that it can use the tool use the function so it's going to try grab that input and then pass it do to our relevant AI tool confirm so our stock beta check is working correctly now that it's been published P refresh let's try another one of these conversation starters um how do I go new chat new chat um what are your top five investing principles for beginners now this should search the knowledge base and give us some Charlie Monga wisdom back that's friendly for beginners but I like to test both the tools and the knowledge base component cuz they they are two separate things that the the assistant has to use so it's good to test both as an investing thing Guided by the wisdom Charlie manga I recommend the following PR investing following five investing principles for beginners patience is key understand what you invest in emotional discipline learn continuously minimize stupid decisions Charlie Super Wise on that last one then minimize stupid decisions anyway so if you guys want to have a play around with Charlie Munga investing assistant and maybe give me a little bit of a usage run up on this so I can get some of that uh possibly non-existent Revenue sh there'll be a link down below if you want to play around with my investing assistant here at any point if you want to go and check on your GPT that you published you can click on explore gbts here and then up in the top right click on my gbts so here I have Charlie M investing assistant and as you can see it has zero chats and it's available to everyone so when you guys go down below and click it and play around with it I'm going to get some of those chats that stack up and if essentially that'll be the usage metric that I assume at some point you're going to get paid for so now that we've created a working version of this agent as a GPT now I want to show you how to jailbreak this GPT and use it outside of the chat GPT site this is a crucial step for you if you want to start selling gbts to businesses and you're going to need to create agents and deploy them to places like their website or their slack or their WhatsApp business line or their Instagram DMS Etc I've added in this next section of deploying to different platforms because this skill is something that very few people know how to do right now and this Market of deploying AI assistance is largely untapped I will mention now that the assistance API which is what allows us to create gpts that we can use outside of the chat GPT site it does cost quite a lot in terms of usage and it can be quite slow sometimes making it not really ideal for all use cases for businesses at the moment but this is to be expected with this new kind of technology and I expect open AI to push an update within the next few months that's going to reduce the costs of running these Deployable agents and to also decrease a response time and therefore making them a lot more attractive to businesses so we are early but I'd rather be early than late to get you started with deploying agents and gpts outside of the chat GPT site I'm going to walk you through how to put one on a website as a chat widget like down the little corner one of those little bubble p knowing how to do this kind of chat widget deployment opens up opportunities for you to sell AI assistant Solutions like customer support or lead generation agents to businesses and normally this kind of deployment takes a lot of custom code and messing around with multiple different platforms to get it done right on this channel I've made a bunch of videos on the topic already but many of these tutorials I found were were far too complicated for beginners a lot of custom code involved Etc which is why myself and my team at Morningside AI created agentive which is the easiest way to create deploy and manage a AI agents and no I didn't wait this long just to give you another sales pitch agentive is 100% free for everyone right now because I want you to be able to learn this valuable new skill without any Financial barriers agentive is in Early Access right now and it only has a fraction of the the features that we wanted to have eventually but it is enough to get you up and running and start deploying these onto things like websites to start with so what we're going to be doing is taking the exact same Charlie manga investing assistant that we created but we're going to duplicate it over to be on agentive so that you can start to deploy it to wherever you want and it on different platforms to start off you need to come on to the agentive landing page here it's going to be link the first link in the description should be the one to agentive so you can come and get this at the moment it is as I said An Early Access but if you apply to the waiting list by clicking this join Weight list button click here if you fill this out you should be allowed into the platform within 24 hours so you receive an email when you've been added to the wait list and then within 24 hours you'll be let in to actually start using the platform so keep an eye on your emails after you've applied to it so I'm sorry I can't leave you in straight away but there is quite a big cue that we're trying to get through and and we don't want to let everyone in all at once because it is an early access once you have got access you'll be able to put your email in here click sign in then you'll be emailed a link to a magic link to sign in automatically so go to your email click on that and it will automatically log you into the platform and here we have the interface for agentive now there's a little light mode dark mode here if you're watching this at night but this is the basic interface we have at the moment what I'm going to do is Click create assistant up here and we're going to go through the process of creating what we just made on the chat gbt site but on agentive and as you can see we've closely mirrored the experience so it's very familiar for you um we do have a bit of an ux overhaul coming in the coming months so keep an eye out for that however in order to unlock this GPT 4 1106 preview model which is needed to create your own GPT on something like ative you need to first go to platform. open.com and we need to unlock your account so you can get full access to all the models available so if I log in here or create an account if you haven't already and then we go to the side here and click on settings we can go to billing and here you'll see that I already have topped up $15 into my account and I've added a payment method so what you need to do to get your models unlocked so you can use ative and start creating your own gpts is just to add a payment method here then add at least $5 worth of credit here using this recharge and i' also recommend you turn on the auto recharge so that if you use up all of those $5 it will automatically top you up so be aware that as you are using this you will be build for API cost so just keep an eye on this to see how quickly you're spending there once you've added some credit you can go to the assistance tab here and we can double check that you have the model activated if we go create by click on the model dropdown you should see GPT 4 1106 preview if you see that you're ready to go back to agentive and get started and creating your own GPT but we can go basically step by step and let's go back here and edit this GPT and now we're going to be duping over this entire thing so I can go Charlie manga investing assistant go it there description copy this come back and get the instructions copy that put it there update that then I'm going to go to the knowledge I'm going to enable the knowledge base going to add files we need to add them in one at a time um this just a feature that we're work on at the moment but adding them one at a time ensures that we don't get a queue which has caused some issues in the past so we need to go through add the second one go down and add the third one and again these are all the same documents that are available on the resource Hub and the fourth one it's going to take a little bit of time for these things to process um but until then we can go and connect our tools once we're here we can click uh choose new tool provider and now we can go back over to relevance Ai and the exact same method we just use for our GPT but this time for agentive so if I come up to the custom acction for gpts click on this generate open API schema copy this head back to agentive paste it in it's already figured out what we're looking for I can go add authentication do the same process if I go custom and I go authorization head back to relevance AI generate me a API key copy that hit back paste it in here update the authentication create the tool tools and now we have our tool created of course it's just one but we can have multiple Tools in here we do have support for multiple tools and if we go to the knowledge I'll give this a little refresh should have processed most of these yep they all been processed correctly so if it says it hasn't been processed just refresh the page and it should update if it's been long enough tool set up everything else set up and we can give it a test one of the things that makes a genive different from any other platform on the marker right now is that we have this testing environment now it may not seem like much but if I go in here and I say hi what can you do not only can we chat to our assistant in here we can test both the knowledge based feature and our tools if you go over to the open AI playground and try to do the same thing you're not actually going to be able to call your tools that youve provided to it it's just going to kind of um okay yes we've triggered it but we're not actually going to process the the request and actually perform that tool what makes this different from the open AI playground which allows you to do a similar kind of setup process that we can actually conversate with our assistant here and test the tools that we've set up so I've provided access to our get stock data what I can then do is say can you get the price of um Microsoft I can actually test my tool in here which is something no other platform offers right now so gent is designed to be this platform where you can come in as either an agency owner and you're selling these other businesses or as a business owner coming in and wanting to test thing before you put it live you can have a full testing experience of how your user will get it here within the within the chat window so how we can see the current stock price of Microsoft it's successfully fited all the information we're looking for so we know that it is working and ready to go here if someone who's going to be building and selling these to businesses I can tell you that this is going to be an extremely useful prototyping and testing interface for you which is why which is how we built it now we're ready to deploy this onto a website up here we have the deploy button which I can click and if I click on the website option here um here's the information we're going to need you're only going to need a little bit of it so it doesn't look too complicated up here we say first time deploying to a website watch our tutorial here so if I click this we have a little tutorial here which obviously you don't need to watch cuz I'm I'm giving you a tutorial right now um but we have two buttons here which is one to download The Voice flow template and the second to sign up for voice flow so our recommended method to get these put onto websites is to use a platform like voice flow we have a template that makes it super easy for you to integrate with so if you haven't already you can click here and sign up to voice blow but otherwise you can click the download The Voice plow template button here and you'll get the template that's ready to upload and make this super easy for you over here I've already got an account for voice of course once you signed in and you've created an account you can click on the import. VF file up here and I'm going to go to my downloads I have the website template click on that I can open the assistant here now that it's created and there just a couple little things that we can change here and then we're ready to go so back in agentive here you can see it says replace your API key here if I go back to agentive here we have our API key we copy this head back to uh voice blow and click on this and down here we just double click here and replace our API key easy as that and replace our API key and assistant ID here click on this replace our assistant replace our API key again and now we just need our assistant ID which we can come back to agenta for and if you come down here you see your assistant ID assistant ID copy that head back to voice blow double click here paste in our assistant ID and everything else has already been set up in this template I created and then we can play around with a bit of this so hi I'm uh Charlie your investing assistant and we can change things around like I can help you with investing advice and also F realtime stock data for you just give me the ticker you're looking for EG ms50 now we should be able to give this a spin and see if it's working to test this you can go up to the top right hit run go run test and then what this is going to do is create a new session and this is basically going to put all messages that come through under this particular chair hi I'm Charlie your investing assistant I can also help you BL how can I help you today and what are your top five tips for beginner B let's check that the knowledge base is working and this of course is connecting to our agentive API here going to be pulling all of this functionality into this here and then we have it so when you see these little Source things pop up that means that it is using the knowledge base and you can remove these with a little JavaScript trick on here uh but for now we'll just keep it as is so here we have it working correctly and pulling our functionality from agentive now we need to take this and put this onto a website which is a fairly easy step from here so if I go back I can go to the publish button I can call this V1 and I'm going to click on the embed widget button and here we can modify the layout and the design of our chat widget so down here you can see a little demo of it um if I go Charlie Monga investing assistant so that's going to change the name over here let's go M so it's not cut off Charlie manga mango investing oh um forget after tab is closed and reloaded let's keep it there let's change the image of the assistant to be it's cool picture we got of CH and we can change the color let's make it oh that icon okay let's change the color to something a bit cooler dark blue maybe it matched that little pink that's kind of a nice looking dark purple I like that um change this assistant Avatar put it time as well and then we have our manga investing assistant ready to go now what we need to do is copy this and put it onto our website now just demo this I'll go free CSS I'll drop this link down in the description below it's just a free template that I like to use in these videos makes it a lot easier um then if I download this up here I can click here un zip it and then if I open up brackets uh you need to go to brackets.io I think you can download brackets the free code editor that you can use to uh pull up these website quickly and then in Brackets what I need to do is go to um open actually we can go to Open folder and if I go back to downloads open up this press enter on that now we have access to this if I click on the index. HML again I noted this is no code but this is literally like just copy and paste to get in So at the bottom of this index.html the same thing applies for whenever you're putting one of these assistants and adding it as a widget going to be scrolling down to this bottom section where you see all these scripts and then I'll just drop into here make a little bit of space and then I can paste in if we go back to this we can copy our web chat can paste this in here I can command s to save and then up here bracket allows us to do a live editor if I click cck this live preview it's going to pull up the web page for me now if I come down to the bottom right we see our chat widget here pops up hi I'm Char you're investing assistant I can say okay let's try do a tool call what is the price of Apple it's going to talk to our agentive system that we've set up and it's going to send back the response using the tool call that we've set up for it current price of Apple is correct great so now let's try it a little more complex what is the current daily price change for IBM and Microsoft compare the two little bit more advanced cuz of having to use the tool call twice so once on IBM second on Microsoft doing a process of comparing those two and its output daily price chain is current daily price at this the stock was trading at this when comparing the two both stocks experienced an increase in stock price Etc so they you have't that is how you can put your own assistant your own agent that you've created um both the same as we did on the chat gbt site but this time being able to jailbreak it out of chat gbt and putting it on a website now this same approach applies for taking it on to things like what's that business or Instagram DMS or slack or any other deployment that you can imagine for an AI assistant that is what ative is all about and that's what we are rolling out in the next couple of weeks it to be able to deploy to the different social platforms and other platforms that businesses use as well so you'll see this deploy tab we not just have website but a bunch more so if you want to get on and have a play around with it then it's going to be 100% free as I said before you can hop on here and follow this along step by step this has been just a super brief taster on how to create gpts and how to use that exact same skill set to create and deploy powerful AI agents for businesses follow we have covered a lot in this video I have had to leave a little bit out when it comes to the complexities of tool creation and Tool usage and APS authentication all this sort of stuff it's a little bit too advanced for this video I wanted to give you a beginner guide to get you started so you can take this exact same method use relevance AR for your tooling you know how to do all that now but if you do want to see me do an entire guide on how to create advanced gpts and AI agents with tons of real world use case examples then be sure to subscribe and let me know down below if the demand is there to create an in-depth guide on Advanced Tool uh creation and things like that then I will make the video for sure so let me know down in the comments below subscribe to the channel if you haven't already and if you've got any value while you're down there please leave a like on the video it would mean a lot to me and also a reminder that agentive is 100% free for the time being so if you want to sign up and have a go you can head down below to the description and sign up there within 24 hours you should be led in so that you can have a play around so what are most important takeaways from this course firstly AI agent creation is going to be one of the most valuable skills to make money online in 2024 and Beyond not only because you can create powerful assistance to make you more productive in your personal life but you can also build and sell AI agents to anyone who needs them I believe if you're going to learn a new skill and go all in on it it is crucial that you're learning something that you can make money from for the years to come since all the big tech companies are rushing to create AI agent platforms by mastering the use of the three ingredients and the five key skills covered in this course will put you right in the middle of the AI industry with a skill that is in high demand secondly while the GPT store may not be a great option for some the bigger opportunity is taking that same skill set and using it to create agents for businesses and integrating them with their operation just as people made million building websites and building mobile apps the same need is there to be building AI agents over the coming years for businesses and finally if you are going to go all in on this new opportunity be sure to check out agentive it is free to start deploying your own AI agents and selling them to businesses as you need I want to end this video with some action steps for each of you so you know exactly where to go after you finish watching this video firstly subscribe to the channel if you haven't already as I said I have a ton more super helpful info coming on things like tool creation GPT marketing that will be hugely valuable for you moving forward and while you're done it please leave a like on the video if you've enjoyed you can join my free telegram you can join my free Discord you can sign up to my free newsletter and get all of my free content whenever you need it across all my platforms all I'm talking about is how you can best make money and build business businesses in this AI boom that we're experiencing so if you want to get all that information be sure to sign up to all those in the description below as I did mention earlier I do have a paid Community if you want to get in there where myself and my coaches and my team teach people how to start and scale their own AI agency where they sell these kind of solutions to businesses over the coming weeks both myself and my coaches will be breaking down these key skills covered in this video within workshops for our paid members so if you're interested in getting access to that you can sign up below and get further help on starting your own AI business secondly start researching some use cases for gpts and AI agents so you can try out that Reddit research method tool I mentioned called Gummy Search or you can go looking for some interesting data or apis that you can use to create tools for your gpts thirdly you should then go and create your first few gpts and get used to combining the three ingredients in different ways to get different outcomes and number four once you've got the hang of it you should consider picking a niche or area that you want to specialize in and go deep on creating gpts that help that one kind of user and number five you can then start to build your building profile around those kind of people and begin building an audience on a platform of your choice by sharing helpful content on things like Twitter or YouTube or LinkedIn or short form content number six you can then look into how deploying your gbts onto websites or other platforms via agentive can help you increase your income further if you want to learn more about starting an AI business you can check out my complete 2hour guide here taking you through step by step but that's all for the video guys thank you so much for watching and congratulations on making it to the end I wish you all the best for 20124 and we'll continue to do all I can to help you make money with AI and build a life you truly love living I'll see you in the next one oh

Channel: Liam Ottley

Views: 428,032

Keywords: chatgpt update, gpts, how to create a custom gpt, create a openai gpt, openai, openai devday, ai tools, build custom gpts, openai gpts, chatgpt on custom data, openai function calling, ai business ideas, assistants api, assistant api, gpt assistant, openai assistants, website gpts, agentive, gpts tutorial, gpt store, openai gpt store, gpt store launch, how to build gpts, custom gpts course, gpts beginners guide, make money with gpt store, gpts course, gps course

Id: Hh2zqaf0Fvg

Length: 81min 23sec (4883 seconds)

Published: Wed Jan 17 2024