n8n Tutorial - Zero to Hero: The Basics (Beginner’s Guide to Automation)

What's going on, guys? Michael Elliot here. So today we are going to be running through NAN0 to Hero Guide. This is part one of maybe three or four further videos that I'm going to take you through. By the end of it, you're going to be able to fully deploy a saleable flow to a client. That's really important. Today's video is about the basics and understanding UI pricing, self hosting of nan. All of these different things that might have confused you up until now. I'm going to cover them all in detail, but I'm going to focus on the 10, 15, 20% that actually matter and I'm going to leave everything that does, everything else that doesn't, so that you can just laser in on where your focus should be. We'll build the two or three basic flows together and that will give you a core understanding and confidence before we go into the next stage, which is further deployment of builds. So let's jump into it. So the, the first thing I need you to do is head over to NAN IO, right? We're going to sign up, we're going to create an account together, literally at the same time as one another. I'm going to get rid of these cookies. There is nothing on this page that you need to know other than the fact that you can plug into over 500 integrations, which isn't actually a lot in comparison to things like make.com but you can plug into them. There's only one thing that I really want to show you which is really important. Well, there's two things I really want to show you and I'm going to show them quickly, but they are important. So from a pricing perspective you have two options when it comes to nan. You can either have the cloud based version, which is by far and away the easiest route, avoids any issues later. I strongly advise that you stick with cloud based hosting of NAN. They charge you 20 pound a month if billed annually and 24 pound a month if build month monthly. Although you get a 14 day free trial and if it's self hosted and you start the free trial here, you can, you can go through this. But there are cheaper ways to be able to to self host nan. And I'm not going to go into it in this video because this is a beginner's guide. If anybody wants a video in the comments, let me know about getting self hosting set up. Self hosting effectively allows you to run as many executions as you want for next to nothing like the cost of the server hosting. So it's It's a cheaper option, but this is by far and away simpler. When you're self hosted you lose loads of individual features like the AI feature. You lose the ability to like just click a button for credentials. I strongly advise you stick with self hosted. That's what we're going to do today. So we've covered pricing. Now the only other thing I want to cover is product templates. If you hit to products in the top left and go to templates, you'll see that there are 5,138 workflows. Right. From a template perspective, a fantastic person to start with when it comes down to using templates is a guy called Lucas Pyrin. He's a French guy. I personally know him. He's, he's, he's sound, he's very intelligent and he has some really good templates. Okay, okay. So if I were you, once we finished building together, importing templates off the bat is probably not the best idea. We need to build something together first so you understand the logic. But then when you're looking at more complex flows, you can import them from here and your brain will already understand how things are working and you can put things into action. The final thing I'll show you on this is that you can search by app. So one thing that's critical is I'm going to go through some generic stuff today and we're actually going to build together. So we're going to build live. But the problem is, is that when it comes to deployment for you as an individual, you're going to have your own tech stack, you're going to have the own things you use. Some of you are going to use Air Table, the others are going to use, you know, Google Sheets, some are going to use Telegram, some are going to use, I don't know, whatever your tech stack is. HubSpot, Zoho, you're going to have them. Well, you can stack them on top of, on, on top of one another and you can find, you see there's no none that use Google Sheets for Zoho but there are four Zoho automations, qualify, meta ad leads and what's with WhatsApp, verify verification, Gemini and Zoho CRM. So there are opportunities for you to find templates that already match your, your, your, your space and where you're at now. And you also have your, you have the ability to be able to, to go through your individual steps and follow them. What we're going to do though is we're going to go Right up here in the right corner and hit get started and we're going to fill this information in. Michael Elliott company email. Michael, what email am I using? Michael, hi. Atomo. AI Michael. This is not my main email by the way. This is a cold email. Add your password, keep it nice and safe, don't let anybody see. And here we're going to call it Michael Elliot Test. Nan. Yeah you should call this your name. Call it something related to your company or whatever. Let's start the free trial. Going to hit save. Then the type form is going to pop up and it's going to ask you a few questions. I'm going to very quickly skip through this so you don't have to worry. What team are you on Marketing? What best describes your company software as a service? Which tools have you used before? Make.com maybe some of you have used that. A. Okay, where did you Hear about us? LinkedIn. Submit Invite Team members if you've got any team members at their email here but I'm going to skip it. Your workplace is starting up. They actually have quite a cute Quick Start tutorial. But I'm going to run you through what you need to know from a basic perspective and there's no need for you to. To watch. To watch that. Although you could watch it later, you could save it. That's how quickly you can start an instance. You hit start automating. Now we have an NAN fly. An NAN platform self hosted in. Sorry, an NAN instance cloud hosted instance from nan. So that's how easy it is to sign up and use nan. We're now going to get into the intricacies of the platform and how we're actually going to use it. Then going on to a build. But I'm going to give you a quick overview of the AI in real life. I never use this tab so let me just get rid of the upgrade thing for now to clean it up. In reality I never use this tab, this overview tab. I just. I really, I just really don't like looking at failure rates and stuff like that. I get alerts anyway. I know if there's a problem so I head over to personal. This is where I live, right? Personal. Inside of here there's three options. There are workflows which are if you've used mate.com there's scenarios if you haven't this there are. Each individual workflow is effective effectively your own, your brand new baby that you're building your workflow out of. They can speak to one another, you can connect them in different ways moving forward when you get more technical. But a workflow is just effectively where we piece together our nodes that allow us to be able to build out things which save us time and save businesses time. That's where we are from a workforce point of view. Credentials credentials now credentials are where you can go and add your credentials in terms of like if you wanted to connect your Gmail right now, you could add your first credential and add Gmail as a credential. If you wanted to add HubSpot you could add HubSpot but you can do this live inside of the workflows themselves, which is how I do it. I prefer to do it that way. And I strongly advise you do because it just, I don't know, it makes life a lot easier. The there are exceptions to this rule but that we don't need to complicate things. Now the final thing in here I'll show you is executions. Obviously this is a brand new NAN instance so we don't have any, we don't have any executions that have already happened but we will do later and I will show you. You can see templates down at the bottom. You could click admin panel. Well let me open in a new tab so we'll have to go to it. But inside of the admin panel this is your dashboard where you can upgrade your plan. Not only that, you can manage it and what's important, the only thing that's important really is obviously you have your billing but you have what nam version you are on, what time zone you are on. And you know a couple of, a couple of options here like when this, when this has it verified community nodes isn't is an important feature for later. But forgetting about that one and looking at these two time zone and latest version you can see we're on the latest stable version and that's the one I strongly advise that you. That you. That you remain on. So if we. Sorry, if we go back to dashboard and click open instance it'll bring us back to where we need to be. We head to personal and there's a project section. If you had a keen eye during the pricing stage, you'll see we can only have one project on the basic plan of 20 pound 24amonth. But you can add a project in here, my first project and you can hit save. A good way to do this would be to use your project as your business, as your name things for you, everything, every flow that you build for you. You can build it inside of your project but you don't need a project inside of personal, which is where we need to be. Go start from scratch. And this is the canvas. So our canvas is. Is. Is here. What you see in front of you is. It's a little bit common sense because of what it says. It's add first step. Okay, so adding a first step is obviously what we, what we need to do. But before we head into that, I'm going to show you a few features. The first one being because I've now moved the note around and made some different changes, you can see that this little button here turns orange. If this little button here is orange, it means that you haven't saved. So command S. If you click on the canvas command S it will save. Now you need to get into a habit after every single change to save. Okay? Because this little button next to it is a version history and you can have a look no matter how deep you go into your. So let's just do something here. Let me add a, a trigger. Don't worry about following along with this, but this is just for use for me to show you something. I'm going to add three random nodes I have to install that. Don't want to do that. I'm going to add three random, completely random nodes nose that we're probably not going to use today. We're not going to use them, but I'm just going to have three random nodes. I'm going to make them look all ugly. Look at that. Look how ugly they look. I'm going to command S and save it. Oh no. I've just added some nodes that I didn't need and I'm not. I've now lost my, my, my flow. Well, no, we, we haven't because we can look at the version history. But before that, see this little brush at the bottom? If you hit that brush, it will tidy up your flow. So it doesn't matter where you drag them, it will tidy up your flow and it will centralize them all. Something which is important when you're dealing with longer flows because it keeps it nice and tidy. The only thing is sometimes when you have loops and stuff and this has got more complex things, but when you had had certain nodes in there or certain structures in there, it can actually leave you in a position where it makes it uglier. So just be careful before you press it because you can always command Z it like I've just done there and go back. So don't be scared. But from a version history point of view, we look at the right. See so I've just made the same mistake. Command S. We want to save it. Now we've done it. I need to practice what I preach. We go to the work version history and if you look, you see the version history. So we can see one as it stands right now, today, we can see. Well, as of right now, this second. We can see one, you know, a minute before it, or whatever that was, whatever that is, and we can see one a minute before that. This was an empty blank canvas. I don't know why this. I can't see anything inside of here. But what we're going to do is we're going to restore to this version, restore back to that version, close the version history, and we're back to square one. And that is such a key and core function because you don't have to be worried if you get something wrong when you add something new to a flow because you have a version history, it will stay for a certain amount of time. You're in a position where the version history will be available to you when you're live editing, so you don't have to worry about making a mistake. Okay, now we've covered all of the. All of the ui. There's. There's a couple of things over here. You've got a focus panel which opens up which, like you can ask an. Sorry, there's a. And there's an AI assistant. Let's close that focus panel. There's an AI assistant that can ask you to do basic things. How can I add a trigger node? It's actually quite useful for beginners to use. It will help you. This will come back. To add a trigger node in any end, click the plus button in your workflow editor. In the node search bar, type trigger. See available trigger nodes. Select the trigger node. So you can see why AI might be of an assistance to you. And the final thing is this button here is a sticky note and that allows you to be able to document what you do along the way, which is quite important. You'll have seen them in a few of my. In a few of my videos, but we're not going to do that for now. So command S. We've got a blank canvas. Let's get ready to upgrade now and let's start building something. Okay, so we're going to build from scratch now. We're going to add a first step and what we're going to do is we are going to have this. Let's start with a simple AI agent because you'll learn quite a bit out of this from an AI agent? Actually, no. We're going to start with an on form submission. So click the node in the middle, click on form submission and just escape back to Canvas. We haven't set up this form submission yet, but we have, we know what it is. And a really good habit to get into is when you escape out of a node, press the space bar and when you press the spacebar, you can read it. Sorry, you can rename it form submission. To clear things up, you should do this for every single node that you add, every single node that's added, you need to rename because it, you want, it might want to add specific things in a complex flow explaining what it does. It's really important you get into these good habits now. Really important. So you'll see now this little, this little icon that we can drag, like you can drag it around. Well, that little plus icon is how you add another node. So we, if you press that plus icon, you're going to have a list that comes up on the right hand side. Okay. And these are all the nodes that you can, that you can add. AI action in a loop, data transformation flow, core, human in loop, add another trigger. So what we do here is we want to go through action in app nodes, what they are. We've added a form submission as a, as a trigger. But I've got something prepared that I want to showcase to you before we add any more nodes. We've learned how to add a node now. Okay, so we know how to add a node. We've just done it together. We've clicked on that we've selected on form submission. Right? But there are many nodes to choose from. As you've seen, they've got over 500 native connections. But in here I have an overview of what flows are to explain to you quickly now. And I'm going to run through them quite quickly because I want to get us building. But there are three different types of nodes, effectively trigger nodes, action nodes and logic nodes. The trigger node, you can see the form submission we just add a trigger is something that fires off a. You know, it fires off the execution. So on the fort, when that form is submitted, the workflow will fire because it has a trigger. We can have a schedule so we can say to run once a day, once an hour, once every three hours, or once every, once every week, whatever it is, once every, every Thursday, whatever the trigger is, we can schedule a trigger. Those are, those are triggers. Probably the most, two most powerful ones are when execute executed by another workflow which we will get into in subsequent videos to this. And Web hooks, which we will get into in subsequent videos to this. But these are trigger nodes. These are things that when something happens inside of these nodes, something happens and that is what starts the. The flow. Action nodes, on the other hand, are individual tools. So when I say tools, I mean Google Sheets being a tool, you know, Perplexity being a tool, Air Table being a tool, Google Calendar being a tool. All of these things, they're action nodes inside of action nodes. They can do many things. They can create doc documents like this is. This is what I've selected for this one. But if I was to delete this and add another node, Oops, not AI Add Action in app. If I just search for sheets, Google Sheets, I can have it do any of these things. Append an update arrow, which is something we're going to use in a moment. And you can see why they're so powerful because they have, you have many actions that are available to us from a, from an AI. From an NAN perspective, logic nodes are where the true power of NAN comes in. Because logic nodes allow you to be able to really do some key and core things. One of my favorite ones is the split out node and the aggregate node. Group all individual items into an array. I'm not going to go into detail about what they are because this is a beginner course, but you can split, split one item, one, one array out one object out into many, or you can do the opposite with an aggregate node. Multiple reasons why you would do either, not really for this, but the Swiss army knife of NAN are HTTP requests, where we use an HTTP request to make a request to a API that allows us to be able to manipulate data that is not inside of NAN when it doesn't, especially when it doesn't have a native integration. So if you think about that as an opportunity effectively, that's the Swiss army knife that allows you to connect to everything on the Internet. That's the way that I would explain it. Okay, now the code node is the Goat node, right? This is this. The power of this bad boy is something that you, you as you begin and as you begin your journey in AI, at any end usage. This code node is a disgusting animal, right? It is so powerful that it will blow you away. Something that could take seven nodes. You can do it with a code node. We're not going to get into this until later on in this playlist and, and flow. But this is the Goat node, I have no doubt about it. Inside of this is JavaScript. But the beautiful thing is that LLMs now like Chat, GPT and OpenAI, Chat, ChatGPT, AI Studio, etc are so powerful from a simple prompt and even a screenshot of your flow, for example, you could have it write the correct JSON and it will just do everything for you. It is spectacular what the code node can do. I don't want to bang on about it anymore, but it is, it is incredible. Now the, the final piece that I want to cover is an AI agent. Now, AI agents are a little bit of a, a little bit of a divided territory. You'll, you'll have seen probably if you're watching this video, AI agent builds or AI agent flows where 7,500 million AI agents can develop your can, can manage your day well, they can, but they're very, very, very, very likely to break the consistency. Isn't, isn't where it needs to be. We're going to use an AI agent today in our second flow. There are use cases where I use them. For example, they can accept binary files, whereas other nodes like the basic LLM chain can't accept a binary file. All of this stuff, you don't need to. By binary file, I mean a file binary is a file in nan. So I don't want to bore you with that stuff, but AI agents can do some really cool stuff and in your position now, you. We're going to use them to be able to do it right. So the boring stuff is now out of the way and we're gonna start with a live build. So we're gonna head up here to the top left hand corner where you can see this. It's a horrible track, that's a horrible arrow. And we're going to hit it and we're going to rename it our first workflow. Okay, so now we've renamed it. The workflow is, is, is named properly and appropriately. It's been saved. If we go to the overview page or the personal page, which is where I live, you'll see that our workflow's here, it's saved. If we added another workflow and saved it and called it our second workflow. Our second workflow and saved it and head back to personal. You can see how they go together. Now what, what I need to do here though is from. I need to create a folder for organization. So YouTube series nan beginner guide create. Now we have a folder and I'm simply going to drag and drop both of those in there. And now I'm organized because if anybody, if I create overflows inside of here. In your instance, if you create overflows inside of there, you can add more folders, you can put them in for the right thing so you know where they are. So now we know we're in our first workflow and here we are, we're going to add a trigger. So we're going to click this big button and we're going to hit on form submission. It generates work forms in NAN and passes their response to the workflow. And inside of here we're going to give it a title. And we're also going to do something pretty cool, which I don't see in a lot of N beginners, but it's going to teach you how to leverage AI. Okay, so what I'm going to do is I'm going to call the form title New inquiry form and the description is going to be a form for new inquiries to submit. And then right here you're going to see add a form element. Don't worry about any of the other options for now. Add in a form element and here we get to add the fields to the form, simple as that. So the field name for the first one we're going to want a name. Oops. We're going to want a name and the element type is going to be text, which means they text, they type it in. If you look at the options, there are a lot. There are individual things, there are checkboxes, there are drop downs which we'll use, radio buttons, passwords, there are all sorts of different things. For this obviously simplicity, we're going to stick with text name. Now, placeholder text is what is shown to the user on the form. So what we want to put is please enter your name. Ticket is a required field, which means they can't submit the form without it and then add another form element. And this, this, this field name is going to be email. Now we're going to leave it as text. Even though there is an email option. Just, it doesn't matter. It keeps it as a string. It's easy, it doesn't matter. And then finally, so we're going to hit add form element one more time. So we're going to have three elements and we're going to call this service. And when we're mimicking a, a businesses flow here of like a form submission, that service is not going to be text, it's actually going to be a drop down. So we click element type, we choose drop down. And option one, I'm going to put ppc. I want to add Another field option as AI. Okay, so the inquiry can either be about PPC or it can be about AI. And then we will add one final field element and we'll call this Description. Description. Okay, and the placeholder. Please tell us about your inquiry. So we have name, email service, which is a drop down and a description. Okay, so we're gonna escape out of that command S to save space, bar to rename enquiry form. Rename. We now have our inquiry form fully built out, ready to rock and roll. All we need to do now to see this in action is hit execute workflow. So we execute the workflow and this is the standard NAN form. You'll see that it tells you when you're in test mode, which is important. And we're going to do this. Michael Elliot, let's just give Michael michael@aitrain.co.uk well, actually let's use. Yeah, we're going to select AI as the service and the description. I'm looking for an AI, a solution to my client onboarding system. Currently it's all manual and time consuming and we submit. The form's been submitted. You can see that around the trigger node of the form it's turned green and there's one item in the output which you can see there. So that means our data is now available to us. So in the right hand side, see, you can pin this output instead of waiting for a test event. So by clicking on it on the node and pressing the letter P, you can pin it. You see how that little icon there, the little pin that shows us that we, we have pinned the, we've pinned the data now. So what that means is every time we're building the flow moving forward and we want to execute, we don't need to go and resubmit a form. So, and that is really, really, really important because when you pin data in a flow, it avoids AI having to use credits into. And I'm trying to keep you in some good habits here, although I am going to unpin it in a minute to show you something. We don't really care about those. We care about these four. We care about these, these four here. These are what came from our form. So we have a fully functioning form now that we could send the URL to a user to submit and we would be able to receive this information. So we've got a fully functioning form. We now just have to build out the remainder of the steps in terms of what we're trying to get this to do and what I want you to do. Here is add a Google Sheets node by clicking the little square, the sort of smaller square, less obnoxious button there. And I want you to go search for Google Sheets. Now, I always search for nodes. You can find the nodes by clicking Action in App and finding it. But I just find that searching for whatever tool I'm using. Your tech stack is going to be made up of like five things, like, you know, seven things maximum. And this is why everyone shouldn't be scared of AI. Because you're in a position here where you, you know, you. You don't need to use every single tool in the box. You just need to use what's right for you or for the individual client that you're using. And it, when you look at it like that, it becomes a lot less daunting when you, when you're trying to learn like this. But we want to append or update Arrow in Google Sheets, okay? Because that means that we're editing a sheet that already exists and we're appending or updating a row inside of it. So we want to, we want to do this. Now we've got the data on the left, and something I haven't yet covered is the three different views. Okay? The three different views that you've got. You'll notice that this is different to what we saw inside of the form a moment ago. This is called schema. And schema allows us to be able to dynamically use data that we've already got from previous nodes in future nodes. Stay with me. It's very easy. This is what we were looking at a minute ago, the table. This is the data as it exists. Very simple. Now we have the schema, which has them all separated out into individual entities. They're individual strings on their own. The key and the value. The key is the name. The value is Michael. That's how JSON basically is structured from top, top to bottom. But this isn't about Jason. Watch my video on JSON if you need. If you want to learn about that effectively. Schema allows us to be able to dynamically input data into specific areas. Okay. There is also a JSON option which allows you to see the JavaScript output and that will show you a similar thing. But this way it's very clear to see the key, the name, and the value, because the value is in. Is it. The key is in brackets, curly brackets. Then there's a semicolon, then there's the value in brackets itself. Those are the three views. Table is much simpler to keep your eye on when you're looking at information and you want to see it. JSON is an extremely important format you'll go on later to be extremely comfortable with. And schema is where we're going to use to be able to dynamically drag and drop information. But we have a core problem here that doesn't exist. We're going to rename this row. Update slash, append row. We're going to rename it, we're going to command s it. And what we're going to do is we're going to go inside of this sheet and look here, create a credential to connect with. Well, we have to connect a credential. At the moment it just, it's just an empty node. It's just. And that's what they're called, they're all called nodes, these little squares. This is node one, this is node two, they're all called nodes. And in this one is broken. And it's broken because of multiple reasons. Doesn't have a document or a sheet, but it doesn't have that because we don't have the credentials. So we need to create a new credential and all you've got to do. Sorry, all you've got to do is hit select credential, create new credential and hit sign in with Google and then select the email address that you're using. Continue. Oh, go back, sorry. Got to take these. Select them all. Apologies for that. Tick them all Connection successful. Accounts created and connected. Absolutely fantastic. Now I have a document that I've already built out that I need to add the description to. Feel free just to pause the video and do this the way that you would do it, though you can probably do it whilst I'm talking, is go to sheets new hit enter. It will open you a Google sheet up and then you would literally just copy these four individual things over. Bang, you've got it. Name, email, service, description, that's what you. What you need. So I'm slowing down intentionally here in case you. In case you need to need to go and do that whilst you're doing it. But the fact is, is that now one thing that we've been through is creating credentials. Creating credentials is something that you're going to do consistently. Doesn't matter what sort of credential it is that you're trying to connect with. You can connect with it and often you can just sign in using Google. Sometimes it gets a little bit more complex, but we don't need to go into that now. It's just based around the basics so inside of this now we have options, right? Append or update a row. You can see in here you can actually change to the other. The other options that were available get rows, update rows, create a row. Sorry, create a new sheet. Clearer, clear, delete contents or part of a sheet. But append or update a row is the most common one that you're going to use. So we're going to use it today. And then there are a few different options. We can either select the document from list by URL or by ID for simplicity. Today we're going to select from the list and we're going to select. My sheet is called our first build. So I'm going to select our first build there. Then the sheet underneath, which is a little bit confusing, to be honest with you. The sheet underneath is the name of the sheet because there's multiple sheets, can be inside of one, obviously one, one. Google sheet. This is called sheet one. And as a standard practice, they're all called sheet one. So we can select from the drop down, which is the easiest way and I advise a beginner to do so. But you can also select by URL ID and, and so on and so forth. But I'm going to stick to the basics in this video. We'll do all of this later on. Now, one thing that is important is selecting a column to match on. And a column to match on is the column that is going to be effectively unique for each individual user. So what you don't want to match on is the service, for example, because the service can be the same for multiple people, whereas an email or a name can be the same. Whereas an email is. Nobody has the same email, nobody on the planet has the same email. They, they, it just doesn't happen. So we want to select the column to match on and we want it to be email, right? Now what we're going to do is we're gonna add these variables and this is where the schema section comes in. Be careful here because you might be on table, you might be on JSON, but you need to hit the schema section and drag and drop the individual options into the right location. So we've got name, we've got service, you literally just click the box and drag it. Service description. And then for the email, I'm just going to show you how you would manually do it if you press the equal sign here. Okay. Because we can see three examples below. Very simple, right? Very simple basic variables. Now if we've pressed equals, so this is now an expression. You'll See the fixed and expression fixed, you drag and drop in. So we drag the email in. Fixed, boom. JSON email. Or what we can do is we can do it using an expression, two curly brackets, JSON.email. literally. Let me run through that one more time for you. I'm going to delete it. You add two curly brackets, shift and where the square brackets are, you select JSON, you put a period and then you select email. You can see it's brought through everything. Description, email, form, name, service. If we select email and it turns green. That's correct. Right. And we're going to append this sheet with this information. Congratulations, you've just written your first piece of code during this individual process. So we've not only dragged the variables over, the parameters over, we have actually also just manually coded in something, which is quite a big feat. You're already ahead of 99% of people by being able to do that. Right. So I just wanted to add that in this video. If you can't get it to work for whatever reason, just drag and drop the email over. It's absolutely fine, don't worry. But, but I think that was quite simple. So now when we execute step, this should update our sheet. We execute step. The node's doing its, doing its thing. The outputs. Now this now only contains. If you look at the differences submitted out and form mode are gone because we didn't add them to the, to the values to send. We've got these and if we go to our sheet, we have a full system. Okay, so we now have a Google sheet which is listening to our flow. We now have our first functioning connection of nodes, the grouping together of nodes that allows N to be able to communicate with other third party platforms via their credentials, which is what we've just done. So congratulations. Right, we're making good headway. But there's something in here. So we want to pin this. There's something so, so click it. Click into it. This little button in the top corner is your friend. But I much prefer to click the node and click pin and it will press P. Sorry, press the letter P and it will pin the data. But you can use that if you, if you must pin the data. We've now got a sheet which updates. Well, let's roll back a minute. What have we got? We've got an inquiry form which has a name and email, but it has a service. But there are two types of services. Well, let's for theory's sake say we have a PPC department and an AI department. Okay. And we're trying to route inquiries based off of what they put in their form. How do we do that? Well, how we do that is by using an if or, well, using a switch node. We'll use a switch here because switch if ifs are really common. We can, you can do it both ways, but we'll use a switch. So what we want to do here is we want to use the switch node with the schema and we want to put it as Service is equal to AI. Okay. And then add another routing rule. Drag the service again. Or we can be cute. Let's be cute guys. Let's do that. Double curly, oops, equal sign. Or click this button here for expression double curly brackets, JSON, period, service green variables pulling through. You can see that by hovering over it. Result AI. When we submitted the form, it was AI and is equal to ppc. So we now can come out of this. Well, we can now execute the step. All right, and what, and what you'll see here is this. We're going to come out, we're going to command S, we're going to save. We now have a switch node. PPC or AI. PPC or AI. Rename. You see, one item got outputted there. Ignore the zeros. That's just the way that they, they do it. You could add a million. Look, add different routing rules. If you add them, it will add more routing rules to it, starting at zero to and going increasing from there. Of course, we only want to keep those two. But this has gone down the this zero route, which is the first one. It can because it's equal to AI. If I drag and drop them and execute the step again, you'll see it went to the bottom one. So you can understand now what is happening. If it chooses AI, it's going to go down the correct route. Whichever way we have put it together. If it's ppc, it's going to go to the other section. Now let's test it out because if we go into the JSON inside of of here and we click this little pencil, we can edit the, the input, the output. Sorry. And we're going to change, we're going to inside of those two curly brackets. You don't necessarily have to do this if you don't want to break your flow. But I'm going to change it to PPC and I'm going to save it. Data still pinned. The service is ppc. Now you see the output when it was AI. If we execute this step again, you'll see it output the item to the PPC root. So I hope that's clear. And that is the power of using logic nodes because you can move different things to different areas. We're using PPC and AI and the theory that there's two different departments. Well we want to do two different things, don't we? Based off of, based off of, of off of that. Now what we're going to do is we're going to add a node here. So sorry, we're going to pin this data by clicking it and pressing the letter P and we're going to add Gmail and we're going to send a message now for the sake of this video. So credential again. Gmail needs a new credential. So we, we've done this before guys, you know, you're comfortable now. You hit credential, you hit sign in with Google, you choose the email address that you're using, you give it full permissions by clicking select all and click Continue and you're connected. It's that easy that this is why you use the cloud by the way because you can't do this when self hosted. So this is why you use the cloud when you, when you're beginning. So we now have a Gmail account. So who do we want to send it to? Well I'm gonna have it send it to my PPC email which is not related to this and I'm gonna have it say new PPC lead. Hi Michael, there's a new PPC lead. Please see below. Please see below. I'm just shrugging. So, so what I'm doing, in fact I'll do this slower because you're going to be following along. So now I've added this, what we're going to do is name JSON.name name JSON name Service drag service over description. Pull this over users email address. Drag the email over. So now we send these emails. It's going to come to michaeladtrain.co.uk. let's execute that step. You'll need to put your, your own individual email in here obviously. Don't send me a bunch of emails please. You don't need to, you don't need to do that. But let me just flick the camera a moment and let me load up my inbox and let me show you what I've received. So this is the email that came through. New PPC lead in the subject. Hi Michael, there's a new PPC lead. You can see the description, you can see the service, the format's off which is, which is fine. Don't worry about that for now. But you can fix the format and that it's sent in, but we're not going to worry about it. This is about building your first flow. One thing we will address though, is this little pesky devil. So this email was automatically sent using nan. Okay, so let's get rid of that now for now anyway. And let's add an option and you'll see append NAN attribution. Just turn that off. And now it won't add that. Set that little line to the email. It's important. So we want to command save. You'll see here how you could do some really cute things if you just think I know that you. That this is a beginner's video, but they have IDs and thread IDs so that you could then go and, and do that and you could, you could update it, but that's not, you know, like you could use the thread id to create an AI that goes backwards and forwards, etc. I'm just trying to show you. So we have that send a message. Well, the beautiful thing is, right, is that we can pin that data now, but we have the exact same thing from an AI perspective. Okay, but we just need to update a couple of things. So we're not going to add another Gmail node. We're going to rename this one PPC lead. Email sent and we're going to rename this one. Sorry. We're going to Command D or Control D which duplicates it. We're going to line them up. You see this little connection? We're going to drag it and we're going to connect it to that node. Now what we have here is a PPC lead. Let's rename this AI lead. Email sent. Rename. Don't worry about the one item. That's just the output because this hasn't been run yet. And the data's. The data's in here. Michael. So two. Let's just say it. Let's just use Michael. What's this email? I'm using Michael at High Automo. AI michael@high automoai.co.uk was it michael atomoai.co.uk.com michael ohio.com New AI AI lead. Change this to new AI lead. Please see below. Escape command S. We've. Sorry, we've already renamed it. Now what we're going to do is we're going to unpin all this data which going to check the routing JSON is equal to PPC is the first one we're going to unpin this, this, this and this. We're going to command s it, okay? And we are going to execute the workflow again. Michael let's just call it David so you can see it in the sheet. David michael@high automoai.com and this time we want to select. Did we select PPC as a service last time? I think we did. No. I don't know. I can't remember which one we did. I think we did that, but this will show you it in action anyway. I'm shrugged. No, no, we did use AI. Let's use PPC just for an example. I need PPC support. My account is not performing. All right, let's just say this is the inquiry now. We're going to submit the form, it's going to flow through, it's going to update the row in sheets. Invalid email address Ah, Michael, you silly devil. Didn't save it. Michael AI.com Command S execute the step. See, so it did go through. I'll show you in Gmail that that went through as a PPC lead and we'll execute it again. Do you know what, we're just going to execute it again and we're going to do another AI lead. Michael Mike Dr. Elliot gmail.com service AI description I need AI support. Submit email sent so you can see how that, how that flow is now fully functioning. Congratulations, you have, you have a flow new AI lead. There is a new, there is a new AI lead. You'll notice that there is no appending from nan inside of here. So we have that and if we look inside of the sheet we'll see that we have three different inquiries that have come through for individual reasons and we're routing them based off of this. So command s this because we're gonna, we're gonna use this, we're gonna use this to continue to, to develop the. The flow. Now I'm gonna take you that we're going to use. So I'm going to head back to my folder that I created earlier and go into our second workflow and it's time to hit, hit it again. Let's close this form down. But we're going to use something different this time, something completely different so that it's functionable. So here we're going to have a on chat submission. So what we've done is we've come back in. I don't want to rush ahead. We've come back in, we're gonna, it's. I've already renamed it my second Workflow. If you don't have the workflow already created, if you don't have it already created, just hit create a workflow to create a new one. But the second workflow is going to be an AI agent. And this is going to tap into a few things that are completely different to what we've just done. Now we're going to have this. When we add a trigger. There are lots of different triggers, of course. Course. But we're going to use on chat message, so you press the node, you hit on chat message and hit escape to come out. Now what we're going to use here, we're going to rename space. Remember, always remember to rename your names. You rename your notes. This will really save you. Not right now it won't, but later, trust me. The habit of renaming your nodes takes a second, two seconds, and it will save your life later on. Command S will also save your life later on. And what we're going to do is we're going to add an AI agent. So we click the little square here, we hit this little square and we search for an AI agent, and we've got him. And there he is. What? What? A little robot. Right. And this time we're going to call him the legendary robot. Okay, Right. So he's a legendary robot. You don't have to rename him that. He's an AI assistant. Okay. For the use case of this video, I better rename him AI assistant. Right, so we've got the AI assistant. Well, now we need to connect a chat model. And in order to connect to chat model, we need to have set create up. Set up an account and create a new. A new credential. So with this credential, this is going to be something where we have to connect it and I'm going to use Gemini. So Gemini. Search for Google Gemini. Well, search, sorry, search for Google AI Studio. And we need to go to settings. Oh, sorry. Need to get an API key. Create. So we've gone down here to get an API key. Create an API key. Oops, I just paused that. Hit create an API key. Didn't think that was centered. Copy this API key. Keep this safe. Don't give this to anyone. I will delete this key the minute this video is finished. But you copy this API key and you go inside of here, you paste the API key in and you hit save and you have now connected your first API. As simple as that. Okay. Literally as simple as that. Now you have your eight Gemini Flash available. There may be a cost to this to run this agent of $5. I'm not sure if you can use the API for free, to be honest with you, because I don't use anything for free. For free and forever. But we may have to. You may have to sign up for this for the $5 of usage. Chat GPT I think definitely does as that. But sticking on task. The chat model is the brain. Okay? So this chat model is the brain. That's. That's behind everything. Whatever you select here in terms of your model is the key. So this is using 2.5 flash. It's fine for what we want. We could use lesser models if we wanted to. You could because they're cheaper. This is why you would look at this. But you can Also use like 2.5 Pro if you want. When it's in production, you want the, the best output consistently. We'll stick with Flash for now. We need to add a memory to this simple memory and we're going to give it the sim. Basically, the memory is exactly what it, what it is. So I'm going to rename the gemini node gemini 2.5 flash because I'm not following my own advice and what I do normally when I build because I'm presenting memory. So this is Gemini 2.5 flash, our brain. That's the memory. And you know what's cool about AI assistance is that we can connect as many tools as we physically possibly want, whether it be another workflow. This is where I have to tap into the. This is where I have to. Oh. This is where I have to tap into the power of, of, of AI agents. AI agents are extremely powerful. You just. And as a beginner, this is really important to understand because you're going to be able to use them and play around with them. But for production usages, when it comes to client delivery, I hardly ever use them. And if I do use them, they don't have any tools attached. Do not get confused by that right now. All I'm saying to you is that AI agents are freaking cool, right? They are awesome. They're amazing. You're going to see why in a moment. Even if I have to copy this over to my main, my main account because of the Gemini thing. But you're gonna see an incredible workflow based off of what we put together for some basic tools and, and information using AI assistants, they're really powerful. The memory gives it memory to be able to, to showcase it to you. And I'll show you that now. But I just wanted to take a second to explain AI agents are really powerful when it comes down to messing around with stuff like this. So it now has a memory. Okay, there's another key, that hotkey that I need to teach you, and it's the letter L. You press the letter L and the logs show. Logs are important because logs show you the history. When you'll see when we fire this in a moment, what happens. Hi, my name is Michael. You'll see the assistant goes through Gemini Flash. There's no system prompt inside the. The assistant, so. But because it was a basic prompt, it didn't have it for your user must use the formal JSON tool to complete the answer. What is. You can't. But what is my name? Your name is Michael. You just told me that. And I remember. So you can see here that the memory is one that I told it my name was Michael. It then remembered that my name was Michael. When I asked it if I detach this memory and I say, hi, my name is Michael, what's my name? As an AI, I don't have access to personal information. So you can see the power of the memory node. Now I'm going to go back into it. I'm going to set it at a thousand windows memory. Okay? But what's important here is the logs. The logs show you what happened. So the chat message triggered the AI assistant then started running. It then very quickly went to Gemini Flash, which is our brain, and it gave us this output. The input. And the output. The input that. Sorry, the. The output. The input. You can see what we're looking at by toggling these. You can see the outputs and inputs. Inputs are what are on this side, what is coming in, and the output is what's on that side that is going out. Simple as that. We're using chat, so it makes it easier. So this beautiful little button here, that's like a reverse icon, we're going to reset the chat session. Start from scratch. Okay, command S. Now what we're going to do is we're going to add some tools to this. Well, we've got Google Sheets, so let's start with Google Sheets. Let's start with the Google Sheets tool. And what we're going to do is we're going to choose from the list, our first build and sheet one. And then we are going to add a Calendar, Google Calendar tool to this. We're going to create a new credential for Google Calendar exactly as we have done previously. Same account, select all, Continue. Sign in with Google just to make sure. Sometimes it bugs out. I Don't like it. That's how I make sure that it does it. It's connected. Okay, select your calendar, Michael, at high autonomous AI, this will be your calendar. The operation here is create an event. Okay, so we have the ability to now create calendar events. So create calendar events, update a row in Google Sheets. So we have those two. What else could we use? We could use Gmail because we've set that up already. So we're gonna, we're gonna add another tool and we're gonna set Gmail tool. And what we're actually going to do is we're going to have it. We're going to. Hmm. What operation should we do? Get many. Get. Get an email. We're going to use it to get an email because we're going to use something funky. So we're going to use it to get an email. So the operation we've checked, we've added the tool Gmail, right? And inside of it we've select the operation get. Now the message ID here, it's asking us to put the message ID in. Well, we don't have a message ID right now. So this beautiful little button ears lets the model define this parameter and if you tick will get the operation get will get the most recent email. Okay, now what we need to do here is give the AI a bit of a brain. This is why AI agents get really confusing and can break quite easily. Even if you group them into sub workflows, they can break, but that's for another day. I'm going to press L to just remove that for a minute. I'm going to hit align center so you can see it. We've got the chat trigger. We've got Gemini flash memory. Create calendar events, update in a row in Google Sheets. Did I have the right thing? No. Operation, append or update row is what we want. So append, append or update a row. Yeah, so I'm gonna. We need to edit this sheet because it was on the wrong one. So you come inside of the Google sheets. Append or update a row. Our first build sheet, one column to match on email. And then you do the exact same thing with these. You have AI be able to identify them. Let the model define this parameter and I'll show you why in a moment. So we can create calendar events. We can get a message in Gmail and we can do that. But we've done all of this and built all of this out without giving the AI assistant any feedback on how it should operate. And we need to give it a system Message right now it's your helpful assistant. But in order to make it use the correct tool at the right time when information comes in from our message, because we've asked, we've given basics of names and stuff and it can do that. But when we're getting it to use specific data types, we need to have a system prompt be able to assist us with it. So what we're going to do is we're going to go. We're going to. I'm going to take a screenshot of this. Apparently I'm taking a screenshot. I've never seen that before. There you go. Take a screenshot about my flow. I'm going to go to AI Studio. Gonna go. I'm gonna go back to. Oh, we were there. AI Studio. I'm going to turn the temperature down. I have attached an image that is a screenshot of an AI agent that can use Google, that can use Gmail to get an email. Google Calendar to. Let's check what we've got. Calendar just to set create Calendar events. Google Calendar that can create events and Google Sheets that can update a row. This is. I need a system of very simple basic system prompt that assists the AI agent in assigning the correct tool depending on the input from the chat trigger. But depending on the input. Let's keep it there. Depending on the input. Now we run. So this is going to produce us the prompt without us writing it. And this is why I'm using it. And you can see what it's spat back out. You, you're an AI assistant and we've access to the following tools used for creating events, scheduling meetings or setting reminders. Used for updating or add information to a spreadsheet. Gmail used for retrieving information from emails. Okay, so we have all of this information inside of here. We're gonna. We've got. The prompt comes from the source of the chat trigger and we're going to replace this with. Let's change it to an expression just so we can get it out by the way. Just so we can get it into this view. In case you didn't follow that literally just check. It was fixed like this when I pasted it in. I'm going to delete it all. It if it's fixed like this. If you change it to an expression and double click that little icon again, you've got the result here and you've got the expression here. We obviously we can edit the expression. So I want to paste this in. Is it going to allow me to paste or Is it going to just do that to me? Okay, so let's just go and grab it again quickly from here let's copy it. Copy text back to nan paste in. Be careful to pull. Be careful here to pull out the. The. The system prompt and text. You're an AI assistant that has access to the following tools. Google Calendar, Google Sheets and Gmail. Yeah, you see here we are used for retrieving from emails. Analyze the user's search request and use their if the user request is create an event use Google Calm to sheet Updating a sheet or adding a row use Google Sheets. So now we've given it some context and a brain to understand which tools to use and they are all connected. Okay so command S. We now should be able to use this where's the chat gone? Or l the hotkey to bring up the chat for our chat trigger. Now when it comes down to this hi, my name is. My name is use and they are all connected. Okay so command S, we now should be able to use this where's the chat gone? Well l. The hotkey to bring up the chat for our chat trigger. Now when it comes down to this. Hi, my name is. My name is Michael. Please fetch me my most recent email. It's using its brain memory node executed. I can only fetch an email if you provide me with its id. Well the ID inside of your. The email my friend is defined. Let's change this to get many let's let's it should. We should work with get and what we're going to do is we're going to choose from the list our first our first build and sheet one and then we are going to add a calendar under Google Calendar tool. Let's just change that to get medi command S. This is guaranteed to be a problem with to this we're going to create a new credential for Google Calendar exactly as we have done previously. Same account, select all continue sign in with Google just to make sure Sometimes it bugs out. I don't like it. That's why. That's why I make sure that it does it. It's connected. Okay, select your calendar the prompt. Yeah, exactly. So let's say this, please can you use Gmail to fetch me my most recent email. Michael, at high autonomous AI this will be your calendar. The operation here is create an event. Okay so we have the ability to now create calendar events. So create calendar events. Update a row in Google sheets your most recent emails. Hi Michael, there's a new AI lead. Please see below. Name Michael Service AI description. So we have those two. What else could we use? We could use Gmail because we've set that up already. So we're gonna, we're going to add another tool. Okay. So you can see that now you've got a fully functioning workflow here. Let's now give it a challenge. Please, can you. And we're going to set Gmail tool. And what we're actually going to do is we're going to have it. We're going to have it. What operation should we do? Get many Get. Get an email. We're going to use it to get an email because we're going to use something funky. So we're going to use it to get an email. So the operation we've checked, we've added the tool Gmail, right? And inside of it we've select the operation get. Now the message ID here, it's asking us to put the message ID in. Well, we don't have a message id. You update the Google sheet right now. So this beautiful little button here lets the model define this parameter. And if you tick will get the operation get will get the most recent email. Okay. With the following information. Name test, email tester. Now what we need to do here is give the AI a bit of a brain. This is why AI agentsmail.com service get really confusing and can break quite easily. Even if you group them into sub workflows, they can break. But that's for another day. I'm going to press L to just remove that for a minute. I'm going to hit align center so you can see it. We've got the chat trigger, we've got Gemini flash memory. Create calendar events, update in a row in Google Sheets. Did I have the right thing? No. Operation, append or update row is what we want. So append, append or update a row. Yeah, so I'm gonna. We need to edit this sheet because it was on the wrong one. So you come inside of the Google sheets, Append or update a row. Our first build sheet, one column to match on email and then you do the exact same thing with these. You have AI be able to identify them. Let the model define this parameter and I'll show you why in a moment. So we can create calendar events. We can get a message in Gmail and we can do that. But we've done all of this and built all of this out without giving the AI assistant any feedback on how it should operate. And we need to give it a system message. Right now it's your helpful assistant. But in order to make it use the correct tool at the right time when information comes in from our message because we've asked we've given basics of names and stuff and it can do that. But when we're getting it to use specific data types we need to have a system prompts be able to assist us with it. So what we're going to do is we're going to go. We're going to. I'm going to take a screenshot of this apparently I'm taking a screenshot. I've never seen that before. There you go. Take a screenshot of my flow. I'm going to go to AI studio Going to go. I'm going to go back to oh we were there AI Studio. I'm going to turn the test Description Description. This is a test also can you temperature down? I have create a meeting at 6pm today with Michael ad train create me in a 6pm today about the sales. The the the the please create attached an image that is a screenshot of an AI agent that can use Google that can use Gmail to get an email Google Calendar to let's check what we've got Calendar just to set create calendar events Google Calendar that can create events and Google Sheets that can update a row. This is I need a system a very simple basic system prompt that assists the AI agent in assigning the correct tool Depending on the input from the chat trigger depending on the input. Let's keep it there. Depending on the input now we run so this is going to produce us the prompt without us writing it and this is why I'm using it and you can see what it's spat back out you you're an AI assistant and we've access to the following tools used for creating events should you a meeting today at 6pm about last week's sales and fetch my most recent email. But what I'm going to do here is so you can see what I've done. I'm going to command C that I'm going to copy it. I'm going to reset chat session so it starts afresh and I'm going to paste it in now we should see all three tools work and the power of of an AI assistant. You should see it. It's using Gemini. It's gone back through it's now done the calendar it's now fetched the email and the output is done. These creating meetings are setting reminders used for updating or add information to a spreadsheet Gmail used for retrieving information from emails okay, so we have all of this information inside of here. We're gonna. We've got. The prompt comes from the source of the chat trigger. I made an email for at 6pm I've created you. We're going to replace this with. Let's change it to an expression just so we can get it out by the way. Just so we can get it into this view. In case you didn't follow that literally just check it was fixed like this when I pasted it in. I'm going to delete it all. If you. It's fixed like this. If you change it to an expression and double click that little icon again, you've got the result here and you've got the expression here. We obviously we can edit the expression. So I want to paste this in. Is it going to allow me to paste or is it going to just do that to me? Okay, so let's just go and grab it again quickly from. From here. Let's copy it. Copy text back to nan paste in. Be careful to pull. Be careful here to pull out the. The an email. Let me drag this up. What am I doing? I have created a meeting for you at 6pm today about last week's sales. Here is your most recent email from michaelmoai the same one that we've got as before. So guys, the system prompt and text. You're an AI assistant that has access to the following tools. Google Calendar well done so far. You've got two flows built. You can now book me Google sheets ins. You can now update rows, you can send emails. But all of you with a keen eye will have spot and Gmail. Yeah you see here we are used for retrieving from emails analyzed something when it comes down to the individual calendar booking as the user's search request and use their. If the user request is creating event use Calendar booking is happening like an hour from now at any time. So whenever you sent that message it would fire off and it would do it there. And the reason for that Google Calm two one sheet Updating a sheet or adding a row Use Google Sheet. Yeah, AI agents like to hallucinate. So that's one piece. So now we've given it some context and a brain to understand which tools to use. But the other thing is is that if you look at the prompt that we've actually put inside of views and they are all connected. Okay so command S. We now should be able to use this where's the chat gone? Or L the hotkey to bring up the chat for our chat trigger now when it comes down to this. Hi, my name is. My name is Michael. Please fetch me my most recent email. It's using its brain memory node executed I can only fetch an email if you provide me with its id. Well the ID inside of your. The email my friend is defined. Let's change this to get many let's. Let's it should. We should work with get the AI assistant. So we go in here, look at the prompt that we've put inside of the AI assistant, analyze the user's request. It's very generic Google Calendar reminders. So what we want to do is we want to say for this is what I'm going to put in here to make this work for Google Calendar. Just change that to get medi command S. This is guaranteed to be a problem with the prompt. Yeah exactly. So let's say this please can you use Gmail to fetch me my most recent email? Your most re recent email is hi Michael, there's a new AI lead. Please see below. Name Michael Service AI Description. Okay so you can see that now you've got a fully functioning workflow here. Your reference point should always be too curly. Let's now give it a challenge please can you Brackets dollar sign now which means now exactly now as it stands should all you update the Google sheet always be now. So if the user requests with the following information Name Test Email testatemail.com Service Test A meeting for this Friday. You know it test Description Description. This is a test also can you create a meeting at 6pm today with Michael AD train the closest Friday in the future. And this is a. This is not how this is good. This is, this is what you would do. I mean at 6pm today about the sales. The, the the Please create a meeting today at 6pm about last week's sales and this is what you would do in my situation now to showcase to you but in production it wouldn't but this one fetch my most recent email. But what I'm online should, should fix it. Let's see if we're going to do here is so you can see what I've done. I'm going to command C that and we can now book a meeting for like for Friday at a different time or whatever. I'm going to copy it. I'm going to reset chat session so it starts afresh and I'm going to paste it in now we should see all three tools work and the power of, of an AI assistant. You should see it. It's using Gemini. It's gone back through it's now done the calendar, it's now fetched the email and the output is done these crimp. So press L, open the chat. There's one other thing actually go into Google Calendar and in here inside of the Google Calendar node add an email for at 6pm I have created you flim ensure that they are enabled to be dynamic start and end times. An email. Let me drag this up. What am I doing? I have created a meeting for you at 6pm today about last week's sales. Here is your most recent email from Michael Atomo AI the same one that we've got as before. So guys, well done so far. You've got got two flows built. You're you can now book me and then what we're going to do is we're going to say command S is what we're going to do. Not that command S is what we're going to do and then we're going to say ins. You can now update rows, you can send emails. But all of you have a keen eye will have spotted something. When it comes down to the individual calendar booking, please book me a a meeting on calendar booking. Is that happening like an hour from now at any time. So whenever you sent that message it would fire off and it would do it there. And the reason for that AI agents like to hallucinate. So that's one piece Friday at meeting this Friday at 6pm but the other thing is is that if you look at the prompt that we've actually put inside of the AI assistant. So we go in here, look at the prompt that we've put inside of the AI assistant, analyze the user's request. It's very generic Google Calendar or setting reminders. So what we want to do is we want to say for this is what I'm going to put in here to make this work for Google Calendar to discuss. Our cue for your reference point should always be two curly marketing efforts brackets dollar sign now. Which means now exactly now as it stands should. Let's always be now. So if the user requests a meeting for go There you go. Six till 7pm Is that the time that I gave it 6pm on Friday to book and discuss your marketing efforts. So this Friday, you know, it's the closest Friday in the future. And this is a. This is not how this is good. This is. This is what you would do. This is what you would do in my situation now to showcase to you. But in production it wouldn't but this would one, that's that that's that fixed you. We've learned a little bit about system prompting line should should fix it. Let's see if we. The now the now time in here was really important for for the AI to be able to understand and ground itself to an individual time. Press the curly brackets press. You can now book a meeting for like for Friday at a different time or whatever. So press L. Open the chat. There's one other thing actually go into Google Calendar and in here inside of the Google Calendar node flim ensure that they are enabled to be dynamic start and end times. That's the dollar sign and type now references. Now you can have a look. If you hover over and then what we're going to do is we're going to say command S is what we're going to do. Not that command S is what we're going to do and then we're going to say please book me a meeting on it. It tells you it should at least well they they are in the output even in the result. Friday at me in it's been a long day. Date and time gives you the date time. Exactly what time it is now. Yeah. This Friday at 6pm to discuss our Q4 marketing efforts. Let's that is at the exact time and yeah go There you go. Six till 7pm Is that the time that I gave it 6pm on Friday to book and discuss your marketing efforts. So that's that that's that fixed you. We've learned a little bit about system prompting. The now the now time in here was really important for for the AI to be able to understand and ground itself to an individual time. Press the curly brackets press. We've we've. You. We've created a full AI assistant. So guys we built the dollar sign type now references. Now you can have a look. If you hover over it it tells you it should at least well they they in the output even in the result. It's been a long day. Date and time gives you the date time. Exactly what time it is now. Yeah that is at the exact time. And yet two flows and those two flows are practical working flows. Yeah we've we've you. We've created a full AI assistant. So guys we built is the start of a flow that you could then expand on further which is what we're going to do in our. In our next steps. Maybe we'll build from scratch again. Who knows we'll see what comes to mind. Two flows and those two flows are practical working flows. One is the start of a flow that you could then expand on further, which is what we're going to do in our, in our next steps. Maybe we'll build from scratch again. Who knows, we'll see what comes to mind. I have a bit of a plan around it, but we if you are, this video is designed I have a bit of a plan around it, but we if you are, this video is designed at beginners and if you're a beginner, you have now just learned the fundamentals of NAN and how to implement them into an actual function in flow including problem solving. Something which I gave you the the answer to yes in terms of the now and it is a little bit of a cheat code instead of rewriting the whole system prompt. The point is is that we saw an issue with the calendar booking and we fixed it and we got it to book future meetings. That's as a beginner, as a new person using NAN for the the first time or building for the first time or one of the first times, you should be, you should be proud of yourself for what you've delivered in this, in this video. This puts you like above 80% of people. So you now have a core understanding around things and your brain will be ticking differently I'm sure like about ways how you might be able to use it moving forward. Beginners and if you're a beginner, you have now just learned the fundamentals of NAN and how to implement them into an actual functioning flow, including problem solve. This is one step of a multi step series. Thanks for sticking with me like comment and subscribe especially on this video because if educational videos are what you want then I will continue to produce them. But I need to know that is what you want. Guys, thanks for following me. It's been a long video. I'll catch up with you soon. Cheers. Oven something which I gave you the answer to yes in terms of the now and it is a little bit of a cheat code instead of rewriting the whole system prompt. The point is that we saw an issue with the calendar booking and we fixed it and we got it to book future.