Memuat
Memuat
Want to go deeper?
Join one of our courses or ask us directly on WhatsApp.
Picked from the software, topic, and level of “Agentic AI Without Coding Class #28: Practice: An Incoming Form as a Workflow Trigger”.
Other ways we can help with Agentic AI.
Learn with a mentor
Scheduled AI & Machine Learning courses with live guidance and a certificate.

MATLAB Agentic AI Course: Building Autonomous AI Agents with MATLAB & Simulink

Midjourney Course: Create Stunning AI Images from Text
★ 4.9

DALL-E Course: Creating Professional AI Images from Text
★ 4.9
Have it built for you
No time to learn Agentic AI? Our team can do the project for you.

MATLAB & Simulink Development — Agentic AI for Simulation Automation and Data Analysis

AI Image Creation with Stable Diffusion — Unlimited Custom Visual Production
★ 4.9

AI Image & Video Creation with Leonardo AI — All-in-One Visuals for Business
★ 4.9
Get notified about new Agentic AI classes
At most one email a week. Unfollow anytime with one click.
The full narration of this video. The current sentence is highlighted while the video plays; click any sentence to jump there.
Welcome to Episode 28. Starting today we build the real Sunset Coffee order workflow, step by step, until it's active in Episode 30. Today we set up its trigger: the order form. We compare two trigger options, build a form in n8n, send a test order, and read the first data in the output panel.
Here's the map of the workflow we'll finish in Episode 30. The order form as the trigger, then the AI step, then saving to Google Sheets, and finally an email to Dimas. Today we focus on the first node. This node decides what data will flow through the whole workflow. If data is missing, the next nodes are short of material too. For example, without a WhatsApp number, Dimas can't confirm the order. Let's picture one order's journey. A customer finds the form link in Sunset Coffee's Instagram bio, or Dimas sends it to them in a WhatsApp chat. The customer fills in the form: name, WhatsApp number, order, and notes. It takes just a minute. As soon as they press submit, n8n receives the data instantly, as one item ready for the next node to process. The customer sees a thank-you page, so they're confident their order went through. The workflow does the rest, without Dimas having to copy anything. Before building the form, decide on the fields. First, the name, to greet the customer and log the order. Second, the WhatsApp number, so Dimas can confirm and send delivery updates. Third, the order, with product and quantity. For example, two one-liter palm sugar iced coffees. Fourth, notes. This field matters most for the AI step later, because customers often write questions, complaints, or urgent requests here. And remember, ask only what you need. Every extra field means more customer data you have to protect. This is the form Laras has been using so far, made with Google Forms. A customer named Sari starts filling in her name. Then her WhatsApp number. In this video all numbers are made up, and when practicing, you should use made-up data too. Sari orders three one-liter palm sugar iced coffees. And in the notes she writes: please just send it tomorrow, I'll pay by transfer. Notes like this used to get missed. Now let's make this form trigger a workflow.
There are two ways to capture form submissions. Option one, the n8n Form Trigger: the form is built right inside an n8n node. Option two, keep using a Google Form, and have n8n watch its response spreadsheet with a Google Sheets Trigger. The difference shows in speed. The Form Trigger runs instantly when the form is submitted. The Sheets Trigger works by checking, for example every minute, whether there's a new row. So there can be a delay of about a minute. The Form Trigger doesn't need a connection to a Google account. The Sheets Trigger needs a Google credential, which is permission for n8n to read your spreadsheet. As for data, Google Form answers are saved automatically in the response sheet. Form Trigger answers flow in n8n, and we save them ourselves to Google Sheets in Episode 30. In short, the Form Trigger suits you if you're starting from scratch and want it fast. A Google Form with a Sheets Trigger suits you if your old form is already used by customers and its link is already out there. For this course, the main track is the n8n Form Trigger. The reason is simple: fewer connections, it runs instantly, and all the settings are in one place. If your Google Form is already being used by customers, don't worry. In the last step of this episode we practice the alternative, and all the nodes after it stay the same. For Make users, the equivalent is a Google Sheets module that watches for new rows. The principle is the same: one form submission becomes one package of data that flows to the next module.
Open a new workflow called Sunset Coffee Orders. Click add first step, then choose On form submission. The node that appears is called n8n Form Trigger. Rename it to Order form. The settings panel opens. Fill in the form title: Sunset Coffee Order. Then a short description that customers will read at the top of the form. Leave Authentication set to None, so customers can open the form without a password. Leave Respond When set to Form Is Submitted, which means customers see a thank-you message as soon as they press submit. Now the Form Elements section, the list of questions. Add the first field: Name, type Text, and mark it as required. Second field: WhatsApp number, also Text and required. Third field: Order, type Textarea so the box is wider, and required. Fourth field: Notes, type Textarea, and not required. Not every customer has notes, and a form with too many required fields puts people off. Besides Text and Textarea, there are other field types like Number for numbers, Date for dates, and Dropdown List for choices. For this practice, four text fields are enough. One important point: the labels you write here become the field names in the data. So write them clearly and don't change them often, because the next nodes will refer to these names. Finally, in the additional options, set the text that appears after the form is submitted. For example: thank you, we've received your order, Dimas will confirm on WhatsApp. A few tips to make the form easy to fill in. First, the title and description name the shop and say when customers will be confirmed. Second, only the important fields are required. Third, give an example answer in the field hint, for example two one-liter palm sugar iced coffees, so customers write in a similar pattern. A consistent pattern also makes it easier for AI to read the order. Fourth, if the product list is fixed, use the Dropdown List type so customers just pick and don't make typos. And never ask for sensitive data like ID numbers or PINs in an order form. Besides being unnecessary, it makes customers suspicious.
At the top of the panel are two form addresses. The first is the Test URL, just for testing. The second is the Production URL, the address for customers. The Production URL only works once the workflow is activated, which we'll do in Episode 30. So don't share any link with customers yet. Later, this Production URL is what goes in the Instagram bio, gets sent over WhatsApp, or is printed as a QR code at the counter. The address doesn't change as long as the form and workflow names stay the same. To test, click Execute step. The node goes into a mode where it waits for one test submission, and n8n usually opens the test form in a new tab right away. Note that this waiting mode only captures one submission. If you want to send the next test entry, click Execute step again before filling in the form. This is the test form built by n8n. Its fields are exactly the ones we set up. We'll fill it in with the same data from Sari, starting with the name. Then a made-up WhatsApp number, and the order: three one-liter palm sugar iced coffees. We fill in the notes like Sari did: please just send it tomorrow, I'll pay by transfer. Click submit. The customer sees the thank-you message we set, and at the same moment, n8n receives the data as one item. Picture it from Sari's side. She just spends a minute on the form, then sees a message that her order was received and will be confirmed on WhatsApp. A simple experience like this makes customers feel looked after, even outside Dimas's working hours. So they don't get mixed up, let's line up the two addresses. The Test URL is used by you while building the workflow. The Production URL is used by real customers. The Test URL only works after you click Execute step, and only for one submission. The Production URL keeps working as long as the workflow is active. Test submission results show right in the node panel, so they're easy to check. Customer submission results are logged in the Executions tab. And in the data, the formMode field tells you which address a submission came from: test or production. Later you can use this field to tell test data apart from real orders.
Back to n8n. The Order form node now has a green check, and the output panel shows one item. This is the first order data to enter our workflow. Look at the field names: Name, WhatsApp number, Order, and Notes, exactly the same as the labels we wrote in Form Elements. The values are also exactly what Sari entered, including the note about sending tomorrow and paying by transfer. This is the data the AI will read in Episode 29. If you want to see the data as it is, switch to the JSON view. The content is the same, just written as structured text. For beginners, the Table or Schema view is usually easier to read. Scroll down a little, and there are two fields we didn't create. The first is submittedAt, the time the form was submitted. This field is useful for a time column in the spreadsheet. The second is formMode, which says test because we used the Test URL. Once the workflow is active and customers use the Production URL, it changes to production. Now a time-saving trick: click the pin icon in the output panel to pin this data. When we set up the AI node later, we won't need to fill in the form again every time we test. Don't worry, pinned data is only used when testing in the editor. Once the workflow is active, n8n still uses real order data from customers. Before moving on, get to know the four most common problems. First, the form was submitted but n8n does nothing. Almost always, the cause is forgetting to click Execute step before filling in the form. Second, the field names in the output are suddenly different. Usually because a label was changed after the test data was sent. Resend the test data every time you change a label. Third, the Production URL doesn't respond. That's normal, because the workflow isn't active yet. While building, use the Test URL. Fourth, not a technical problem but an important one: don't use real customer data for practice. Made-up data like Rina and Andi is enough to test every possibility. One test entry isn't enough. Send four test orders that represent four categories. First, Rina: two one-liter palm sugar iced coffees, deliver in the afternoon. This is a regular order. Second, Andi: one hampers package, with the note is there a sugar-free version? This is actually a question that needs an answer. Third, Joko: one cold brew, with a note that yesterday's coffee spilled on the way and he's disappointed. This is a complaint. Fourth, Putri: five hampers packages for an office event tomorrow at eight in the morning. This is an urgent order that needs priority. Why these four categories? Because each needs a different action. A regular order just gets processed. A question needs an answer. A complaint needs careful handling, and an urgent order needs to go first. The last column is what we expect. In Episode 29, we check whether the AI gives the same category. If it's different, our prompt needs fixing. Remember, click Execute step before each test submission.
Now the alternative. If you keep using a Google Form, every answer goes automatically into a response spreadsheet, one row per submission. The first column is Timestamp, filled in by Google Forms itself. The next columns follow the question titles: Name, WhatsApp number, Order, and Notes. When Joko submits the form, a new row appears at the bottom. This new row is what the Google Sheets Trigger will capture. The advantage of this track: Laras still has an archive of every answer in the spreadsheet, even when the workflow is switched off or a node fails. No order gets lost. In n8n, add the first step, search for Google Sheets, and choose the trigger for when a new row is added. The Google Sheets Trigger node appears on the canvas. First, the credential. Click create new credential, then sign in with the Sunset Coffee Google account. As in Session 5, read the permissions requested, and use the business account, not a personal one. Second, Poll Times, which is how often n8n checks the sheet. Every minute is enough for coffee orders. Third, choose the Sunset Coffee Order Responses document, and the sheet Form Responses one. Fourth, Trigger On: Row Added, which means only new rows trigger the workflow. To test, click the test button in the panel. n8n fetches a sample row from the sheet, and the output panel shows the Timestamp, Name, WhatsApp number, Order, and Notes. From here on, all the following nodes are exactly the same as on the Form Trigger track. One important message for this track: don't change the question titles in the Google Form once the workflow is built. Question titles become column headers, and column headers become the field names that the next nodes refer to. Before Episode 29, make sure these four things are done. First, choose your track: the n8n Form Trigger, or a Google Form with a Sheets Trigger. Second, set up four fields: Name, WhatsApp number, Order, and Notes, with consistent names. Third, send four test entries: Rina, Andi, Joko, and Putri, so we can test the AI with a variety of orders later. Fourth, pin one output as sample data, and save the workflow. For your own business, replace the data with the orders or requests you receive most often.
Let's sum up. We set up a trigger from the order form, chose the n8n Form Trigger as the main track with the Google Sheets Trigger as an alternative, tested it through the Test URL, and read the order data in the output panel. In the next episode, we hand this order data to AI. The AI node will summarize each order and give it a category: Regular order, Question, Complaint, or Urgent. See you in Episode 29.
Try these on your own after watching to lock in what you learned.
Type a question or keyword; we find the part of this episode that explains it and jump the video there.
Notes are saved with the current video time and are only visible to you.
Stuck on a step? Ask here; the instructor and other learners can answer.
No questions yet.

Cerdas Inovasi Digital Team
Software developers, BIM & 3D engineers, and AI practitioners
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.
See our clients →