How to Create a $32,000 AI Chatbot in 18 Minutes (No-Code)

Hey everyone, Liam Mley here and today I have another super valuable video for you all showing you how to build a $32,000 AI chatbot that's been built and sold successfully within our community. So I'm going to be walking you through step by step. I've basically done all of the work for you. Everything is templated. All you're going to need to do is click a couple buttons to integrate it all together and you don't have to write a single line of code. So it's going to be super quick. I'm going to keep it really brief. But you guys can take this and sell it. It's being sold by people successfully in our community which I'm going to go into in a second. But you can take this and sell it to your own clients and start marketing it as well. So, great thing to get started if you're just getting into the AI solution space. And let's jump into it. If you are new to the channel and don't know who I am, my name is Liam Otley. I've been building and selling AI solutions like this through my own AI automation agency, Morningside AI, for the past 18 to 20 months now. I've been really screaming this from the rooftops that this is the way to get into the AI space for over a year and a half now. So, that everything I do on this channel is built around teaching you how to build and sell these solutions and build a business around it. So, in this case, I've done literally all of the hard work for you. I've sat here for a few hours today rebuilding the system using no code software so that you can just click a couple buttons, download the templates, and connect it all, which I'm just going to show you how to do in this video. So, there's no excuse now that uh I don't know how to build things. I don't know what to sell. I'm giving it to you right here. So, let's jump into

it. Okay. So, what exactly are we going to be building? Well, this is an AI market research assistant. And this was sold by uh Abby Aumu. I think I'm pronouncing that correctly, who is a member of our community. And a couple of weeks ago, I put out a form, which is this form here, um, where people could submit their solutions and let us know how successful they're being, what they're selling, and providing the proof as well. So, Abby was one of those who applied on that form, let us know about the solution says built. And I'll put her proof of payment up on screen here now. She said she got paid 75% upfront. So, this transaction is the upfront amount. I assume she's still in the building process for this now. So, Abby is the one who came up with this idea. So, if you want to get in touch with her, you can work with her agency right here, or click on this to be taken to the LinkedIn. And if you want to be featured in future videos like this and get your name put in front of thousands and thousands of people then you can use this form here. So you guys will get access to this Figma on the school in a second. And so what we're going to be building is an AI market research assistant which is an agent that takes in questions from us the user and then sends those to three different agents which represent different personas or different users of a certain product or service. So different businesses have different personas. Say we have the this type of user in my case myself regent. We have AI newbies. Some of you watching this might be completely new to building AI solutions. So that's one of the people that will use my software. We have developers who want to use Agent to streamline the development process. We also have business owners who have had maybe a developer work with them in the past and the developer is transferred onto their account and now the business owner is a user of Agent. So three different types of um people who or personas who use the software. And this system basically allows us to send questions to these different personas and ask what they think about things like feature requests or pricing. And so inside each of the prompts and knowledge bases, if you want, uh you'll see in a second, I've written these quite in-depth prompts on each of these to tell them who they are, what they use the software for, what they really want out of this, what they want out of life, um they are able to respond in line with how a as close enough as possible to a real person would reply. So, it allows for very rapid uh market research and instead of having to get send out or send out emails and ask for surveys and or like get people in a room and do little like test groups and ask them a bunch of questions, which is expensive, takes a lot of time, you're dealing with people. This is a far more rapid way of doing this and therefore can be very valuable to businesses if they're looking to get a soundboard on new features and new things like that. So, um walking through the system here, we're going to be creating three different agents on Agent My Software. Don't worry, it's completely free apart from your Open AI usage cost. So, we don't charge anything. It's just the OpenAI usage cost. So you're getting the lowest cost possible. So we're going to be creating three of these. I've already built them, already written the prompts. So you guys are just going to be able to snag the templates for them and be able to mimic those prompts when you build your own versions. So we have these three agents. And then we have a master agent, which is the one we're actually going to be talking to. And this is what we really hand over to the client as the product. Um, and we're going to be able to send messages to it. It's going to take those messages, send it to a relevant AI tool, which I've built a template for. You don't need to worry about. um that's going to manage the sending of a request to these different agents. And then when it all returns to the AI market research assistant, it's then going to aggregate it and make a summary of all the different responses from these different personas and what the AI newbie thought about the feature or the developer and the business owner. And then give us a response saying, "Hey, this is what each of them thought about it. Maybe you could do this, this is this. " So that's the system. This is the step-by-step process we're going to go through to build it. So all of the resources, there's quite a lot cuz I've put quite a lot of work into this for you all. It's all going to be in my free school community. It's going to be first link in the description. It'll need you to sign up to Agent, my free software. You'll be able to clone my agents, all of these that I built into your account in just a couple clicks. You'll then be able to sign up to relevance AI, which they do have a free plan as well. You'll be able to clone my relevance AI tool and connect it to the main agent, this one down here. And then you'll be able to test it done. So, there's barely any work to do here. It's really just connecting the different pieces that I've already built for you. So, let's give you a little demo of what

it looks like at the end. If I say I paste this in, I've got a little message. I have a question about a client feedback mode feature that we've been thinking about for Agentive. So, this is my Agentive market research assistant. It's giving me feedback on from the newbie, from the developer, and from the business owner. I can now ask and say, "Hey, what do you think of this feature? pricing model? " So, I'm going to send this and see what it says back. You can take a pause of the video and just read over what I've asked about there. It's really interesting feature. I would love to hear your thoughts on it cuz I think it could be very handy for AI automation agency owners because I know from running my own agency how difficult it can be to send chat bots over for client feedback and then they send looms back and then they deal with this anyway gets really messy. That's why we're thinking of doing this client feedback mode with an agentive. Okay, so we have three different perspectives here. The developer sees it as highly beneficial. They appreciate the dedicated feedback page for client feature is really aimed at developers. Um they have the business owner perspective. They are expressing reservations noting that the main concerns lie usability and performance rather than developer client dynamic. They prefer features that will empower them to self-manage and optimize their agents without relying on a developer. So business owners understandably more focused on getting maybe better costs or making it easier for them to self-manage. Not too worried about that developer client relationship. The newbies are pretty into it as well, stressing the need for user friendliness, especially for non-developers as they're typically not as technical. So you can see that we get three different perspectives and the prompting that I've done allows us to have quite like adversarial perspectives and what different people in different personas want out of the application. So this kind of solution can be super helpful for getting rapid feedback from your different personas that are end users of your application without spending thousands of dollars on focus groups or without having to go through the hassle of actually organizing them.

So enough talking, let's get into the build. So we have a couple different agents. We have firstly our central agent, the agentive market researcher system, which is the one that we're going to be sending messages to. Then we have our three different personas. The AI newbie persona, the business owner persona, and the developer persona. So in each of these, I have written a detailed prompt. And you guys can take this and apply it however you want to. You can throw it into claude or chat GPT and say, "Hey, take this same style and I want to do it for this. " So everything I've put in here about giving some information on the product, providing some details on what they want out of it, etc., and their background and some general context. It's all really good prompting that you can take and apply to whatever you want. So, the first step in you following along with this build is going to be going to the description, joining the school community, finding the YouTube resources. There's a little tab for YouTube resources. You'll find this video. You see the thumbnail on it. And inside that will be all of the links you need to clone all of these agents. So, go now and agents into your Agentive account. You will need to sign up if you haven't already. So, when you click the link, it'll take you to the homepage. You can sign up and register. It's going to ask you for your Open AI API key. You need to put that in because everything we do on Agentive is built on top of the Open AI assistance API. So, we're basically allowing you to build on top of that API without having to code and do a whole bunch of unnecessary work. So, that's what we've built the gender for just to make it really easy to interact with that API. So, once you've gone and created your account, added in your open AI API key, and then cloned each of those agents into your account, you'll have this, this, and this. And if you go back to our Figma here, you'll see these are the three agents, the AI newbie, Persona 1, Persona 2, Persona 3, and the AI market research assistant. Now, we just need to fill in the connecting part

which is this relevance AI tool. So if we go back to here, there'll be another link on the school community to get this template for an AI persona query. So what this takes in is the question that you want to send to the agent and then it also takes in the assistant ID. So on agentive each of your assistants has an assistant ID which can be found by going here to make sure you've already published the agent and then you can click deploy. You can go to custom code and then scroll down a bit and you'll see the assistant ID is available here. So each time we send a message to our executive agent here, it's going to be taking the question we provided and then the assistant IDs for each of these agents, which we're going to include in the prompt. I'll show you that in a second. And it's going to query each of these agents through our relevant AI tool. Cuz you can see here, it takes in the query and assistant ID. And if you go to the prompt on the market research assistant, if we go to edit here and you scroll down, you'll see that I have all of the keys in here. So that's the next step for you now is to come into your agent market research assistant. to the main one we're going to be sending these messages to. Go to each of your agents. We can start with the AI newbie persona. You can go deploy custom code and then copy this assistant ID. I'm going to go back here. Scroll down and then AI newbie. Replace this. Paste it in there. So, you need to do this and repeat the process with each of them. Make sure it's published. Click deploy. Go to custom code. Take the assistant ID. Take it back over to the market research assistant. And make sure you've pasted in all of your assistant IDs to replace my old ones. Once you've got the prompt all set up, ready to go, you'll be able to go onto the school community again and find the link to this exact relevance AI template. There'll be a button up in the top right that will say clone into your account. You can edit it. You'll be able to clone it onto your relevance AI account. And then you'll see a screen that looks a little bit like this. Now, the final step of getting this relevance tool set

up, I've done most of the hard work for you, is to open up each of these API blocks. And then there's going to be an API key here, and that's your agent API key. So, to get that, we can come back to any of our agents. um in this case just on here. And if we go to deploy custom code again, we can see that our API key for Agentive is here. Copy that. Head back to your relevant AI tool. And come to the API key and paste it in here. Make sure you delete that and replace that with yours. And the same for the next block down here, which is going to be this one here. Can delete that and paste it in. And just a little explainer if you're not sure what's going on here. When we send in the query or the question that we want to ask to this panel of researchers, we also send the assistant ID which our uh our ma our main agent is going to be sending for us from the prompt. It's going to be calling each of your agent assistants first. It's creating a new session. Then we get the session ID out of that. Then we pass it into our chat endpoint which is going to send off the final payload to actually generate the response. So don't want to get too stuck in the weeds there. But once you've cloned this template, you've gone through and added in your API key from this from the deploy tab here in the custom code section and you've added that to both the API blocks, you're then ready to give this a test. If you'd like, you can come back to any of your agents and we can grab the assistant ID from one of them and then we can go back to relevance. We can pass in a query um thoughts on a $100 plan with unlimited usage. Put that in there. And if we go run all up in the top right at the bottom, we should see the final thoughts of in this case the AI newbie agent. And there you go. You can see we're getting good responses from our tool. The last step here to integrate this AI persona query

tool with your main agent is to make sure you've saved it in the top right here. Then we can go back. We can click the new button up here. And then we'll see this custom actions GPT button pop up here. So this is why relevance AI. If any of you guys are just getting started in the space, highly recommend relevance AI is an excellent place to build tools because anything that you build in here and they've got a bunch of templates as well. This is not a sponsored video by the way. You can then click this custom actions GPT button which allows you to put it onto GPTs or in our case onto agentives so we can put it in different places as well. So to do this we can just scroll down and click on our AI persona query. You can select multiple which is really handy. They'll bundle it all up into one schema. Then we want to select the custom or here. Um, I'm just going to generate the schema first. We can click this. Now, we get our schema. This is needed to connect the relevance tool to the agent. So, we can copy this. We can come back to our agentive market research assistant, our main one we're going to be chatting to. And click on the tools tab. I already have a tool built in with this template, but you can just click here and delete it cuz you'll need to update it to your own. We can click upload tool here. Paste in that schema. It's going to add this AI persona query tool. And then to make sure we're authenticated, to actually call relevance AI and be approved to use it, we need to click the authentication button. We can click API key here, we're going to go custom, go back to relevance AI, scroll up, and generate our own API key here. Copy the API key. Come back, paste it in the API key section, and we're going to write authorization with an American spelling with the zed in it. Authorization. Then we can click update authentication, create tools. And now we have connected our relevance AI tool which can take any assistant ID into it and generate a response. And we've got all of the updated assistant IDs in here. If you take a moment to read through the prompt, I've basically instructed this agent to hit that same tool, the relevance thing that we've just set up. Hit that thing three different times using the assistant IDs in here. So every time you get a message from the user, make sure that you are taking one of these assistant IDs, taking the user query that we just sent to it, sending it to the tool, getting the response, and then doing that two more times with the other two tools in here. You could set it up on three different tools, but for simplicity sake, and for you guys watching this and following along, it's actually easier for you to just use one tool and ping the thing three times. So now that we've got all that set up

we can click publish in this case, and we can start sending it messages. So, um, what do you think about a $100 per month, um, analytics plan, which gives much deeper insights on, uh, your agent's usage when processing a lot of volume. So, now it's going to take that query, it's going to go to our relevant AI tool, which you know by now takes in a query and an assistant ID. And so what this is going to be doing is taking that query, taking these assistant ids that we've provided here, sending it off and getting a response from each of our different personas, bringing it back, and then the magic of plugging it all into one agent here on agentive is that the LLM behind the entire system is then going to take each of those responses and then based on our prompting here, rewrite that all into something that we want and that's much more processable. So instead of having these big blocks of text, it's going to write a nice concise summary. And here we go. Here we have our responses. I can probably zoom this up a little bit. Developers see the potential and benefits in the analytics plan. The business owner acknowledges that deeper insights could be valuable particularly for operations involving significant volume. However, they express concerns about the cost effectiveness of the plan and desire a clear demonstration of its potential return on investment. So, the business owner of course very worried. You can see in the prompts of each of those in the templates I've provided how I've prompted them into business owner being worried about cost and managing the thing themsel. The AI newbie is very interested in anything that can allow them to prototype and test, but they're not really interested in the latest stage features just yet. they're more selfish and want uh the user friendly, beginner friendly stuff to be able to build and prototype quickly. So you can see here the AI newbie is interested in the idea of the analytics plan but feels it may be premature for them at this stage. They recommend a gradual introduction to analytics features such as tiered pricing or a free tier to facilitate their learning without immediate costs. So this is just an example showing it on my own business. It's I built it cuz it's going to be helpful for me. I can now send things to this um and my team can use it as well. I can provide that to the dev team. And personally as a business owner using this for myself, this is extremely helpful. And I can only imagine if I was actually having to pay thousands and thousands of dollars a month to get this kind of insights from focus groups, then I'd be definitely willing to spend a few thousand on a system like this for internal use. And it's so much faster than having to use humans for it. Just quickly, I forgot in the original recording to show you how to get that final result that I showed you, which is a full page way to chat with your agent. So, you can come up to here, click on the share button, you can click the make public, then you can click this. I've copied it. Now I can open up a new tab and here I have my full page agent market research assistant where I can say hello and send messages here. So this link that you get um via this option here you can then send to your clients um for them to be able to chat with it and use it as they wish. You can also send the template via email and you can generate a public link to share your templates around. But this is the main one in order to get this full page chat widget. Just going back to the Figma here. I want to say again thank you to Abby for sharing this with us. If you want to be featured in more videos like this and some of the other list videos where I'm sharing AI solutions, you get a massive shout out on my channel, more traffic to your agencies, more clients, more people coming to work for you and in general, if you're looking to replicate this build and do it yourself. Um, this Figma board will be on the free school community as well. So, everything you need to do this will be on there. And you can see the step-by-step process here if you get lost at all. So, I know some of you might be saying, "Oh, $32,000. There's no way I could sell it for that. " That to some degree is a skill issue. Aby's found someone who's willing to pay that. And in your case, if someone's willing to pay $32,000, I bet you they're going to be willing to pay three or four. So even if you don't think this exact product that I've shown you here is worth $32,000, you might need to take another step further and take it from Aentive and put it onto a custom front end that seems like more white labelled and looks better for the client or connect it to say Slack or something there. You might need to add more value onto it. But there's definitely people in businesses out there, as Abby has proved, who are willing to pay for these things. And customer research and market research like this is a massive function for businesses, especially as they get larger. So take this and do with it what you will. I can't do much more than spoon feed this stuff to you without no code, no nothing, tutorial, everything templated. Please guys, I'm doing everything I can here to give you a leg up and get into this because it has changed my life and help me transition into a career where I love what I'm doing every day and working with people that I really love being around. So you guys know I don't want to hop on for too long, but this has been a tutorial showing you how to use my self-regentive to create a market research assistant for my own self-reentive. All of the links available in the description and there's going to be a tutorial up here showing you how to take exactly what you've just built here and put it onto WhatsApp using our no code WhatsApp integration on Agentive and just a couple clicks in a few minutes. So check that out if you want to add on to this and create more value for your clients by putting it on WhatsApp. But aside from that, that's been all for the video guys. Run out of daylight here.