At most one email a week. Unfollow anytime with one click.
Transcript
The full narration of this video. The current sentence is highlighted while the video plays; click any sentence to jump there.
Show transcript
Introduction
Welcome to Session 6, the most hands-on session in this course. From now on we build visual automation with n8n and Make. In this episode we start with the basics: what triggers, actions, and workflows are, how n8n and Make differ, cloud versus self-hosted, how to create an account, and a quick tour of the canvas.
Step 1: From asking AI to a workflow that runs itself
In Session 5, we connected AI to Gmail, Calendar, and Sheets. The results were very helpful, but there was always one condition: you had to open the chat and ask first. In Session 6 the pattern flips. A workflow runs on its own when something happens, for example an order comes in or an email arrives. AI is still used, but as one step inside the flow, not as the place where you type instructions. Before we get to the terms, let's look at Sunset Coffee today. Orders come in through chat. Rina orders two liters of palm sugar iced coffee, delivered in the afternoon. Budi asks whether there's still cold brew. Dimas has to reply, then copy each order into the spreadsheet one by one. Then a complaint arrives: yesterday's coffee spilled on the way. When the chat is busy, a message like this easily gets buried among regular orders. And there's an urgent order: three hampers packages for an office event tomorrow morning. This is why we're building a workflow: orders get recorded neatly, and the important ones don't get missed. Let's do a rough calculation. Say copying and sorting one order takes about four minutes. With thirty orders a day, that's about two hours every day, just for copy-paste work. The workflow we'll build processes one order in a matter of seconds, day or night. These numbers are just an illustration, so measure the savings in your own business. Those two hours could let Dimas serve customers better. Here's our project for this whole session, using Laras's Sunset Coffee in Bandung as the example. Step one, a customer fills in the order form. Step two, AI reads the order, writes a short summary, then gives it a category: Regular order, Question, Complaint, or Urgent. Step three, all the data is saved as a new row in Google Sheets. Laras gets a neat order recap without copying by hand. Step four, Dimas the admin receives an email notification. The email subject already includes the category, so complaints and urgent orders stand out right away. We'll build it bit by bit over five episodes. This episode and the next are the foundation. Episodes twenty-eight to thirty are practice, until the workflow is active and running on its own.
Step 2: Three core terms: trigger, action, workflow
Now for three terms you'll hear again and again. First, the trigger. This is the first node, which decides when the workflow starts running. Here, the trigger is: an order form comes in. Second, the action. Every node after the trigger is an action, one piece of work being done. Our first action asks AI to summarize and categorize the order. The second action saves the result as a new row in Google Sheets. The third action sends an email through Gmail. Each of these boxes is called a node, and the arrows between them show where the data flows. All of them together are called a workflow: a chain of trigger and actions that runs automatically from left to right. A green check mark means each node did its job successfully. One more thing. Without the AI node, this flow is ordinary automation that only moves data. With an AI node in the middle, the workflow can understand what the order says. This is automation with an AI step, which we covered in Session 1. Triggers come in many kinds. Two are the most common. A form or message coming in, like the Sunset Coffee example. Then a schedule. For example, every Monday at seven in the morning the workflow prepares last week's sales report, without anyone pressing a button. It can also be a new email arriving in Gmail, or a new row in Google Sheets added by a coworker. There's the manual trigger, where you press the button yourself. This is very useful for testing. And there's the webhook, a special address where another app sends a signal. For now, just know the name. To make it more real, let's look around us. For Laras, the trigger is the order form, and there are three actions: AI gives a category, the data is saved, and an email goes to Dimas. For Dewi at Maju Bersama Co., the trigger is a quote email from a supplier. The actions log the quote in a spreadsheet, then remind the purchasing team. For Mr. Hendra at Cerdas Tutoring, the trigger is a student submitting homework through a form. The action: AI prepares first feedback, which he still reviews before it's sent. And for Nadia, the trigger is a schedule: every Friday at four in the afternoon, the workflow builds a recap of next week's posting schedule. Think of one trigger-and-action pair from your list of five tasks in Session 1.
Step 3: n8n and Make: two apps, one concept
Now, n8n and Make. Both are visual automation apps, and the concept is the same. Only the terms differ. One automated flow in n8n is called a workflow, while in Make it's called a scenario. One block in n8n is called a node, in Make it's called a module. So if you read a Make tutorial that mentions a module, just picture a node. One run of the flow from start to finish is called an execution in n8n. In Make it's called a run, and each module that works inside that run counts as one operation. This difference matters for cost. n8n usually counts quota per execution, while Make usually counts per operation. The counting can change, so always check the latest pricing page before choosing a plan. Visually, n8n nodes are boxes, while Make modules are circles. Both are arranged from left to right, following the flow of data. This is the same workflow built in Make, where it's called a scenario. The modules are round, but the order is exactly the same: form, AI, Sheets, then email. When this scenario runs once for one order, all four modules work, so it counts as roughly four operations. In n8n, that same single run counts as one execution. To test a scenario, Make has a Run once button at the bottom of the screen. It does the same job as the test button in n8n: it runs the flow once so you can see the result. In this course we use n8n as the main track. The first reason: its AI nodes are complete, from chat models all the way to the AI Agent node we'll use in Session 7. The second reason: n8n can be used in the cloud or installed on your own server. We'll cover the difference in a moment. The third reason: the concepts apply to Make too. If you're already used to Make, just follow each step and translate the terms: workflow becomes scenario, node becomes module.
Step 4: Cloud or self-hosted?
A common question: cloud or self-hosted? Cloud means n8n runs on n8n's own servers. Self-hosted means n8n is installed on your own server. Think of it this way: the cloud is like renting a stall in a mall, you just open up and start selling. Self-hosting is like building your own shophouse. With the cloud you just sign up, while self-hosting means renting a server and installing the app. For maintenance, app updates and data backups in the cloud are handled by n8n. When self-hosting, it's all your responsibility. For cost, the cloud uses a subscription, and there's usually a trial period for learning. The community version for self-hosting is free, but you still pay for the server rental. Prices change often, so check the latest pricing page. In short, the cloud suits beginners and small businesses. Self-hosting suits you if you have technical help, or if there's a rule that data must be stored on your own server. My recommendation for this course: start in the cloud. Your energy is better spent learning to build flows, not looking after a server. Don't worry about being locked in. n8n workflows can be exported and moved to self-hosting at any time. And for Make users, the app is already cloud-based, so there's no choice to make. One exception: if your office already runs n8n on its own server, ask the IT team whether you can use an account there. The steps in this course work exactly the same. So when should you move to self-hosting? There are three signs. First, your workflows run so often that the subscription costs more than renting a server. Second, there's a rule requiring customer data to be stored on your own server, for example from a corporate client. Third, there's a colleague or vendor ready to maintain the server: updating the app and backing up data regularly. If none of these signs apply yet, stay in the cloud and focus on building useful workflows.
Step 5: Creating an n8n account safely
Now let's create an account. Open the n8n website, then click the button to get started or sign up. Remember, the screens in this video are simulations, and button names may differ slightly because the app is updated often. Fill in your name, email, and password. My advice: use a business email, like the Sunset Coffee email, not a personal one, so the account is easy to hand over if a new admin joins later. Then choose an account name. This name becomes part of your workspace address, for example sunsetcoffee.app.n8n.cloud. Pick a name that's short and easy to recognize. Open your email inbox and click the verification link. Without this step, the account can't be used yet. There's usually a short survey about your needs, like your line of business and why you want automation. Answer briefly, and you'll land on the n8n main page. If you've chosen Make, the steps are similar: sign up on the Make website, verify your email, then go to the scenario list page. Before moving on, secure this account. Later, the automation account will be connected to your business Gmail and Sheets, so it's as valuable as the office key. First, a strong and unique password. Second, turn on two-step verification if it's available in your account settings. Third, use a business email, so access isn't lost when people change roles. A business email also makes it easier to keep shop matters separate from personal ones. And what you must not do: share the password in chat with lots of people. If you need to add team members, use the invite users feature, if your plan offers it.
Step 6: A tour of the canvas
From the main page, click the button to create a new workflow. You'll see a dotted canvas, like a drawing board. In the top left is the workflow name, which you can change right away. In the middle of the canvas is one box with a plus sign to add the first step. The first step is always a trigger, because the workflow needs to know when to start. Click that box, and the panel on the right shows the trigger options. There's a manual trigger, an event in an app like Gmail or Sheets, a schedule, a form submission, and a webhook. Notice the On form submission option. This is what we'll use in Episode 28 to capture Sunset Coffee's orders. For this tour, we'll just choose the manual trigger. Now the canvas has a manual trigger node. Its left side is rounded, a sign that it's a trigger. On its right side is a connection point for adding the next step. Click the plus sign at the end of the node, type the action's name in the search box, then choose it. For example the Edit Fields node, used to prepare data. The arrow connects automatically. On screen there's a button to run a test. It's called Execute workflow, or Test workflow in other versions. This button runs the whole flow once so you can try it. Now notice the Inactive label next to the workflow name. While it's not active, the workflow only runs when you test it. Once activated, it runs on its own every time the trigger happens. In the latest version, this button may be called Publish. Let's sum up the five parts of the screen. First, the canvas, where you arrange nodes from left to right. Second, the node panel, which opens when you double-click a node. It holds the settings, the data coming in, and the data going out. This is the panel we'll take apart in the next episode. Third, the test buttons. One runs the whole workflow, and one runs just a single node, called Execute step or Test step. Fourth, the Active toggle, which turns the workflow on so it runs by itself. Fifth, the Executions tab, a history of every time the workflow ran, with a success or failure status. This is the first place to look when something goes wrong. Let's peek at the Executions tab. Every time the workflow runs, a log line appears here: the number, the time, whether it was a test or a real run, and its status. The top two executions succeeded, each in under a second. Nodes that succeed also get a green check mark on the canvas. The first execution failed. Click its row, and n8n opens the canvas exactly as it was at that moment, showing which node failed. We'll practice reading error messages in Episode 30. Before the next episode, do these four small things. First, create an n8n or Make account, and verify your email. Second, create a new workflow and name it Sunset Coffee Practice. Third, add a manual trigger and one Edit Fields node, just like on screen earlier. Fourth, click Execute workflow once, then open the Executions tab. See the log of your very first execution for yourself. Congratulations, you've officially run your first workflow.
Done!
Let's sum up. The trigger decides when a workflow starts, actions are the work in each node, and the workflow is the chain of both. In Make the terms are scenario, module, and operation, but the concept is the same. In the next episode, we open up the nodes: how data flows as items and fields, how to read the input and output panels, and how to map data with drag and drop. See you in Episode 27.
Practice exercises
Try these on your own after watching to lock in what you learned.
Write three trigger-and-action pairs from your own work, like the examples from Laras, Dewi, Mr. Hendra, and Nadia.
Create a workflow called Sunset Coffee Practice with a manual trigger and one Edit Fields node, then click Execute workflow once.
Open the Executions tab and find the test execution you just ran. Note its status and how long it took.
Ask the class material
Type a question or keyword; we find the part of this episode that explains it and jump the video there.
My notes
Notes are saved with the current video time and are only visible to you.
Questions & answers
Stuck on a step? Ask here; the instructor and other learners can answer.
No questions yet.
Your instructor
Cerdas Inovasi Digital Team
Software developers, BIM & 3D engineers, and AI practitioners
53 video classes1,286 episodes
We build software, BIM models, games and AI automations for clients every day, and turn that hands-on work into step-by-step classes: every step is shown on screen and comes with a PDF guide you can keep.
