Master Ai Agents: Build Multi Channel Customer Support Ai Agent Using N8N

1. Introduction & Overview of Class: Hi. Welcome to this amazing NTN master class in which you are going to learn AI automation by building no code AI agents in the NNTen platform. So my name is Shek Saiful and I have one year of experience as a prompt engineer, freelancer, and I am very excited to share how we can build AI agents without coding by using NTN platform in this class. So in this class, you will learn how to build multichannel customer support AI agents, including a telegram, email, form submission, and feedback collection all using simple nocOTol that is it. I will show you step by step how to create five different AI agents for real business scenarios so you get practical hands on experience in that. You will also discover different ways to build AI agents using AI models like Cloud by writing the prompts, create the AI agents for you, even if you don't have knowledge in the AI agent. And I will also discuss with you the different ways of monetize your AI agents building skill in which you can sell your services online. And I will show the different ways to monetize your new skill that is AI agent skill, whether as a freelancer, business owner or automation consultant, I will discuss all those things in this course. This course is beginner friendly, packed with project at workflow Temples, you can use immediately for real results, whether you want to automate your own business customer support or start your freelancing journey by building AI agents for your clients. By the end of this course, you have the ability to create AI agents for the customer support and other AI agents also. And we will master the AI automation using NATN and you will learn the different ways of creating AI agents using AI models and different ways to monitor zeer this skill easily. Let's get started building AI agents and mastering automation together, and we will jump into our first session and we will start from the basic that is what is an AI agent. Let's dive into that. 2. 1. What is an Ai Agent?: What is an AI agent is? An AI agent is an autonomous system that can look at what's happening in the automation. Okay? It will understand the task. According to our instructions, it will understand the task, it will make decision, and it will choose the actions very specifically to complete that particular task. So unlike we have seen the regular workflows like we have got inme.com. So there are a lot of platforms already we have seen their EI automation which connect the different apps together to automate the repettd task day by day, right? That are the workflows but in the AI agent, we will choose the actions by taking the decision according to our instructions and the given task. And not only that it will learn, adapt the situation, or it will adapt the task, and it will choose the actions according to our task and instructions to reach its goal without needing and constant main guidance. Once we write the full instructions to AI, and when we assign some actional data to the A agent, it will choose the actions according to the task, and it will make the decision, solve the particular task very effectively without need and constant humane guidance on it. Just once we need to write the instructions and we need to assign task rules to A agent, then it will automatically take the decision. It will choose the actions according to the task, and it will solve the user task or problem. Not only that, it will learn. And it will adapt that particular situation, and it will choose the actions to solve the particular task. By solving the problems again and again, it will learn Okay. And when we give the feedback to AI agent, it will automatically it will improve itself over time. That is simple. So you can see the example. It's like a digital assistant that works independently to help people with different jobs and challenges. There are more examples you can find in online for better understanding. You can find it more tutorials on that. Okay, for better understanding, let's see what is the difference between AI agent and EI Bflows. You can see difference between AI buflow and AI agent. You can see we have divided seven different aspects of features like structure. It is a predefined sequential task or steps. You can see it is a simple line diagram of EI workflow and I gen. When you see this here, it is a simple steps and logic predefined task. Dia overflow means we will just connect different apps and different APIs like that to gather the information, and it will just done the task from start to end. Okay? You can see here. You can see here. Such a predefined and predefined or sequential task or steps. We can see it is a simple predefined start and end to sequential. Okay, sequential steps. When you see in the AI agent for the structure, it is a dynamic planets own actions to reach the goal. So according to the task and our instructions, it will take it will choose the specific actions to complete the specific task by itself, by itself, and it will complete that particular task very effectively. Okay, according to our instructions, you can see. So in the A workflow, it is a predefined and fix logic which works from start to end to complete the task. But in the AI agent, it will choose the actions very dynamically, according to the task that we will assign to A agent, it will choose the actions according to a task and it will complete effectively. That is simple. You can see what is the autonomy. You can see Low follows fixed rules and pass and the A agent makes independent decisions of the intell prom. Wore than intell prom, it is instructions that will guide the AI agent to work in this context. I hope understand this point. It is a high because it will take the decisions independently to complete the specific task by itself, adaptability. For the adaptability, you can see for the AI workflow, you can see unlimited and handless expected variations needs a bit for new scenarios. So for every time you will try to connect the different apps, you need to update the workflow, right? You need to update the workflow in order to change the scenarios. Okay. But in the AI agent, it will highly adaptive handles new or unexpected situations. The AI agent will automatically handles new or unexpected situations. And in the decision making point of view, the AI workflow is embedded in the workflow logic step by step, as we have seen in this one. Get the decision making, which is, you can see emband workflow logic. It is simple workflow logic, not all other independent decisions taking like that. But in the Agenttegrate chooses actions in real time. You can see the task is anything, but according to our instructions and the tools and the data we will connect to Egen, it will choose the actions very smartly according to the task, and it will complete that task very effectively without any human intervention in that. Okay. In the real time. So we do not need to change every time for the different tasks. It will automatically take the decisions, and it will adapt that particular situation and it will complete that specific task by choosing specific actions that we required to complete that particular task. That is simple. And in the memory and learning, so in the AI but flow, it doesn't use a memory or learning between runs. It is simple. Pixel logic, right? It works from the start to end to achieve particular goal. There is no intervention of a memory and other things. It is simple logic base. I will work from the start to end to achieve a goal. There is no reasoning in this AI workflow. But in the EI agent you can see, can retain memory, learn and improve over time. So according to the working of AI agent, right, it will adapt the situation. It will learn, it will improve over time by solving the task. You can see here, adaption, according to the new situation, choosing the action to update the particular task, it will observe and it will take the feedback again and it will improve over time itself. Okay, I hope understand these points. And the transperse you can see it is easy to edit each step, right? So because it is a simple pixel logic, we can check and we can edit every step very easily by just seeing what happening from start to end in every step. It is an easy one, right? But in the A agent, it is very hard errors, issues, and logs. You can see full reasoning may be hard to trace. In that control, controlled externally via workflow platform. It can be anything make.com Enten. There are so many workflow or agent builders in the online. So it is Air workflow, which means it is controlled externally via workflow platform. So it is a simple thing. In the air agent, we can control is internal to the agent needs special guardiels for management when we create the AI agent, it also agent is developed in the int and you can host in the different cloud services. But when you control the AI agent, we need to have the special guardls for management, which is required to protect your private data. You need to use the special Gudils for management. So we have seen these two differences between AwfroE agent. Let's see the examples how the AI workflow and AI agent looks like. You can see it to a simple A workflow example. You can see to the simple logic base, logic base, and step by step workflow, which automate your task. It is simple overflow example. You can see. In the next one, you can see this agent example. So you can see it is a action particular task, we will ascend to the AI agent. We will write the instructions in this A agent, and we will tell to A agent use these particular tools according to the action sent to AI agent. No, it is an output. Okay? Now, when any action occurs here, according to the notes, it will take the input. It will use the different tools to achieve the specific goal and providing the output in this form. So the decision it will take according to our instructions, according to the task, it will take the actions, and it will just provide a output. That is simple. We will see all those things how its how to write the instructions, all those things in upcoming session. This is difference between IIbflow and AI agent. For more information, you can check in the online or YouTube tutorials for better understand. In the next session, we will start from the NA ten basics. Okay, let's dive into that. 3. 2. Getting Started with N8N Platform: Share, we are going to see how we can use NA ten, the basics of NA ten. So for that, just come to Google search and write the Nt N. You can see this is the first website you can go with that that is NNI ten. It is a AI automation platform, click here, and you can see this is a web page of NNITen software. You can see flexible A workflow automation for technical teams. You can see some different EI agent templates, IT Ops. You can see DevOps, sales, you can see this is a security purpose. This all the AI workflows, you can build on this platform by just dragon drop and connecting the nodes easily. Okay, there is no limitation on that. You can do so much things, you can connect the different apps, all those things in this NA ten platform easily without any coding. You can see some reviews, all the things you can explore by yourself. If you are new to NA ten, you can just click on the guest Chatter for free, or you can go the tases. I have already created the account, so I will just go with the sign in button. You can see this is my simple name. I'll just click on the sign in option. Now it directly takes me to the interface of my but flows. This is the simple my list of workflows. So if you are created the new account, you will instrp this workflows, you will get the two tabs. One is to create the workflows and another one is to create to AI agent. So they are not different, but they are the same. You just go with creating anything, Workflow, AI agent, it will take you to the same dior. There is no problem with that, right? And you can see this is simple ten, this is the ten homepage. You can see you can just click Here plus button. You can create the workflow as well for the personal M project or cringe and you can use. Okay, you can do things. What the personal this is simple, you personal projects, credentials, executions, you can see how we can create the but flow directly from here, all those things. This is a MI project. You can start from the scratch, you can click here to create the project as well. You can see this is the admit panel. This is simple admin panel. We can just create it, right? So for more about this one, you can follow the simple tutorials on YouTube for more information about this ten, right? This is the variables, all the things. You can create the variable. So it will use to store your data across multiple workflows. It can be anything. If you click here, you can write anything that is simple, you want name. Okay. And a well, you can give your name so on. Like there, you can just save it here. You can use this particular variable in any workflow. You can do these things also. You can see the insights, how much you overflows, which workflows are working Dilly, what are the execution, all those things you can get this one failure production execution, failure rate, time, save run time. You can get all this data once you make your workflow active. Okay. Can see this simple documentation can find u star documentation form course, you can get the course from it about NNN. You can see this is simple What's new. So you can get directly some update recording N ten. You can see g and to node. You can get the directly updates in this particular what's new in the tab section. You will get all those things from here to make sure you get updated through this technology. Okay, I have to understand this points. Come to your home overview. So these are some my workflows that I have created earlier. These are the credentials. So the credentials means we need to we are here to automate our workflows, to create AAgens to use my other apps, all those things like Gmail, Google Sheet. So we need to connect that particular tool to this particular NATN. So for that we need and credentials of these particular apps. But by using that particular credentials, we can connect NATen with other apps like Google Sheet Gmail, open any account, open rooter, anything you are looking to automate, you need to get that credential, save these credentials in this entertained platform in order to connect all those things to automate our workflow or to save our time. Okay, I understand these points. These are the executions. So what is your workflow tally executions you can track from here? Where's the time? Okay, where is all those things from? Okay. From here, you can just click on the Plus button from here or you can get from create workflow or you can create the create connection as well. Okay. Let's go to our create workflow editor and we'll learn some basics. This is a simple workflow editor. At where we are going to connect the different apps to automate our Dali task. Okay. You can just click here to add the first tab. You will get another section, sign par, you can add anything. This is a triggers app event. I will explain each and everything. The simple how we are using to connect all those things in a few seconds. Before that, we need to learn the interface. As I said, you can learn this. So what is edited, this is the editor tab and this is the execution tab. So when we execute this particular workflow, you can see the executions, how much executions have done. This is the evaluation. You can learn how to evaluate our overflow by using multiple inputs. You can see add evaluation to the right workflow outputs back to data cell. This is simple tutorial. You can see in order to evaluate your workflow, it is working in the right manner or not with different multiple units. That is simple editor executions. You can come here. You can just click here. You can easily add the different apps and connect all those things from here. Otherwise, you can just click here. Otherwise, you can just click on the tab option in the keyboard, you will get these things. I hope you understand these points. You can add the sticky note here, the sticky note, you can just double tap here. You can write anything you want. Like hi. This is for tech. For example, as take in. I will just help you to write the different nodes, all the things from here as well. You can use the AI ask assistant. If we got any issue in this workflow, the AI assistant will guide us and we will help you to resolve that particular issue very smartly. You can use this one as well. Okay? Now, you can see you can save the particular workflow as well. We can share it with your colleagues or users that we will add here, right? And not only that, you can activate activate your workflow. All these things you can do this. You can change your workflow name. You can tag here for a better understanding that is customer support or Achatbod like that, right? In order to manage our buflows very easily to save a lot of time on searching for it. You can count here three tots. You can duplicate this particular workflow. You can download the JCNFle that you can share with different acadics or students or friends to try your buflow as well. You can rename it and you can import from URL. You can take the NA ten templates here. Let's say NATen templates. We can discover there so many workflow templates here. Let's take this one. Learn JCN basics. When you click your US for free, you will get this one. Import template when you have the account and directly in this particular chrome that have created the account in Int and it will directly tell you Import template to AI Prompte. I will show you a user name, you can just copy the template to clipboard JSN. You can do all those things by itself. When you click on the input Template, it will directly import to this R workflow editor that you can see here. No you can see. This is a whole workflow template that we will get from this particular one. This is how you can use these things as well. You can live without seeing. We have different ways we can use the workflow deter, click here. We have different one. Even if you just download the JCNFle you can input from file option. We'll get all the JCNFlow you can see. You can add the JCNFle can see here. This is a simple Zoom to fit option. How it works, I will just show by simple adding the different things. Okay. Let's take any at one. Okay. So this is simple thing. If you see here, click here. These are the two nodes, which have the in different positions. So when I click Zoom to fit, it will directly show the correct one. When I just Zoom in Joomod you can Zoom Zoom out easily from the tabs here. Otherwise, you can click, you can enter press the Control button on your keyboard and with the mouse, you can just scroll up or scroll down. You can use anything like that. You can come to here, just click plus button. Just come here. You can directly see this one. F. You can directly change from here by scrolling air up. I help understand this one. You can do for better experience, you need to try self to get how it works or not. Okay. Now, so this is a Zoom mode or Jo Min. This is a simple tidy up like that. So this is a tied, what happens at tidr? When you click here, it will just the Nitin will automatically put your workflow sequence in the correct manner. As I said, this workflow is in this position, this workflow is in this side. When I click a tidy up, it will just take to the correct sequence of this particular. This is simple basics that you can learn. You need to learn in order to create the workflows in efficient manner. Now, I will explain what is a trigger. Always remember we have we need to learn some basic trigger. So what isn't trigger. Trigger is simple. When this is a first step or first node, like, Okay, when it will run, the rest of other workflow will start working. That is simple. So when clicking isecut workflow, when I click here, Run button, it will just run it because it is simple when clicking executive workflow button. It will work like that. You can directly come here, you can deactivate it. Okay, you can delete it, or otherwise, you can click here, you can execute, activate from here, copy that. You can do all those things. When you click here again, it will just activate. Okay. You can run this one when you double click on. You can see if you have any doubt regarding this one, you can see the parameter settings. You can just click her the dogs. It will help you to learn more about this particular node. Okay, these are simple basic things that every node, if you don't have any knowledge about this one, for example, let's take an AI agen. You can see the different AI templates recommended. You can just go and you can see different AI agents, templates you can directly input from here and you can learn how to make it very easily in order to save a lot of time to creating AI agents. Okay, you can learn so much things. For example, we can take AI agent node. If you don't know what is this one nodes, what is this node? You can just click Double click here. You can go the Docs option in order to learn more about this node. I hope you understand these points very well. You can just click here, Control plus A. I will selected. You can delete the option from the keyboard as well, or you can just right click it here and I can tell you delete three nodes that is simple. It will automatically delete all those things. This is a basis one, so you can try yourself. Not only that in an empty workflow work flow. If you just click on the right click, you will see you can easily add the node or sticky node or tied up for flow. By easily instead of seeing all those things by itself. Okay. I hope to understand these basics very well. In the next session, we will create simple AI gen, for basic task. Let's dump into that. 4. 3.1 Building Basic Ai Chat Agent: In this session, we are going to discuss to create the first AH, we are creating CharBT AIH. Let's see this. You can come here to add the first step or the node. You can click from here or you can go with here. So I'll just click in the add first step. When I click care, the sidebar will open, you can use the trigger manually or all these things. The trigger means adding the first step. So when the trigger will run, the rest of all other steps, the all overflow will run. Okay. So that is simple. That are the triggers. The trigger, it can be a web hook. Okay, it can be a simple form submission or sending an email, Okay, receiving an email or some chat received from other channels as well. So in this, we are going to create an AI agent first. Okay. So for that come here, just write AI. Okay, you can see this are the sum A, you can find this AI agent option here. Just click on AI hen. No, this is the simple chen node comes with the chat receive. Sometimes you can get only the Ichent node as well. But just come here, tell chat trigger. Just try hat trigger. You will get the chat trigger option here. That is the same. Come here. I can add the directly from here. This click, when you click here, it will show the other nodes to connect it. You can directly come from here or you can directly just drag and drop this one. That is simple. Okay, you can use this chat as a trigger. So now you can directly come here open chat. But open chat, you can send any message you want. For example, I will write high. This please enter. Now we can see error. A chat model subnode must be connected and edible. Now you can see. So when you get any error, you can use the ask assistant chat. Okay, directly come here, just click ask assistant. When you get any error, take help from the AI assistant in order to solve our issue. Now you can see. This is simple solution for our error here. The error, a chat model sub node must be connected and enable. Okay, you can see it means that your Agen node is missing a required chat model, subnode. Okay, anything like that. So you can get the instructions how to resolve all those things. So I will explain what is this one. So this AI agent is simple AI gen, but there is no brain in it. So we need to attach a intelligent brain in order to do our task. So for that, we need to add the chat model. Okay. Just click C plus button, it will automatically show you the available list of different chat model. You can see anthropic chat model Azure AWS, tap C, Google Gemini chat model, Group stroll this all or the chat model that our brain, which we are going to provide to AI agent to done the task, to take the decisions, all those things. Okay? So for that, you can add any chat model from here. So in my case, I am using open air chat model, just click here. Okay. Now, I have already different chat models as well, but I'm using the Acht Open AI chat model. So when you are using this chat model for the first time, that means you don't have a credentials, right? So to create new credentials, click on the Create New credentials button, you will come here. So you need to add your open AI API key. That is simple. So if you don't know how to get the open A APN K, just come here. Platform. API. You can directly go with the open A APAKes or all those things by yourself. Okay? So already I have my open A Aky. I will just already paste it here. Okay. Otherwise, if you have any doubt on getting the APQ and just click on Open Dogs, you will get how to get the open AIE credentials by simple following this step. Just go to open account, create an IPAkey. Okay, you can follow the open AIP documentation. It is very easy to get the open API. Or you have the different tutorials on the YouTube to add the open A APK in the NTN credentials. It is very easy. You can done by yourself very, right? So I have already an open A account. I have already just connected my open A APK to the NTN, so now it has done very perfectly. Okay. Now, let's see whether it will works or not. So when I just connect my opene chat model to the AI agen, when I send hi message, just enter. Now you can see we have got our response from the AI. That is hello, how can I assist you today? So when you open the open a chat model, you can change R models. Okay. I'll just go with that 3.5 turbo which is best for experimenting whether our workflow is working right or not. Okay? That is simple. Now, major problem with this is when I tell, for example, my name is My name is Si. Okay. So when I tell to AI module, my name is SIP. Okay, it will recognize Hello SIP. How can I assist you today? That is great. So right now, the AI agent is not working chargibty. Okay, so for example, when you use the HGVT it will recognize our name. When we just provide our name to CHGVT, it will remember our name up to our chat goes on in that particular context window. Okay. So in this Al the drawback is, so it will never recognize my name again when I ask to chat this AI model, what is my name? Let's try it out. So when I tell to AI what is my name, you can see, I'm sorry, I don't have access to your personal information, so I don't know your name. Even though I have just told to AI, my name is safe. Okay. And the next chat, I will just ask you to AI, what is my name? It is not recognized my name because the previous data or chat is not stored by the AI agent in order to recognize my previous data. So for that, we need to add memory to store our previous chat or data to recognize in upcoming chat. For example, you can just click your memory here. The New will automatically show the available memory, you can add it. You can order Mongo to chat memory, motor heead, postgres chat, memory, all those things by yourself. In this case, we'll just go with the simple memory that is stores in NNTNd memory. Just click here. I will automatically create it. This is the memory. It will store my previous chat or data to get the answer per me. Okay, Let's Now I will tell my name is. Si. What happens here? It will just store my name as well. Si. Hello, Si how can I assist you today? Now what I tell to AI? Let's take what is AI. Let just give the answer what is A. You can see I stands for artificial intelligence, right? It refers to the simulation of all those things you can get the data. Now, I will ask to AI W is my name? It will remember my name and it will say, your name is sip. No, it is recognizing my name very easily. Because we have just connected our memory to recognize the previous chat, all those things here. Okay. Now what we do? We'll just look inside of this particular notes how it is working. Okay, come here, just hide a chat. Now, just click on the web chat received. Okay. Now, when chat received, you can sit test chat button. This is a session ID. For every chat, it will create the session ID and the action is sending message. You can see this is a chat input. This is our chat message. What is my name? You can see we have different three ways to know this data. You can go with that table, you can go with the aschemaO you can see that Jason output. It is simple. Okay. You can make this chat publicly available here. Just click here. Now, you can see you have the two different modes, hoster chat or embedded chat. Poster chat means the chat is hosted by the Netin itself, or you can embed this chat in your websites, all those things, buy yourself easily in order to save a lot of time in creating the chat section. Focus on simple first creating your Pflows. Okay. Now, when I run this one, so we need to stop this for flow, open H hat. Now we'll come to the AI agent. Now, when I have just messaged in the chat input, what is my name? It's the session ID from the chat the three output come from in one item. This is the whole output. Okay, come in one item. It is a pack of this whole data. You can see in the table format, all those things. This is a one item. Okay, it contains this particular data that is session ID action chat input. It comes under the one item. Okay? Now, it will just go to the AI agent. You can see when chat message received. Okay, this is a one item. This one item contains the package of this output. Okay, you can see this re comes from the chat node. No, this is session ID. Now, this is action chat input here. No, this is a AI agent, right? This AI agent is using our chat model, this is. Right to done the task. Now, this AI agent, you can come here. You can check the settings, all those exhibit ones, all those things. You can go the docs option to learn more about this AI agent. You can exhibit step here to see whether it is working or not. Now, remember one thing. You can see the prompt user message. Now what happens here, you can click here. You can define below or you can just tell to AI agent comes the basic default prompt user message with chat input session, this one. I will show in otherwise, you can go with a defined below. You can add the session d here, all those things. But we are using this chat input as our question to AI model to get the response from is just come here. I will automatically recognize our chat that is what is my name here. You can see here. What is my name? You can define below or you can just tell to connect a chat gerne, it will automatically recognize this one. Okay, otherwise, you can come to define below in order to customize our prompt Baself. Okay, come here. You can drag and drop here. It will automatically recognize our message. Okay. Otherwise, you can come here table. You can just click here. You can take this one as well also. Okay, you can come here JCNFle as well. Just click here and rock and drop with the works likes that only. Otherwise you can come here, add option, that is system message. You can add the system mese you can see, you are a helpful assistant. This is a we have asked in the chat interface. Now this is that you are a helpful assistant. Now, the A agent works based upon our instructions and the chat we have given in the chat interface. That is simple. You can execute step directly here, you can see the output. We have got some that is error. We will see all those things here. You can see model. Now, Okay, before running this, we will see so much things here. So mapping from here. You can see these are chat received all those things. This is a memory. Okay, I will open chat again. I will just write high message. It will run, you can see. The AI agent simply just given the output Hello. How can assist you today? You can come open a chat model. So this is a message. You can see a system message you are helpful assistant. Who made my y is This is a previous data. These are the tokens, how much you are using. This is the output. You can see her complete tokens. You can see this is a response from open I chat model. You can see hello icon you assisted today. We can see in the different models. It will just go the I agent. It will give the answer in our chat model. That is simple. In the simple memory, it will just store our all session ID chat inputs or output. You can see this input. Just messages you are a helpful assistant. You can see the output here. Hi, you are a helpful assistant. This is a chat history. Okay, all the things will be stored in this memory. Okay, you can directly connect to chat trigger node. Otherwise, you can go with the defined will option. Okay, what is the meaning of context window length? As I said, up to five iterations, it will recognize our previous chats. Above five, that how much you have just fix it the number here. You can go with 500, you can go with 1,000 that is up to you, right? So for example, if I take five up to five times I message the chat model, the AI agent will recognize my name. After five iterations, the AI agent will not recognize my name. Okay, for that we are using the memory option to store our previous chat or data. So for that you can place here, you can tausend iterations, all those things up to you, you can use, you can test out to make our A agent works perfectly. Okay, you can run here, you can hide a chat, all those things from here. This is how you can create simple EI agent, simple chat GPT model by chat receiving all those things itself easily. To know about how it is working or not, just come here activate. Got it. Just click on the Hi. No it's working perfectly. You can see hell it seems like you said hi twice. Is there something specific you would like that? What is working in right now. Now this is how we can create the simple AI agent which respond our queries. It is simple. It is working like an AI agent, which receives any quotien and it respond depending upon the AI agent knowledge base, all those things. Now in the next section, we are going to see how we can connect GML, to send an email for every time that AI agent respond to user. Let's dive in. 5. 3.2 Customer Support Chat Ai Agent with Gmail: Previous session, we have just learned how to create a simple AI agents sends or which respond to our query for any question. In this session, we are going to see how we can connect GML, Okay, to send an email for every time that AI agent respond to user. For example, so when I just open the chat, I'll just to AI hi. Now it will just start working. You can see. I got an response from AI agent. That is, hello, if you have any question or need assistance, feel free to ask. So I am looking, so now it's working well, but I am looking. So for every message I have sent to AI, the AI agent is responding very well, right? So for every responding to me, so I'm looking to get the email for every respond of AI get. Okay, for example, if I just tell to AII, it will just respond with the message here. So for every time I message to AI, I should get the email off with the subject of my message and the response of AI agent in my mail. I hope I understand these minds. For better, we will see how we can implement the practical for now. So this is the input section. AI agent will use the brain memory. So even if you can add anything do, so we will talk in upcoming sessions about these things. But just understand the basic structure. Okay, this is the input or this is the output. Okay? We can do anything. We can add any other step, any node to automation task. So know what I'm just looking to connect with Gmail account. But every time I send a message to AI agent. So I need to get an email with the AI agent response and my subject. Okay, my message, all those things. Okay, for that I'm looking to connect with my Gmail account, come here and just search for Gmail. Okay. You can get this Gmail directly here, just click here, and you can see there are different actions you can take label actions, trapped actions that action triggers outswell. Okay, now I'm looking to send an email for that. You can see send a message, just click here. Okay, now, if you are new to connect with a GML, so you need to create a new credential. Just click here. Okay. So you can directly sign in with Google. Okay. Direct Lum here. Just click to connect with, and that is simple. Now, remember one thing, it is for me because I already created the credentials for my app. So if you are new to one, so you need to get the secret key and client teddy from Cloud counsel by creating the project. Okay. So if you don't know how to do that, you can follow the docs as well. There are a lot of more tutorials in YouTube. You can find out you can just easily connect your Gmail or Google Docs, all services to the NAD very easily. I have just created and send a message. So just click here. Again, you need to mention your email ID that you are looking to get the email from the AI agent, with the AI response and your message. So for that, I'll just write my email for checking purpose for test purpose. Info dot A Cgs. So this is my Gmail account. I just mentioned here. What is the subject for subject? I'm looking to get my chat input here. Okay, just click here. Dragon drop here. So for every time I message to AI, it will send an email with my chat with my message. Okay, that is high in previous case, right? Now, for the AI agent output, but the message, you can go with a HTML or text format. For the message purpose, I will just take the output from AI agent. You can dragonrop here. So this is the output. Okay. You can see here it is already showing Hello if you have any questions or need an assistant field for you to ask. This simple one. When you click her Execute step, it will show the level that is sent. Okay? For checking purpose, come to your Gmail account. Now you can see it has now you can see this is a message I have given as a subject, and this is a response from EI agent. You can see this is one. Okay, so for checking for test purpose, come here, whether that is walking or not, come again, open a chair. Okay, now, tell, ask any question that is what is an AI? So this is a message I just Q to AI. Now it is working. Now it has done. The answer is, you can see. What is an AI? Now you can see, this is a ID. Okay, thread ID. That means the output is email. Instead of getting the chat, answer, I have got the answer through email. Okay? If you remove this email, it will work in the chat. This is the basic foundational of NAT. Vu don't kept anything in the output format, it will get the answers in the chat as well. Now let's see whether the email has come or not. No, you can see this is the email that you can see what is an AI? This is the output of AI agent. An AI artificial das refers to the simulation of main. This is the response from AI agent I have got from the AI H this is how you can do so much things. You can automate so much things by yourself. You can connect the Google Sheets to create a new sheet. Okay, you can create a Google Docs that will save you all the agent responses in the document Google Sheets wherever you want, there are a lot more things you can connect. Okay, you can do so much things automation in the editor as well. Okay? Now, if I remove this one, just delete one. When I give the directly question here, what is an AI. No, it will directly give the answer in the chat only. You can see here. In the previous one, the output as given to the GM. Through Gmail, I have gotten response from the AIHen. But when I remove this one, the chat, the output of the chat is here only. So this is the inbt function of the chat chat node. Okay? So this is simple how we have just created all the AI agent for our different Rs, even though you can add so much things. In the next session, we will see how we can connect the tool and what is the process. Two. 6. 3.3 Adding knowledge base to Chat Ai Agent: So now in this session, we will see how we can add this simple specific application document. For example, I will just delete it. When we open a chat here, let's say, I can ask any question, right? So it will give the response because let's say what is AI. Okay, it is the tech basic answer, right? So A stands for artificial intelligence. Not only that, I will ask another question related to the healthcare. It can be anything. Okay. Let's take can you tell me about Human body. So it is like a biological question. So it will do the answers for any type of query, right? So you can see. You can see the answer here. Certainly, the human body is a complex integrated biological so. So it is not a specific one, right, it is working like an chad GIP. So what is the use of this? Okay? So we have already had GPT. So why we are creating this A agent for all of queries. But we are here to create AI agent for the specific application. To implement the AI agents in my business to handle the customer queries or anything. Okay, right? So as an example, we are taking customer support agent. Okay? Now, this AI agent is just know how to respond for the user queries. Okay, we have seen in this one. What is my requirement is? I am looking to implement this AI agent for my E Commerce website or services website, to solve my customer queries. Okay. Then this AI agent should know about my company, what is my company name, details? What are the products that I am looking to sell through my Ecommerce store or website? Okay, now, I'm looking to implement this AI agent to work on behalf of my instructions, and according to my company details, products FA o, frequently asked questions. AI agent should work behalf of that particular data. Okay, in which this AI agent can help me to solve customer queries instead of human agent. Now, when I add my company details, product details to EIGN, as in knowledge base, the AI agent will works with the provided data only, not for the all queries like chatbot or like HAGPDCloud. Okay, now, I will add my knowledge base through tool here. Okay? Tool means you can use the tool to done the task. So it is best for get the information from the different APIs, to change the particular data in Google docs or things you can do it by just come here, click on the press button, you will get all the tools available in here. Even if you don't find the tool you can use the HTTP request tool in order to gather the information about your company details. So for best and easy integration, okay, I am using the HTTP request. So to g the data to particular this AI agent, we need to use some vector database. We need to collect, we need to take that particular data from that particular document, knowledge based document, and it should be retrieved to the user is. Based upon the user quotien, it should retrieve that data from the document, call company document company data in which it will give the perfect answer for the user query. Okay, to implement that, so I am using the pin phone. Okay? It is a vector database for scale in production. So you can see it is a vector database. So you can learn more about this pine cone in the YouTube all itself. It is easy. Okay? It will just take something it is something called it vector base. Okay, I will just take the information that we want. It will retrieve the information from the document. Retrieve the answer from the data to the user. Okay? That is simple. You can come, you can just go with the login or sign up, it is free. You can come to the sign up button, just sign up. So I already have an account so I will just go to the login. After login, you can see it is a home interface. Okay? After creating the account, you will get the default ABI key. You can just copy and paste in Notepad as well. Okay. Now, welcome here left side corner, that is assistant, clicker assistant here. So I already created the support agent assistant here. Okay, I will show this one. I will delete and I will create the again. Just I will write things again here to delete. Now my assistance is deleted. I will just create a new assistant. Come here assistant tab, you can just here, create an assistant, here, and name you an assistant. I will take the Support Agent. So you can take any name you want, just click on the create Assistant. Now the assistant has created. After creating, you need to add your company data file. It can be the JCNFlePDF docs, anything you want. You can see this accepted files, PDF document, text JCNFm. I already have my document that is PDF format. I'll just upload input. Now, it is processing our PDF file. No it is updated. To check whether the assistant is working or not perfectly, just the ask a question about company details. Let's take what is your company name? So the AI assistant will retrieve the answer from this particular company data PDF. No, you can see it is. The company name is ACMAHme Appliances private limited. So this is simple. You can check here the corrected details. Okay, it is a document preview. We can see our company name is ACMAHme Appliances Private Limited. Okay, this is a simple I have generated using the AI for better explanation, you can do it by yourself as, right? This is simple my products catalog, frequently asked questions. Okay. All those things I just it is retraining the answers perfect one for my question. Now, this AI instead is working in the pinecd. So I am looking to implement this assistant. Okay, this process in the tive. So how can I do that? So for that, you need to come you need to use this HTTP request tool, click here and we'll get the request to, right? As I said, double click here. Now, this is simple, make HTTP request that is simple. Now, we need to import cur layer. Okay. For that, go to Pine cone. Okay. Come here. Just click on the get started. Okay. Now, what do you need to do? Go APK, create a new ABI key here. So I will just take support AI. Support agent. Okay. So remember one thing you need to use the create a key. So I have created the key, just copy here. Now, come. I will take the noteboard and I will just paste Okay. Now, what I can say, I have just taken the API. Now go to Gtarter or come to the dogs. Come here dogs. Now, come here. Pine Cone dogs, click on the new tab. You will just change this Pinece database to the pinecone assistant. Come to the standard interface. Okay, come down. You will get this Python JavaScript curl option here. We need to import this curl into the STT request here. Click Import curl option here. Come to the pine code, copy this one from curl option, this curl, time to here. Just copy this one and come and paste the curl here. Just click on the Import button. Now it is automatically import your all things that is easy. Now what you need to do, come here. We can see URN. We need to update our assistant name in Pine code here. You can go check Pine Code assistant name is. Let's go here assistant. This is our AI support agent. We need to replace this chat assistant. You can see from here with our AI support agent name. I can see. We need to change this one. AI support Agent. So we need to replace with our agent, assistant name in the pine coot the next one is EPT. We need to replace our API key. I have already taken that EPT. I will just paste here. Sorry. We need to go to this one. So I will copy that ABG and place it here. Now, it is done. No, afterward, you need to see here. So in the JSON body section, you need to replace this quotien with our quotient with the chat input we have seen in the previous one. So far that what to do, you come here. When chat received, right, just take this chat input. Remove this one. Okay. Just take this chat input, come here, this past here. So it is taken here. Just copy this one again. Pace in between here, content. That is simple. Now, it is replaced with our chat input instead of there. Now, when I click here, with. Now, this HGTBRquest tool will work. We'll retrieve the data from my document. Let's see whether it is working or not. Come here. Remember, we need to add the instructions in here. You can see you are helpful assistant. For my Company. Next we'll take use TDP tool. Otherwise, we can rename our tool for better understanding, right? So for then we are using name as the stake, company details or company doc. So it is a company dog that is simple. It is a tool, our name. So we need to write our instructions as you use company, Doc tool to retrieve or to solve user quotienqery user query based on company details. Okay. This is a simple I have just told you AI. You are a helpful assistant for my company. Okay? Use company dog tool to solve user query investor on company details. This is simple instructions. Okay, the AI agent will automatically use this company dogs API. HTTP, we have called with the Pine Poe. Okay, it will retrieve the answer, and it will just u to the user. Let's check whether it is working or not. So for that, I will just click on the open share. I'll just create the previous session. I'll just start with what is your company name? No, you can see it will give our company name as APMA let's. Let's say it is working. No, you can see the company name is APMA home appliances Private Limited. So we have successfully retrieved the data from our company. Okay? Now, the AI agent will not work like the simple chat GPD, okay, like other chatbards. But now our AI agent works with based upon my company details, instructions, all those things. So to make these things, I will just tell to AI agent Company details. Don't do other tasks. Don't do other tasks. Now, let's see. When I ask any question out of my company details like explain like explain AI in few sentence. You can see here. Now, if you see here, explain AI in few sentence, it has generated the output. Artificial intelligence, all those things. But even though we have adding the company details, we need to clearly instruct the AI not to do other tasks. Don't do other task rather than, rather than provided provided data company in company do. Let's see whether it is box or not. Again, I will clear the session. Now we'll try again. What is AR? Now you can see here. Now, when I just told just instruct AI, don't do other task rather than provided data in company doc. AI When I just ask what explain AI in few sentence, it has generated even though I just added my company details to the AI agent. But when I just instruct the AI agent, don't do other tasks rather than provided data in company dog, it is a company dog. No, it will just refusing to give the answer, you can see the company, let's take this one. You can see the company documents do not contain information about what AI is. If you need information especially from the company documentation, please let me know if you want to search for data topics or within the specific context in a company materials. Now the AI agent is simply refusing to give the simple answer out of my company dog. Okay. That is how you can use you can create the A agent for your specific application to create for your support assistant, customer support assistant, website support assistant, whatever it may be, you can create this particular A agent by simply just giving that. Okay? I hope you understand these points very well. Just we will see another example. So can you give the details of your products? Let's see whether it pulls the product details information or not from the company top. You can see how is running now. Like no, you can see it is taking my product details from my company dog and has simply generated the product details. You can see ACM Smart Cool, Acme ICO Wash, PureTaste ACM HartQuik. It is on some products that have just added in the document to check whether it is working or not. So simply it is now working like support agent. Okay? You can ask any question regarding your company document or regarding your company details, product details, I will give you the answer based upon your company dog or based upon your details that you provide to AI agent. Okay, it will retrieve the information from that particular company docs, and it will analyze and it will give the answer to the user very effectively. Okay? Now, we have seen only one chat that is we can implement this whole chat directly in our website. Okay. So whenever we can use this one. If sometimes sometimes our complete details, compare dogs doesn't have the customer questions answer, for example, if don't we can use another method, that is, we can add simple email Email, which will send the AI agent. The AI agent will send an email to the particular support team. Okay, human agent team, when if the AA agent got the quotien from user, which is out of your company docs, it will simply send an email to our support team, Okay, through Gmail, in which you can just solve the customer quotients by human agent instead of A agent. Okay, we will implement in the next session. Let's dive into that. 7. 3.4 Query Escalation to Human Agents using Gmail: Session, we are going to see how we can escalate the user queries to human agents or human customer support. Okay. So when the EI agent doesn't find the required information from company dogs to solve the customer query. Okay? So to implement that, we just come to your tool, click on the Plus button and search for that GML click. We can see here. Okay, come here, click on the GML tool. Okay, Info. A proms. Okay, this is my simple human AI agent mail, right? So in which it will send the agent will send mail to this email, okay, to take the query by humans and to solve the customer query. That is simple. Okay, for the subject, I will just show right now. So now, you need to do that two things. Okay? We need to come here, okay? We need to rename this particular tool with the simple send mail. After that, come to the AI agent, just update the system message. So I will just show No, I can see. It is a simple instruction that I have just written here. You can see send mail. So what is the send mail? So I'm just telling to A agent to use this particular tool that is send mail. Okay. And what is the instruction? You can see it is instruction here. Send Email. When you don't find answer in company doc and send email. We send mail. When you don't find answer in company dog, send mail to team with user name and email ID of a person who is required. So it is a simple instruction that has just guided the AI model, AI agent to use this particular tool when you don't find any answer for the user, for the customer quotien, just send a mail to the human agent, okay? To that support team with their a query is, what is the question with the name and email ID of a person who writes a query. You understand these minds. Let's see whether it will works or not. Okay. Now I will start. Before that, we need to come here. We need to change this subject, all those things. Okay. Come here, you can go with the hemo text, I'll go with the text and just click here, let the model defines the parameter. Now for the subject, you need to use you can write by yourself anything like Nik support team. Okay, you can do anything you want. So which best suits for you requirements, you can add and number of emails here. So I'll just add one. Okay now, what we can do. Execute the step before we need to start from the scratch. Let's open the chat here. No, I will just ask AI out of my company dots. For example, so I doesn't mention in my company do about the offers. Okay, now, I will ask a AI agent about my offers. Let's example, can you have 50% all products. It is a simple question that I've asked to AI agent. Now, the AI agent will give the answer, let's give this one. No, you can see it is working right now. It is out of my company dogs. So there is no details about the discounts in my company dogs. Let's see the Agent. You can see here. There is no information in the company documents about 50% off or on all products. But details about promotions or discount, you may contact, so you can see here. So we need to go. So I need to contact to this particular support team with the mail or through their live chat on the website. So what happens here? I can tell to A, please, can you send the message? Can you send this message? No, I will ask, I have sent message to Team regarding a query about the 50% discourse and all products. If we have any more question, need for this feel free to ask. No, you can see t has generated. Now we will check our mail whether it will send it the message or not. No, you can see the mail has come. Please needed support team. If you see this is a query. Now if you see customer name not provided, customer even not provided. So it has simply just given a customer name not provided, customer even not provided, but it has given the question about provide the information regarding possibility of offering 50% discounts on all products. It is a customer query. You can see it is a customer query, right? But the customer name and customer email is not provided. So for that, we need to change our stat instructions. You can see. So come here, when you don't find answer in company dog, send mail to team with user question, name and email ID of a person who is a query. We can write the gather. Okay. Gather? User name and email idea of a person who writes a require. Let's see whether it works or not. You can try by different instructions. So when you've got the working AI agent, you can set the whole instructions as same. Okay. So you need to try the A agent's instructions whether it's working or not. Okay, you need to try different so many times in order to make AI agent to work perfectly. Okay. Now we will do that again. This will say, can you tell me about offers on all products. Let's say whether it will give the answer or not. No, you can see it is working perfectly right now. The complete documents do not provide any information about all products of my company name. If you would like, I can help you to draft an email to the relevant team to inquire about the current offers. So could you please provide a name and email ID? No, I can see. When I give my name and email ID, Okay. This is my name and email ID. Let's see. Now it will start sending an email. You can see here. Now you can see here, that it has done. No it has send email to my GML. You can see. I've sent your inquiry regarding the offers and all products to the team. They will get back to you with the information. Is there anything else I can assist you with? So you can see the email has sent. We will check our email again, come back. Now you can see the GML has you can see it is a subject Inquiry about offers on all products to your team. I would like to know about the current offers available on all products. So could you please provide the details of directly that I can find information, right? It is a simple you can see it is a side. It will come directly from the customer, right? Now you can see. When I reply this particular one, it will Now, if you think here, the question is good, but I doesn't find any GM ID from the user, all those things to save the time, so we need to come again, do the perfed instructions for AI, like gather user name and let's dig. I run again, whether it is taking the information from the users or not. Let's go to or not. Now, you can see, it is not working perfectly. You can see user query. Can you tell me about offers and all products? This is a user quotien. It is a user name that is Sive and it is a user email. This is the email that I have just given in the here. Now it is perfectly working. So you need to try to change the instructions as our requirements, and you need to check whether the AI agent is working based upon our instructions or not. So you can try at different times, you can change the sentences. You can try, you need to execute the step and you need to check the action of AI agent. So now we have got the user quien and the name and GML. Okay? So I can send the answer about this particular question to this GML. Okay? This is how you can escalate the user customer question. To an agent. Okay, when the EI agent doesn't find the relevant answer from the company details dog. Okay, I hope you understand these points very well. So this is how we can create the amazing AI agent for our customer services. Okay? So in the next section, we're going to see how we can track our AI agent progress, whether the AI agent is solving user queries very efficiently or not. Let's jump into. 8. 3.5 Adding Ai Agent Logs using Google Sheets: So in this session, we are going to see how we can track our AI agent progress, whether the AI agent is solving user queries very efficiently or not. So to do that, we are using Google Sheets to track the AI agent progress. Okay. So for that, you need to create the Google sheet in your account, which is connected to the NTN. That is simple. And I have just created the simple Google Sheet with the AI agent progress log. Okay, title and with the two columns, that is user query and AI agent answer. That is sheet one, right? So come to the atm. Okay, now we need to connect another tool that is Google Sheets tool for dot come and just click here and just a search for Google Sheet. You can find this one. So interrupt that just come here directly click in the tool plus button. You will get the exact tools you need. So just a search for Google Sheet. You can find it here. Google Sheets tool. Select that. Now we need to connect our Google Sheets to this Aten within this. Okay, Google Sheets account set automatically St sheet within document. Okay? This is whole document, and this is a sheet one. Okay, next get rows. So we are looking to do the automation operation is. So we are appending a row. Okay, that is create a new row in a sheet. So what happens here for every Interaction between customer and AI agents automation, this AI agent will create a new row with our user query and AI agent answers. It will create a new row which contains the parameters user query and AI agent answer. I hope you understand. So far there come here apenErow select the operation that is append row, that is, create a new row in a sheet. Now, from list, you need to choose which document or which sheet you are looking to connect. So you can see you can find our created Google sheet that is AI agent progress logs. Just click here and from the sheet. So we have only sheet one. Okay, as you can see here, Sheet one, so we need to select only Sheet one. Not connector. No, it will automatically a map each column manually. So for that, you can see this is a two column names. We have already seen in the Google Sheet in previous right, user query and AI agent answer. So we need to set the parameters, the val we need to send to this Google Sheet to fill here, right? To do that, come here and just select. We can see there are any option from AI. So to do that, we need to execute the previous nodes. For that, we can just select let the model define this parameter. Now we'll simply do these things. Now we have connected the things very. So another thing, so we need to rename our tool name to recognize by the AI agent. So for that, I will just write A agent logs. Agent logs is our tool name. So you can kept anything if you want. Okay? So this is a tool which is the name is Agent logs. So now we need to update our AI agent instructions, system message. So you can see, I have updated that is Tard tool. So agent logs. So this is the name of our Google Sheet tool that is Agent Logs. So I have just written the simple instructions, right under this JCNhat input. So what the chat input the quoti which is asked by the user query in the chat interface is called an chat input. You can see here. You can find it here, right? That is where your company located. You can see here you can drag and drop directly here, right? Now, if you focus for here, let's see this one. Now you can see. Agent logs, this is our third tool, which we are connected to the AI agent. You can see enter this. It is a chat input that is user query. Okay, enter this one. Enter this message or user query in user query column. Okay. And you are actual response. Okay? You are actual response in AI agent answer column. So we have just guided the AI agent. So you need to for every interaction between AI agent and customer query. So you need to enter the customer query as well and your actual response in the Google Sheets that we have connected to you with the name agent logs. Okay, I hope you understand these points. Let's go and check whether it works or not. Okay, to do that, just come here. We need to run this again. So let's take Open hat. I will clear the section. Now, I will just write. What is your company name? Take whether it will works or not. No, you can see it is running. That means our flow is connected very perfectly. Okay. Let's take whether our AI agent is, you can see, I'm unable to disclose that information. But we will see the answer in the chat, we'll get or not. But if you see her, we got the name of company in the chat, but she doesn't get anything in sheets here. Okay, you can find this one. So you can see if there is a problem in it. Just click her again. So we need to just click this one, remove this one, and you can see you can find the correct answer. Let's see. We'll try again. Let's say we have now, you can see we have got the correct answer. That is, what is your company name? The company name is at my home, Appliance is private limited. So sometimes the agent lets so we need to write the correct instructions as well, right? And so we need to update the notes as well in the Google sheet to do the work correctly. Now, you can see this is one, right? Okay, you have already successfully connected the Google Sheets to track our AI agent interaction between customer and AI agent to track the progress of AI agent, whether it's solving the customer questions very clearly or not. So we have successfully connected this one, right? So you can do by yourself, you can add the Super baase instead of Google Sheets or any other things you need to just come here. You need to click on the plus button. You need to find the relative Okay, you can call into the HDTV request tool as well. You can send the parameters, all those things. You can save the progress of AI agent in database or anywhere to track the progress of AI agent. I hope you understand this 9. 3.6 Adding Supabase Database to Ai Agent: Now if you are looking to track your customer interaction, storage for permanently in your database, not in NTN. So in your database, we need to use the superbase. Okay, Superbse is a great database for better databases. Okay, you can learn more about this superbase in YouTube tutorial for more information. We are looking to connect this superbase with our AI agent for our memory instead of Enten memory to permanently store our customer queries data in our database. To do that, just come here, delete this simple memory. So click here. It cannot find that directly super baase here. Connect our super baase to our AI agent for the memory. You need to use the post grace chat memory. Okay. Just click here. Now, we need to connect our superbase with this post grace chat memory. I have already connected. Okay. Now we'll just connect this one. This is our so I have just connected, so I will show how you can connect this one. So before going to that, let's take. When you click here, come here. Now we need to go through Superbas just to go for Super Base and just click on the Start word project. So you need to click on the Start Word project and just sign up with your Github, if you have or email with the password. After successful, you will get this creating the new project or create new organization. Okay. So I just give the name ad that is project name as N and Demo. Okay. And I'll choose always use a strong password to remember BU. So we need to copy this one. Strong password. Okay, done. Copy. Now, remember one thing you need to save your database password, so we are required this information into connect with NNDon. Okay, so just copy this one and just click here. So you can choose your near regien time as well. Okay? Come, click Create New Project. Okay, you have successfully created the new project. Now, we need to connect our tone with the Superbase. To do that, just click on the main that is Connect button can find this stop. Just click on the connect. We'll get this transaction puller section. We need to use this particular one to integrate with this one. So to do that, come to our Natan so you can see host. We need to get this host from this one. We can see you can use this host name as well, just copy it and paste here as the database, host. Port number, copy this one port number, copy and come to use this port number as well here. Use the database as a postgress come here and just paste in the database as a postgress and use the user, just click here and user past here. Okay, for the whole mode transaction, we doesn't wonder. So for the password, just paste here what we have just created the password in previous. Just copy and paste here. That is simple. After that, you can save it, our post address account is successfully connected to this post address chat moment. So our superbass connect to the AI agent to check whether it's working or not, just double click. And we need to find this one key, right to do that, come here, execute a previous node. So we need to write the simple high Now, what happens? Let's see. We need to go here, we need to dragon rob. Let's take session ID. Just come here, we need to give the session ID. Let's see whether it is ox or not again. Now it is running perfectly, right? So instead of NTN memory, we can easily save the chat interaction between agents human and the AI agent in our super base permanently without relying on the NTN memory, which is temporary one, right? And even though you can add so many extra features this particular AI agent, adding another google sheets which gather the user feedback from the chat itself, right? So after the solving customer query, so you can ask a feedback from the user to provide it about how the customer support is going on. Okay, like that to improve your customer support system very well. You can implement this, so I'm giving an assignment to you. Please try this whole AI agent by yourself. After you can get the template, also, you can directly input the file from here. I will provide the Jason file. Now, you can add the extra feature that is extra tool, come here, add Google Sheet, create another Google Sheet. After that, connect here and write the prompt message that is system message about how to gather the feedback, try yourself can expand this particular AI agent for more advanced features. As you want to open the next session, we're going to discuss how we can create the EI agent for the form submission, email, and telegram as well. Okay, let's jump into that. 10. 4. Building Form Submission Customer Support Ai Agent: In this session, we are going to see how we can create the same AI agent for form submission. If you are observed, so many websites have their form submission, in which they will gather your name or email ID, anior subject, anior message. So like that, we can create an form in NADN itself and we can host a form in our website as well. Okay. So for that, we are going to duplicate this exact AI agent. So instead of a chat, we are us the form submission method. Let's see this in the session. Now, what we need to do, come here. Just copy this one again, right? Just paste again, COP. You will get the same thing again here. I will drap upon here. So as the default, the chat, so we can integrate the chat, only one in the same workflow. In the same ter, we can integrate only one chat as a default. We cannot connect two charts for the two different workflows in the same Na ten editor. I hope you understand these points. Now, our goal is to create the EI agent for the form submission. Whenever a user submit their query using form submission in the website, they will get the email from the EI agent. Or from the human agents who are taking that particular query according to our instructions as we have seen in this particular chat AI agent. Now in this we will try to connect the form submission. To do that, come here, just click on the plus button. Search for the form trigger. You will get this NATN form, even though you can add so many form that is like a telegram form trigger that you can easily integrate. But I recommend while you are learning this particular basic A agent creation, just to go with the NAN form. Now you need to click on the triggers. Just click on new N a ten form event. So let's see. So this is the Form trigger. Just connect this one. That is fine. Now click on the arm form submission again. This is a simple URL. We will get the two different URL that is test Ud in production URL. So during test, we will use this particular form link in order to test our workflow whether it is working correctly or not. Let's see. Let's reconfigure this particular form. So our authentication is done. So our form title is let's go with the contactors. That is simple and you can write a form description, so you can keep this as it is. We will get back to you soon. This is a form description. Now, let's add form elements. What is a form element? That is about name, fill name, element type, like what is your name? You can take Enter your name as well. Element type is text. We will just pick this required film. Now we can add another form element. That is what is email ID or Enter email ID like. That is simple. Now we can select the email type here and just click on the required field and add another email, sorry, another element that is subject. That is we leave that is add text element type as it is. Just click on the required field. Now we will add the same again. Another is message. So we will leave the text as it is. Enable the required fill. Now you can see the respond when you can choose buffalo finishes or form submitted. It is a trigger one, we will choose form is submitted. Now we can execute the step. When you try to execute step, you will see this is the form looks slide. Okay. You can see here it is a test version of R form. It is our headline. This is our description. This is our form elements that is we have earlier created. You can see what is your email ID, subcheck R message. So when I click in the submit, the AI agent will take that particular message and it will respond as we have seen in the chat according to that. But the response is go through the email ID. Okay, I hope you understand these points. Let's check with that in walk or not. So to do that, so we'll simply fill this one like my name is Leste I tic simple mi. This is my GM ID. Let's take the subject. I need you can take anything, right? Can write anything. No my message is, what is your product or where your products manufactured like that? In my case, I will take just one simple quote. That is, what is your company name like that? Or what is your company history you can take like so this is simple question. Let's submit this one. Now the form has submitted. Now, you can see this are some JCNFle data. You can see here. What is your name side? Okay, email ID subject message submitted. So this is simple, we have got the things. And remember one thing, you can use this pin data. So when you click on the pin data, this will be a store and fixed here. So even though our workflow is not started from scratch, no, I have pinned this one, so we have the data here on. Now what we will do just click on the Agent P. Okay, now we need to so our main objective is to solve the user Cody right. So what we need to provide to AI agent is, we need to provide the user message here, right? So we will just remove this one. Now, come here. Dragon drop this UR message and just keep here prompt user message. That is simple. You can see query that we have cut in the chat earlier, right? You can use the same thing here. Just come here, drag and drop your UR message here. And next one, we need to do the same thing in the prompt as well. So just come here, just remove this one. Even though you can just copy this one from here. And remove this one. Past here. That is simple, right? So we can do the things as we can change system message prompt instructions as our requirement. Let's see whether it works or not, execute the step again. So we have got some error. Let's see what is that. Now we have got the AI agent. Error in this one. Let's see, open error now. We need to remove this particular one. We need to change this database also, click here and we need to change this session with the UR message. Let's try whether it is ox or not. Or you can come here define below again, just keep it here key and just take this UR message again, this past here. That is simple. Let's see whether you walk or not again. Just click on the run again. So we have got some errors. Okay. So we need to change each and everything, when we have corrected on form submission, right? So to do that, so we have some major issues, right? When the users submit the form, so they will get the email, right? So we need to set up the email first here. So let's connect to the Gmail. Okay, send a message. Now, we need to just go with the Gmail account B send. So we need to take this m form submission that is their user email ID. Just dragon drop, what is your email ID come from here. For the subject, you can write anything you want, right? But the subject can write your solution. Okay, support from team like that. You can solution for I'll drag this message or subject. I'll take this one as set. We can see solution for Okay, you can see. So I will copy this one again. We'll write in this way. No, you can see it. So it is a solution for this particular subject. I have just written. You can write whatever you want. Next, we will take as a email type text for the message. We need to take the output from AI agent. That is simple, drag and drop this one, I put it here. That is simple. Okay? Just leave it here. Now our message is sent very clear. So what happens when the user submit their response here with message, the message is supposed to agent, the agent will retrieve the answer from the document we have given. After that, it will send the message to user. Okay. So in previous case, we have seen the agent will retrieve the answer in the chat itself. But here we are using the form submission, the user get the email what they have submitted in the form. With the AI agent answer. Let's check either box or not. We need to check some errors before going to that. Just click on the company doc again here. So we need to change each and everything here, right? So if you see here, Jason, in the company doc, TTB tool, we need to change this one with this parameter. What is a parameter form submission? That is message. We need to change this one. Just remove this one. And just drop and drop your message from here and just keep it here. Now we have done clearly. It is right. Now, just click on the execute step. Okay, we have got the output. Now let's try from the scratch in order to see whether it will put in the information from company dogs very well or not. So to do that, I will just unpin the data and I will run the thing same again. Just click on the execute workflow again. It will start showing the form. Let's fill this fastly. Findings Let's take in This is simple, my email ID. Now the subject is I need product. Now your messages. Let's take what are the products of your company? This is a potion. I will sup it simply and you will see the workflow is running, you can see here. You can check. You can see it is retrieving the data from the document. Now, generating the answer. Now, you can see it has tented successfully. Let's check the email again. Now you can see we have got the email from the AI agent. Now you can see solution for I need product. Okay, this is a subject, and this is the answer from AIgen. Okay, that is simple, right? So we have got answer from the AI agent to the mail. Okay? This is how we can automate the things with the form submission as well. Even you can see the agent logs as well. Okay, whether it has doing the things or not, right? So you can connect this one, and it will work same like the chat AI agent. But the input and the output process is different from the chat itself. Okay, future marketing purpose as well. Okay. So not only that you can implement much more things with this system, even though you can add some Google shat to see whether the email has sent or not. Okay? So to implement that you can go, you can create another sheet, connect the Google sheet here, and trap the AI agent whether the email has sent or not. Okay? To do that, you can do easily by just click here. Try to connect the sheet. And you will get the output from the GML, that is you can get the label as you can drag and drop this label in the similar column you have created in the Google Sheet as status of email sent or not. Okay? You can drag and you can track the progress in the Google Sheet as well. So I'm giving the assignment. So to by yourself, when you change the input and output, you need to change the chat database and the company do input parameter as well. Okay? This is how we can use to implement the on form submission method solve the customer queries. In the next section, we will see how we can create the telegram agent. Okay, let's dive into that. 11. 5. Building Telegram Customer Support Ai Agent: In this session, we are going to see how we can create a similar AI agent chat bood for our telegram. Let's see in this session. So for that, I will just copy the same AI agent. And I will paste it here. Now you can see. So we have got this one. Now, what we need to do we need to connect our telegram with this AI agent. To do that, just come here, click on the plus button, search for the telegram. You will get this. Now, we need to take the trigger. To do that, just click in the trigger section, come to on message. You can find it here. Now, we need to connect our telegram account. Just come here, click here. I will delete. I will start from the scratch that I will show how you can create the telegram. Now, come, set a credential, create new credential. We need to get the access token. To do that, go to Telegram just log in your account here, and we need to go to the bot further. Just come here, search for the board further. We will get this bout further. Just click here. Now you can create a new bot. To do that, just come here. Now, Anew choose your name. So we need to create a new bot in the bot father by giving our name. So I will take name as AI agent. Let's take 02. Okay? This is my AI agent name. Let's do this one. Now, right. A new bot, good. Let's choose your name. So we need to create the unique user name for our bot to do that. So we'll choose user name for it. So remember, you need to use the bot at the extension to create this board. Lest I will take the name as AI agent. AI agent, 02 bot. That is done. Sorry, the username is already taken, please try something again. I will take this one again and I will write anything like AI, let's take agent. AI NN 1002. The name is already taken. We will take another one, copy this one. So we need to write this bout in the last. Now you can see, we have created our bout in the telegram. Can see perfectly, you can see the De congulations on your new bot, you will find it here. This is we need to get this TTP API in order to access this spot. Just to copy this HTTP, come here NA ten, just paste here, access token and save it. That is simple how you can connect the telegram with your NA ten. Okay. Now we have successfully connected the telegram to our A. No, come here, just connect this one here. After that, we need to run this one. To run this one, go to just click here Exhibit step. Now come to the bot further. So you can find it to assistant here, just click here. Now you have start this. Now we can write the answer like, what is your word company name. So we have sent a message to or what. Now what we need to do, you can see here. So let's exhibite the button again. So you can see we have got the parameters from the pot, you can see this is our message that is what is your company name. Now, we will configure each and everything AI agent tools according to our telegram knock on the AI agent too. So we need to remove this one, right? We need to update this prompt according to our input. So in this case, our input is telegram, right? So what we are looking to get this is our quotient, that is text, you can see, what is your company name? So remove this one. So the prompt is the text that we have send in the telegram bot, so you can find it here. We need to change this particular input in every case that we have used in in the previous AI agents. I will replace each and everything. Now let's execute this step. We need to remove another thing here. Let's go with the come here, we need to change this also, right? That is done. Now we need to change each and everything with the input we have given. Okay. You can see we have changed all those things here. We need to remove that is simple. Now we need to check whether it is working or not. We need to also change our input in the Google agent logs tool, send email tool. We need to configure each and everything according to our input. Let's take a message from the AI agent and the N support team that is good. Now we'll try to run again. No it is running perfectly. Not executes successfully. But remember, one thing, we doesn't get any answer from the EI agent. Now, to get the answer, we need to connect the output node also to send the text message for the given input. Okay, to do that, just come here, click on the plus button, search for the telegram. You will find it, click on here and just go to the message access, send a text message. You can find it here, right? Connect this one. Okay, we have collected the telegram account message, send message. So we need to get the chat dy. How we can get the chat dy, come to telegram trigger. You can find the chat. Here are ID. Come here and just dragon drop it. And for the text for the text, you need to take the output of a agent, come here and just paste in this text field. That is simple. When I execute this butter, you can see this is the text is sent to our agent. Let's see. You can see this one we have got the response from nt. That is simple. When I give a list your products, So we'll get the answer directly from here. Okay. Let's see working for doing this. Let's take. We need to execute the workflow, right? Let's see this one. No it is retrieving the information from the document. You can see it is logging. Now, you can see, we have got the answer that is here. You can see our AI agent board giving the response. You can see we have got the response from this particular question list or products. Here the list of products offered by ACM Home appliances you can see these are the products list. So this is how you can simply connect the telegram for your customer support AI agent. Not only the telegram, you can try for the Whatsapp, Facebook messenger not only that, you can check the logs of AI agent and message sent. Okay, to do that, you can come here, add a new Google sheet, track the user question and the answer, but just come here, create a new sheet and connect the Google sheet with the two different columns that is user query and the AI agent response. You can expand this particular telegram AI agent for you or at use cases. I hope you understand these points. So I am giving an assignment for you. Okay? You can create the similar AI agent for your WhatsApp and Facebook messenger, go just copy this, try it by yourself, connect with the Whatsapp. It is similarly same process. If you have any issue, you can find the templates in the Internet. In the next session, we will see how we can create the similar AI agent for our EML trigger. Let's dive into that. 12. 6.1 Building Gmail Handling Customer Support Ai Agent: The previous session, we have seen how we can create the similar customer support AI agent for our telegram ball. Now in this session, we are going to see how we can create the same AI customer support agent for our email trigger. Whenever I got the message in my Gmail inbox, the A agent will take over it and it will give the answer in the email itself. Okay. Let's dive into how we can create this one. I will copy this whole AI agent, again, and I will paste this here. Now you can see it has simply pasted this one. So I will take and Dragonrop here, and I will just take this one. And now, we need to delete this one telegram. Now, we need to connect the GML. To connect the GML, come here. A click on the press button, search for the GML. Now, we'll get this one, just click here. Now we need to go to the trigger section on message received. Okay? Select the on received message. Okay, just click on the simplify button. Now, whenever I got the message in my inbox, the AI agent will take that question and it will solve the customer query, and it will send the same email. Let's see how it will works. So I will just take this one Gimil trigger and I will connect this one to this here. Now what happens? We need to execute this Wplowo before that, we need to execute this one here, right? Now it is executing now it is successful executor. Let's say, let's check. We need to update this prompt user message wherever it is, we need to update this one. So to do that, we need to find that text message, right? So we have got the text in our GMA, so we need to find it. So we can see how where it was. Okay, before that, we need to send message to this connected GMA. To do that, I will send a simple message that is I Nudes solution. I need solution, and I will ask you a question that is, what is your company name. This is a simple question. I will ask to the agent whether Ts works or not. I will send it. Now you can see it is successfully sent. No, I will check my email whether it has received or not. Let's refresh this one. Now you can see we have received the message, that is, what is your company name? Now I will come to my workflow, and I will run this one again. It will receive the message or not in the workflow. I will just click a agent. Now you can see we have got the answer like Okay, it is very more technical bit to solve this issue. We need to click here the simplify. Let's execute this one again. Now we will see agent. Now you can see this is a simple one. Now you can see. So we have got the text message. That is what is your company name. We have got this, right. So we need to change this particular prompt user message with the text we have got in the Gmail. So just drag and drop that here and just paste it here. So we need to change each and everything where it was required, right? And I will take this one again, we'll copy this one. And I will change that is simple and we need to change in the database as well as a key. Now we need to change in the company doc to retrieve the answer from the document in the content session. Now we need to change in send mail. Come here and just so it will automatically stake that. In AIgent log we need to change the user query with our message received that is from the mail, that is simple. So we have updated each and everything. Now, we need to connect the output. When the Agent retrieve the data from company dogs, so it should send the solution, right to the user. So to do that, we need to click here and we need to search for Gmail. Now we need to send a text message. You can see send a message. Now, before connecting this one, we need to execute this one. I will just delete it. I will execute this one from here. Let's see. Not is working, it as taking the the foundation it was saying. Now what we need to do come here again, just click at the GE, send a message. And now we need to take so we have got the agent three answer as well as griger. Operation is send, so we need to send the message right, take the gtrier, so take that email from where it was received, just from that will respond to this particular email from where the message has received. So take the text. And for that message, we will take the AI agent output. That is simple. For the subject, we will take you can take anything you want. So I will take solution for your question. Or you can take anything you want. I'll just take solution. We have successfully connected the Gmail. Now we will execute this workflow again. So it is taking answer. Now it is retrieving the information from the document. Now you can see it send the message. We will check again with the sending a message, that is, I will send the message to this particular EBL which is connected with NAN, right? Now I will send the message like I want product details. Take the question as last product details. Now we'll send the message to particular company email. Now, let's say we have got the mail from the user. We can see. You can see I want product details. List your product details, we have got this. When I click here, execute the work flog in. You can check this one taking time. That's whether it will work or not. Now we'll check our mail, whether it has come. No, you can see this is our solution from the AI agent. That is solution for quotien. This is the answer from EI agent. But the question that we have rise, we have sent to our company mail. That is you can see this is our product details. So when I send the email to this company mail, okay, you can see the company mail received this particular question. Now after that, after a few minutes, I got this particular solution from the company mail. You can see this is. Okay. This is the answer from AI agent. So this is how you can connect the mail system for your support AI agent. So connect the third party services like Bravo or email marketing campaign tools. In the next session, do you see how we can add so many features to this particular email AI agent to classify the high priority, low priority, medium priority message. Okay, to prioritize, let's dive into another session. 13. 6.2 Adding Advanced Query Priority System in Gmail Ai Agent: In the previous session, we have seen how we can create the agent for our GML system. Now, we will try to integrate some advanced future for our GML, customer support agent system in which we can classify the EML as a high priority, midum priority, and low priority. Let's dive into how we can create such a system which we can handle the customer queries according to the priorities of their queries. Let's dive into that. Just copy this one. Just drag up here, right? So now we will try to integrate advanced system here only. To do that, we have the unique node that is test classifier. So we need to use the test classifier in order to divide the specific messages or queries into the high priority, medium priority, and low priority. To do that come just click here and you can search for the class textifier you can see text classifier. You can use this node, right? Now, that's serious one. So this is a text classifier. So we need to integrate this one here. I will delete this one. Classifier, add it here. You can directly click. You can add. Now, after that, we need to just open now what you need to do? You need to just click in the add category. Before we go into that, just click on the Run button, Jimel trigger. It will run. After that, you can click on there. So you get some details from the geometric. Now, we need to classify the text. So in what basis this particular text classified will dive or we divide the category and send to the AIgen. So to do that, we need to use this nippet and you need to mention here. Okay, now we need to add some categories. That is category number one, that is high priority like that. So just take high. Okay, I will take that description as I'll take high priority, high priority. So we can use some keywords in which this particular node C can classify and divide the category as well. I will use these keywords. When the email contains this particular keywords, the AI will automatically assign the category according to our keywords. Okay? Let's see. I will take product damage. Okay? Now we can take with the meetings now, even you can take any of the keywords, when the email contains this particular specific keywords, the EI will automatic ascend this message as a high priority. I understand these points. Now I will take with this same category that is another category, just copy paste here. I will take this one as a medium. Now, Medium. That is fine. Now for the keywords, I will take refund, or you can take off words like that. You choose you can enter as many keywords you want according to our requirements. Now let's take the lowest category that is low, and I will copy this again and I will just placed here. Now what? I will just I don't enter the keywords because this EI will automatically solve every issue. Okay, it doesn't contains refund offers or product damage or meetings. Now, you can come here, add option. You can see the system prompt template. You can see. It is already default or prom template. This test classifier will use, in order to assign a category, let's see. So this is simple. Now what we need to do? We need to connect the chat model. So you can directly connect this from here. Okay, you can direct it from here. Otherwise, you can choose another chat model as well. So I will use another chat model, I will take the GPT four mini. This is a simple chat model we have connected to it. Now, what we need to do. According to the Gmail priority, it will goes to this agent. Now, we can add this one directly here. Directly. According to the type of any customer send a query through the Gmail. So we will get the Gmail in our Gmail inbox as well, right? Come here. When the customer will send a query through the email, we will get the email here, right, inbox in the primary. So to create this, so we need to create some labels. I will delete and I will show from starting how you can create the labels, how you can do all those things. I will clear this one No, let's come to the Gmail, which is your company email, come and just click on the labels plus button. You need to write the priority emails a label name. That is high. Just create another one. Let's take medium. Next take another low. That is simple. No, we can easily add a label color. That is for the high priority, will take the red, for the low I will take as a yellow, for the medium, I will take as a orange. That is fine. So we need to send an message which contains the high priority to check. So to come just add here and just set for the G. So we need to add a label. That is, can add a label. Okay. Add table dream we need to use this one in order to check whether it's working or not. So just come here, just click and just select this one. Now what we need to do. When you run this one, you need to run this as well to get the output from this one. Okay. Now we have got the our quotient is here. Let's see. This is a low pratty right? That's why it get the output from low priority. Don't worry I will show each and everything just follow my instructions. We can do such a thing easily. Now come here. At each thing here, let's take this one. Now, let's see this one. Now, we have got the output from the text classifier. For you to add a message ID, you can take this one ID from the text flashifer. For the labels node, ID, come here and select. It is a low priority because it is coming from the loop. When we will get the output from the text classifier. If according to our message, it will categorize, which it is a low priority or medium priority or high priority, you will get the message, our GML get the label. Let's see. I will run. Now we will see our GML here. I will run this again. Now, come here. That is low priority. You will find this one. Okay. This is our solution we have got right now. Okay. So you can see the low option here. Like this. It comes from the low. You can check here. We have got the answer. You can see it is ascent. It is a low priority. Okay, you can check. We have got the low priority in the low label section, low priority one. Okay? Now, what we'll do is just copy and we will duplicate it. We will connect this one very easily. I will just click here and I will duplicate again and you can add the same thing here. We have connected the thing for every message. When we got any message to our inbox, the text classifier avail categorize based upon our priorities and we will get the Gmail, which is one according to our basic thing. I will just connect this one. That is simple. Now, let's try whether it works or not. To do that, just come here. We get the answer from test classifier. So we need to do things Jim and trigger as well. We need to change each and everything. Let's take the snippet, we will change each and everything from here as well. Let's take control plus C. That is done. Now I will change in the database as well. Company D HDB tool, I have done this one, now come to the logs that is simple. Now we have done. Now we will run this one again. That's working perfectly. But let's see if we will get the output here or not. Now, you see, we need to remove this one because we have error. So to do that, take the Jimel trigger and send from where the message has received, right. So to do that, we will take this one as from yeah. We need to send to the user. Next, the output should be there from AI agent. We'll do this. Now, we will try this one again, we have another problem here, the subject. Now we will remove this one. That is simple. Now what we will do. Let's check whether it works or not, from another male, I will send the query to the company mail. Let's try whether it will work or not. I will send a simple message which product damage. I will send this particular company mail. But the subject is I need rg support. That is subject line. No, I will write the message. I got product. Which damage. Okay, you can write anything. Okay? I've got a damage to product like that. Okay, let's send this one. So you can see it has successfully sent. Let's check our inbox whether the message is received or not. Come here, come to the inbox, check primary, let's refresh it. Let's check to execute workflow, right? So we need to execute executing. We can see it has taken in the higher priority because our keyword product damage has entered in this text classifier, but can see retrieving the information from here. You can see that's good. Now you can see it has sent the mail perfectly. Let's check. Now you can see, I need gen I need urgent support here. I got damaged. So you can see it is a high priity but we have got in low. Why? Because we have doesn't change the label here, label names. Okay, we need to change this one as a high. Okay? We have simply forgot it. Okay, let's change according to our node. Okay? This is a medium. Now it will works perfect. In order to check whether the workflow is working correctly or not, you need to come here, you need to enable the active. Now we need to go to the customer mail and we need to send the message. You can see I'm sending the message to the company email that is Aponte atregmil.com. I just send the IN answer. I will take as a low priority answer like I need some product or companies information like where your company is where you company is in India. I'll just send this question to the company mail. You can see the message has sent come to the mail. We will see, we will run this again. I will refresh it. Okay. So we don't find the incomes. You can see I need answer what company is in India. Now we need to wait for 1 minute in order to see it is working or not because our workflow is run answer for every minute. Okay. Now we need to fetch that test event as well. So we need to activate this one, execute the workflow. It will take on the thing it going to the load that is working in the perfect manner. Okay. Iterative the answer from here, it is adding logs, and it will send a message to our return. You can see now we can check our customer mail. You can see solution. The company is in ACM Home Private Limited, headquartered in Bengalur in India. So we have got successful the answer from the workflow as well. We can check the company mail Okay, but simple refreshing, you can see our label has low. As successfully, we have created a workflow in which it will categorize easily. You can see you can find it here. Again, I'm good to Heritage. Okay, that is a company, customer mail. I will simply ask the question about Let's take, we can take the same thing like Promptek as a company, mail, right? I will take this one, right? And for the subject, I will take as I need offers. And I'll just write the message as, can I know more? Can I know more about the offers you Okay. Can I know more about the offers? I will send this one. Now I will wait. Now you can see the message I sent successfully, now, I will check the company mail, I'll refresh it again. Now we will see the message as received or not. I again you can see we have got the ideal offers. We have got the message received right now. Come here. Okay, wait for 1 minute. Now again, run this one, execute workflow. Now we'll check whether it will works correctly or not. Now I can see, there is a problem in message too. We need to change the label, sorry. I will change it to medium. That's simple. As you got the step. Let's execute this again. You can see it is working right now. It is taking the information from the company dog, I will just give the answer and it will send it to the customer. Let's check our customer mail. You can't get the solution. I couldn't find the information about the offers, and email has sent it to the team provide further assistance. Okay, that is simple how we have got this one. So we have already done the previous one. If you follow on the tutorials, you have find our function of this AI agent in previous right. No, it is working profitly because for the offers, we will take the medium category. Now we will check this one and we will refresh this one. The alt should come. This query is labeled with the medium because our medium we have added some keywords, you can see in the workflow, we have added some keywords that is for the medium category offers we have ordered as a keyword. Whenever our customer message contains offer related keywords, it will just categorize into a medium, and it will send the mail and it will create the label as medium in our company mail, and automatically the Aagent will send the answer to the customer mail. Okay, you can see this one we have boards. As in. Now we will try for the higher priority email. Come here, the customer mail, I will send to the company email that is IEI prompte and I need the Adgent support. Okay. Let's take this agent support. Sorry. I got a product. I got a damaged product. Okay? I will send to the company mail and you can see successfully sent. Now we will check whether our inbox get the message or not. Let's refresh it. As refresh, you can see, we have got the message from the customer. Urgent support, I got a damage pronuc. We need to wait for 1 minute in order to label this message as a high priority because our workflow has a 1 minute trigger, We can find this one here. Now what we need to do? We need to execute the workflow again, right? Let's execute this one. It will take the question. It will classify this one. You can see it is taking at the high priority. Okay, we need to check this one as we need to select this one high priority. Sorry, we forgot this. Now we'll execute the step again. Now you can see here, it is working right now. No run this one. No it is running perfectly. It is taking this one. It is taking this one. We have send mail to the company. Can see it has sended this one. So you need to run this last step. Now, we have successfully done this process. No can see here. When I refresh this, now you can see we have got the high priority label. Agent support Kot damage product because we have entered this damage product keyword in the text classifier that AI can easily classify and divide into category, which is the high priority low priority and medium priority. I hope you understand this point. Okay, now, our company inbox has you can see, high priority medium priority, low priority label. Even you can see here what the message we have got in the high priority labels, medium label low as well. And we can see that we have got an answer through our customer mail that is solution we have seen this is our answer. I have logged the user query regarding the damage and send an inquiry to the team for further assistance. This is how we can create this categorized EI agent. Not only that you can add so many advanced features like we can add here this same feature you can add in the company mail as well. You can also track the agent response, log whether the email has sent us successfully or not, you can track any errors as well. You can do so much things by yourself once you learn this particular A agents building by USL. We have seen four different type of AI agents. You can add this same feature, right? Advanced feature in this form submission as well in the chat. So I am giving an assignment to you. So please try to integrate this type of advance feature in this type of AI agents as well in order to learn the advance NAN A agent building. Now in the next session, we will see how we can gather the user feedback about our product or company or customer support. Let's champ in in that next session. 14. 7. Building Feedback Collection Ai Agent with Sentiment Analysis: In this session, we are going to see how we can create the user feedback system in which we can gather the feedback on our product or services or for any purpose, we can gather there, and we can easily align sentiment analysis, which is whether the feedback is in a positive nature or negative nature, and we can send a message according to it to the user or customer. Okay, let's tap into that how we can do in this session. So for that come here, I will just scroll below and I will just click. First, that is Form submission. You can gather the feedback from email as well. You can through chat. Otherwise, you can take any social media account like a telegram, Whatsapp. Now in this session, I will show how you can do with the form submission. You can gather a feedback from user in any way. You can implement this exact system. I will show how you can do with the NIDNFm. No, I will just search for the form. No I will get this Notin form from here and choose the trigger that is or new atan form event. This is the thing here. Now, let's tragon drop this one here. Now we'll just configure this one. Let's open this notin form. We will give the form title that is feedback form. Feedback form and the form submission you can take anything like submit your valuable feedback. No improve. You can write anything you want. Now we will add the form element. Let's take if we will gather the information about user that is name like that. Just required fill. Now add another element that is email. I will just select the type of element type that is emailed, enable the required film and we'll add another element that is feedback. Very simple and we'll just enable the required. Now, we have successfully created this form submission. Now we will execute this one first. Then we'll open this. No, we will write a feedback and we will submit all the details for our testing purpose. Let's take Jnames email as, let's take feedback is you can take this your product is good, it comes under the positive or you can take your product is bad like the incomes under the negative. Now we will check for first positive. We'll just submit it. Now we have got the information here, so this is done. Now we will fill this. Now we will add the sentiment analysis node. Come here just a search for the sentiment analysis. So we'll find this one here. Hey, come here, just select the sentiment analysis. Now, we need to submit this one here, text to analyze. So what type of text you are looking to analyze in order to categorize in the positive or negative? To do the sentiment analysis, take this er feedback as a text to analyze because based upon our feedback, this sentiment analysis will divide the category that is you can see sentiment analysis, category can come and adoption. So you can go with a positive neutral or negative. So we will just in this session, we will see only the positive or negative. Even, you can use the neutral as well. Option, you can see the system prom template. You can see how it works. You are intelligent, accurate sentiment analyzer. Analyze the sentiment of the provided chat text category. So this is the category of the particular sentiment analysis. Okay. And it should come only output in the JSON form. I hope you understand one. So this is the prom template, which is a default the deten has wrote in this particular system prom template. Okay? Now, we will execute this term. Before execution, we need to connect the model here, right? I'll just search for the open AI. You can find this Open AI chat model. Now we will select the model of our open AI, let's take I will take with the GPT four mini. No is done. Now what we do? We'll just run this one. No CCT has come in the positive image because our feedback is your product is good. No CC this is the answer. I should become in the positive branch, right? In the negative branch, we doesn't have any output because our feedback is in the positive. Now it is done. Now what we need to do? We can send this particular feedback in anywhere if you want. Now what we will do. I will just connect another AI agent. Let's take AI agent. Now, what happens here? When the person submit and feedback good feedback, now, the user, the customer should get another mail from our AI agent, which contains thankfulness to the user for submitting the feedback as well and asking the permission if they want the future update from our company or not. So we need to connect another AI agent. Now what we'll do? Just connect the chat model. You can do from here, but I recommend use two different chat models for better operations select p41 minute. That is fine. Okay. Now what we will do, come here. So we need to select this one. So connected chat defined below. So we will got the feedback. So we need to take this feedback and we'll need to dragon drop in the prompt user message. Now, we need to just click here and we need to write the system message. Now I will write based around feedback, based around feedback, crap and images which contains thankfulness to to user to user. And take permission if they want future product releases or services like that. Product updates or less. That is simple. The system message you can write by yourself or more. Now we will execute the step again. Let's see what happens here. Now we have got the output. If you see here, subject and the dear user name, all it comes in the information like passage. The problem in here is, I will add the Gmail to send it. Send a message. So we will choose the Gmail account. Now we will send a message. Now, we will take the email from the on form submission email. Now for the subject, for the subject, we can write our according to the feedback, the subject need to be changed, but it doesn't have the separate subject section here. So you can see the subject thank you feedback. It will come under in the message. For example, I will show how it is I will just keep thanks. Now, for the message, I will just send this, I will execute this step and I will check with the mail. Now you can see it has sent. Now I will open my mail. Problem of this workflow is you can see we have gotten mail. The subject is just a thanks, right? Now CC, the email contains a subject, but this subject should be come here, right? You can see here, username, also doesn't get the username as well. You can see we have doesn't get any your name, your position, your company contact information like that. But it should come with the company name, a position, company, all those things, personalize it to the customer, right? So to do that, what we need to do come to the Mr flow, come here. With just write system message d use below details for mail grafting. What I do for the user message, you can use this one or for the user name, you need to drag and drop this from here. Let's take from the on form submission, your name, and you can paste here. We will see, this is the sympathy. Let's go to the enter. We need to tell to AI, so use our company information and add it in our email. To do that, we will use the company information. So let's take name is AI prompt. No, email is that is sympathy. So what it will do the AI will automatically take username and it will send in the user names place, right. Best records name, that is company name. Instead of your position, you can see your company as the AI prompting gmail contact information like this. Let's execute the step again, whether it works or not. You can see we have got the dear user name from here. You can see here, dear side. After the best records, we doesn't get any user name, but we have written the AI prompte name and contact information. Even you can write the name as John, that is simple. According to our requirements, you can edit this particular system message in order to get the best output from the AI. I hope you understand these points, know what we'll do. Let's let's run this again. If you see the Gmail, so we have got all the things. But what if I need the subject here. The subject should be changed according to the feedback as given. You can see that thank you for your feedback I need here. So to do that, what we need to do we need to use another node that is required specific output format. So when you click here, it will show to connect a output parser. Just click in the output parser, select the structured output parser. Okay, now we need to change this one. We are telling to AI to extract the specific value from this whole paragraph. To get the output in the specific format. I hope understand these points. Let's do JCNFmat that is subject of the eat. That is simple. Now body of the. Now you can outfix format as well. Now we need to add the model again so you can directly connect this one from here to here as well. You can connect this one model from here. You can do all the things. Run this again that is running right now. Now, let's see what this output will be. Now, if you see we have extract the subject from the simple whole paragraph, you can see, we have got the specific output format according to our Jason we have given in the structured output parser. We can easily just dragon drop this particular subject here and it will automatically send the message according to our email body and feedback. Let's test out here. Execute the step again. Now we have got some error. Let's see where this node. Let's see what is the problem. Nobody to remove this one. There's an output. Come here, just dragon drop this email body here. That is simple. Now we have got this will run this again. Now it has done perfectly. Let's take our mail. Now we have got this mail. Now, you can see it is done. You can see we have got the subject from the email that is thank you for your feedback. This is a name. This is a whole mail from the AI agen. This is the best records company details as well. See the difference between this one. We have got the template right in the subject, you can see we have got the template, email from the previous one. But after connecting the output parser in our flow, we have got the correct and profit mail from the workflow. Even you can add the Google sheet which you can track the feedbacks, a agent answer. What is the type of sentiment analysis? It is, you can do so much things from here. You can add the different Google Sheet database, just to create the columns and rename the columns like user name and user feedback and a answer and sentiment analysis type. You can create a Google Sheet. Rename the columns, and you can just click here, add it. You can see the collect the information in the Google Sheet as well to track your A agent progress and the user personal information and feedback and sentiment analysis. Now, what we'll do we'll just take this again. Copy this one. I will paste again for negative type of sentiment analysis. To do that, I will just take this one here. I will here. Now, we need to check that take sent dances. We need to craft this system message based on feedback, craft and email whenever we got the negative feedback from the product or support or services, we need to appreciate it. Okay, the AI agent should be appreciate user frustration, all those things, and the A agent should tell them, we will solve the UR issues appreciate and not to go the customer from our hands. So we can write this one. But that just come here, craft an email, which contains a apiation and support support to user in order to solve users. But the negative feedback, let's take you can write the negative feedback, craft an email, which contains aperation and support to user in order to solve, these are issues. You can write the mail as you want for better system message working, all those things. Okay. Now, we'll just keep this one as it is. Now, we have just edited the system message for the negative. Now, what we will do we will just click here, we will do for the negative, right. Okay, unpin that. Just come execute the step again, a submit and feedback which contains negative words. Just run the on form submission again, click on Execute Workflow again. You can see we have got the form open. Now, I will write the name. I will take the same email. Now feedback is. You product is bad. The feedback is anything, it can be automatically, the sentiment analysis will categorize it. You can see it is working, it has come in the negative AI agent. Not is taking Johnson has successfully let's check how the meal let's refresh again. Let's see if we have got the feedback. Dear, sir, I hope this message function, first and foremost, I will look to express my sincere apetiation but taking the time to share your feedback regarding your product, we genuinely value our piece of input. So this is one, you can see the best email. You can say, I'm sorry to hear that your experience has not met your expectations. This is how you can see this one. It has simple appreciate the user feedback, and let's say sorry to the user for a bad experience of product like that, can see. This is how you can do the create the exact system for your product and services, e commerce. Okay, you can implement this particular system chat based AI agent, email AI agent, and you can take from the telegram Watsap, you can integrate this particular form in your website as well to gather the feedback from user and you can track them. Okay, you can improve your product, your company, your business. So we have successfully create the feedback system and different AI agents, multichannel AI agent. Sure of form submission, you can implement the webhook system. Webook system means you can call the feedback from your website. You can take the parameters like feedback, user feedback, user name, email. You can add the Webook simply come here. Web Hook. You can see this is a web hook. You can add, you can use this particular copy to URL. You can page this particular web hook in your website. You can ask the Chat GPT or Cloud to get help with that. Even if you have so many online tutorials, how you can create the web Hook, how you can create to NTN, and how you can send the information from your website app to the NTN. You can do for each and everything if you want. You can just come here. You can add this one from this form submission, you will gather the information from your website. From outside of the ten. Let's dive into our anos session in which we will see what are the different ways of creating workflows by using AI models like Cloud or habit. Let's dive into that. 15. 8. Different Ways to Build an N8N Ai Agents : No. In this session, we are going to see what are the different ways you can create the workflows by using AI modules to see what time come to Cloud. I have just written this whole prompt. I will provide this prompt in the document you can get from that. You can see this is a whole prompt I have used. It is a simple prompt. You can see you are an expert automation developer with advanced experience in tools like NTM. Your task is to design a complete efficient error free automation workflow based on the user specific requirements. The final output should be JSON format or blueprint file compatible with the seleor platform. Ready to be inputed and executed? No gather the information by asking subdivided question, which is the most important pattern. In order to get the information from me in order to create the personalized workflow. Based upon your requirement, it will create the workflow. So for that, we need to write this particular prompt pattern. You can see asking a subdivided is to user. Based upon the answers of a user, it will create the personalized workflow for our requirements, which helps you to generate NIDN workflow JSNFle. Remember the user is beginner, user didn't know about technology, just know the task or automation description based on to creative. Workflow. When you don't know about how we can create an AI agent or workflow in the NI ten, it is a beginner friendly prom. You will just come paste this in the cloud. It will give the answers like this. Hi, I'm here to help you create automated workflow. Don't worry about the technical details. I will guide you through this step by step with the simple questions. Let me start by answer understanding what you want automate. This is the information the AI is looking to gather from me in order to create the workflow based upon my requirements. So when I give the answers for this particular questions, Cloud will automatically create that JCN file, which is ready to import file that I can use Import NIN Editor, which create the workflow. Focus on here. I ask you some what triggers you want, what is your main goal, what triggers this task is required. What types of services do you use? This is the all words. AI is looking to gather the information from me. When I write the answer, you can see I have just written the answer. That is AI agent knowledge base, question and answer board. I just I have written a simple description. When I write that description here, the cloud has automatically based upon the AI agent description, it will automatically think what are the tools required to create this particular AI agent? It has asked simple that is create you want to create an AI PO chat boord that can answer questions using your company's knowledge base. Okay, where is your company knowledge tool? So I need to write this one. We can gather the information from a PDF doc that is about company doc. We have created through the HTTP tool to the AI agent in order to retrieve the information from that particular document to answer the user's query, right. So you can write this one here. So in this, you can upload company dot on Google Drive, SharePoint. You can do all those things from here. You can write all the required ments of these particular questions. You can response, you can give the answers for these particular questions in order to create the workflow without knowing the technical knowledge. You can see all the questions that EI has asked to me in order to create that one. Can see I have just told that I have using the Google Drive to add my company docs to the AI agent. Just have taken as an example, you can see it has created. Again, it has asked some more questions for more advanced thing. You can see Google Drive setup, how you can set up your Gold Drive. You need to write what is a folder name. Okay, list the main folder names, go logs, PDF, all those things because this information is required in the node. Okay, we need to change all those things we have seen already in the previous sessions, right? Email configuration, what is the email? Lauds required information from you in order to create the workflows. This is all the questions that I have given the answers for you. After that, it has asked you some more questions. I have given the answers more depthly. Then Pufate I have all the information I need. Let me create your workflow for the AI knowledge base and question board. Now it has generated the JCNFle Based upon my requirements and AI agent building, it has simply just generated the JCNFle by gathering the information from me. In order to create the personalizer AI agent according to my requirements. That is great right. You can see this is a whole JCN file of workflow. Now what I will do come here, you can copy it or even you can download the JCN. I'll just download it here. Now, I will go to workflow. I will just come to the home. I'll create the new workflow. Now, we have in the new workflow editor. I will just come here three buttons, import from file, select the file here now we have just download it open. Now you can see it has simply created the workflow, simply created the workflow, right? Even there are so many mistakes in that. If you are using the Cloud latest version by upgrading the account, you will get the more accurate EI agent workflow JCNFle. No, I'm just using the fee account. That's why I just returning the previous versions. Instead of taking AI agent, it has taken the simple hat GPT Noo. Instead of that you can add by yourself. That's why we need to learn the basics of the nit and how we can create the when you have that basic knowledge, you can edit this whole workflow according to our requirements. So this is a simple example I have shown. You can do so much things by just describing your requirements by using this simple prompt. It will simply ask questions, just give the answers. I remember one thing, give the answers based upon your requirements. In that case, you will get the exact answer. I've just given some basic description of the AI knowledge base. When you give the clear answers for this particular questions accurately, then you will get the best output from the AI model. You can use the same prompt in the char gibt itself. Not only the cloud can take any model, but it should be the latest version to get the best. Now, you can come here. I paste this. Okay, that's C. No, it will ask the simple questions. You can see. Let's begin. I will ask a few simple questions to understand automation tasks. Please describe your task automation in simple words. We need to write all those things here. Where your data comes from. What should happen automatically? Do you want conditions, filters or delays? Do you want to receive any alert or backup of this automation Fina checks? When I give the answers for this particular questions, it will automatically create the JSNFle. Try by yourself, prompt, and you can do so much things. We have another way that is going to entertain a template section. You can find any templates that already creators, developers create the workflow. You can find them here. Just go to the Naten dot Iolash Workflows. You will get the workflows for every type of rows use cases, industry that is EI, sales, IT, whoops, marketing, document oops. You can see there are a lot of more templates you can use if you want to use this particular template, just click here. You will get here, you can use it for free. Just a click use for free. It will ask input the template to this one because you have already created the same account it will show you this thing. Okay? You can copy the template to Clipod or JCNO you can directly input template here. It will just automatically take to the inner ten editor, in which you can, I hope, understand these points. This is another way. So we have seen three different ways you can do equal. Now, not only that if you have the issue in writing the prompt for the system message in the AI agents, just take the screenshot. So this is our recent chatbot right. I will take the screenshot. I will copy this one. Now I will go to the cloud, past it here, tell to cloud. So please write the analyze Analyze, give you an image. Analyze, give an image, and write system message, write system message to use in AI agent to craft email for positive Feedback. Okay. When you have a how to create the AI agent, just take the screenshot of this particular A agent, come Cloud or Chatb use this analyze, give an image and write the system message to use in AI agent to craft email for positive feedback. You can use for any type of workflow, any type of requirement, any type of A agent that same similar that I'm going to show you. Now, what the cloud will it will analyze the image. It will write system message. You can see system message, positive feedback email response, role if you see here, it is whole AI agent message because we have just given the image. I have just shown as example. You can write even more simple prompt in two to 3 seconds. If I write this one, the cloud will automatically write the four to five prompt to use in the agent for crafting the email for positive customer feedback. You can copy this one, come to the workflow, good and you can just come here agent four. You can change this whole thing here. This is how you can write the system message as well if you don't know how to write a prompt for A agent to work in. You can create so many workflows by describing your requirements by using the prom that I have shown in the different AI models like Cloud Cha JPTy as well. You can come here, you can download it, just click on the download. The whole AI agents JCNFle has successfully downloaded it. You can share with your friends and you can save your computer as well. You can do so much things from here, change Owner, rename it, okay, duplicate. You can activate if you want. Okay? You can see the evaluations, execution, editor, all those that is how you can use Senten for your creating the next agent. So go watch it and learn by yourself, create just learn how to connect the different nodes for the specific automation. Okay? So you can do so much things. I hope you understand these points. I will provide this whole AI agents template JCNFle, come here, create the new workflow button, import from the file, and you can change the credentials you can use for your businesses as well. We can sell this template to your clients can connect the so much apps here, all those things by yourself. Okay. I hope you understand this session very well. Okay? The next session, we will see how we can monetize this let's dive into that session. 16. 9. Different Ways to Monetize this Skill: Now in this session, we are going to see what are the different ways to monetize our AI agent skill. Okay? The first ways you can become an Enten affiliate that you can earn passive income by promoting Enten clot to your audience, receive 30% on all your referrals for 12 months. That is simple. You can see here. You can learn more about this affiliate program, how our affiliate program works, and you can see what are the high ending affiliates. You can see the information here, and then you can come here just a start building. You can just click here and you can join the affiliate program easily. You can get started here. It will show sign up and you can go with that. That is simple. If you have any doubt, you can check it in online like YouTube tutorials, how you can become the Nten affiliate and what are the different ways that I can promote it. Second way is you can become an freelancer. Which you will sell your AI agent building skill by creating the gigs in the fiber Upwork, aguru.com, People Perard. These are the different freelancing platforms. You can submit your projects and you can make the money from your clients. Just post your projects, what you have done earlier, and what you can do, what are the real problem you're going to solve for the businesses. You can create the gigs in the five upworggo.com and PeoplePerHour. There are a lot more freelancing platforms you can search and you can create the account for free and you can just post your projects and services in that when the client approach to you, just talk with them, close the dy, solve their problem, you will get the money from these platforms. It is a simple way to start your flancing from the different platforms because if you are a beginner to flancing platforms or if you are beginner, how you can communicate with the clients, you can learn from this fiber, right? Fiber through these different platforms, you can start learning the communication between client and you. When you gain the experience, you can start your own agency like that. Okay. Now, before approaching to the client, you need to have your Portifilio website. So for the Potifili website, you can use the AI tool as well, lovable dot bole dot new to create beautiful Portoflio of your projects you have done in the NAN. You can use that particular one other than you can use this systemitiBO. It is the easiest Al in one marketing platform. You can create the account for free forever. You can start for free. You can build the amazing landing page. You can post your projects. You can use this systemi dot IBO to create the Portofll website or agency website like that. You can sell your ten templates through this system dot IVO platform. Come here just to create the Landing page, you can sell your AI templates as then. What you can do more with this system dot IVO. Now, the third way is you can build your own web or mobile app in which you can connect the ten as a back end and you can build the front end, connect them to and web is going to live. Want the particular course for creating whole Webb that is from starting to advance, like adding the pricing, adding the authentication to the Webb, just creating the full stack web app or mobile app using this lovable bool dot Nu with I automation tools, I agents like nitenm.com or like that. And follow my profile, and I will try to create these particular codes, amazing codes in future. When you gain the expertise in the building agents, you can take the consultancy service for the specific businesses, small businesses, and just to show them how you can help them, you can take the consultancy or you can create a new agency in which you will sell your services to the particular client. Okay, you can make the money or as a freelancer, you can expand your team by hiring the different freelancers to done the work for you and you can make money through agency model as well. Now, Everyone is becoming the general AI agent builder. But the thing is here, you need to become the specialist in specific niche in the creation of AI agents. For example, if you gain the expertise in creating the AI agents for the customer support, you need to create the gig, like I am the expert of customer support AI agents like that, you need to go with the specific one. Okay. Don't just become the general that you can create the agents for everyone. You need to solve the real problems in that particular businesses. Show them your results instead of benefits to the clients. So for that you need to learn the communication skills. You can go in the YouTube, just learn what are the strategies to close a deal with the client. You can create the course. You can coach in the online. I hope you understand these points. 17. What's Next!: We have learned how to create the five different type of A agents for the customer support like chat based EI agent, telegram based AI agent, on form submission AI agent, email based E agent, which contains the priority system. And lastly, we have seen how to create the feedback AI session for the customer support in the previous sessions, and we have discussed what are the different ways of creating the workflows using AI models also we have seen what are the different ways of making money online with this AI skill. Know if you've got any issue in creating the AI agents in the Enten, please check it out the online tutorials or NTN dogs for more information and don't forget to follow my profile for more advanced courses like this, I will try to create more amazing course for you to get help in this AI technology and you can move forward in one life. I will try to create an amazing course like creating the Android apps, mobile apps, webaps by using these cutting edge technologies in the future. You find this course valuable, give you feedback and review, which helps me to know whether this course is valuable for you or not. I will try to create the event more advanced course on AIS. Always keep stay updated with this AIS technology because every day the EI field is emerging and something new things are evolving day by day, you need to stay updated with this AIS technology right now. You can follow some YouTube tutorials influences or the newsletter, you can join different forums to get stay updated with this AI. Market to adapt the new things very fastly and you can grow in your life as soon as possible. Always keep learning, keep moving and keep growing in your life. I will meet with you in the next amazing course till then, good luck. Thanks for joining.