Updated . Built against the tool’s current screens and checked step by step. How we check this lesson
In this free Make.com tutorial, you add AI to a working scenario so every new request gets a warm, personal reply within minutes. Brookside Dental, a small family dental practice, already saves routine patient requests to a Google Sheet. Now the OpenAI module writes a reply from each patient's own answers, and a Gmail module sends it straight to the patient.
At a glance
| Question | Answer |
|---|---|
| What you build | Connect OpenAI to your Make scenario, have ChatGPT write a personal reply from each patient's form answers, and send it with Gmail. |
| Tools | Make |
| Time | About 12 minutes, 30 steps in 5 parts |
| Level | Beginner, no coding needed |
| Cost | Free |
| Ways to learn it | Practice it step by step |
The 5 parts, in short
- Add the AI module: Add a module after Google Sheets, Search for OpenAI, OpenAI, Generate a response.
- Connect OpenAI: Create a connection, Paste your API key, Save the connection.
- Write the prompt: Model, Pick a small model, The instructions, Map the name, and 3 more steps.
- Send it with Gmail: Add the email, Search for Gmail, Gmail, Send an email, and 8 more steps.
- Test it: Save the scenario, Run once, Look at the output, Check the sent email.
Start here first: Your First Make.com Scenario: Google Forms to Google Sheets (module 1 of Make.com Automation Basics)
In the course: Part 8 · Make.com Automation Basics (Module 34 of the full course)
Part of: Email on Autopilot (step 4 of 5). Next up: AI Email Responder: Draft Gmail Replies Automatically with n8n
What you will build
You continue the scenario from the previous two modules of this course. A Google Forms Watch Responses trigger picks up each new patient request, a router sends urgent requests to the front desk by email, and the fallback route adds routine requests to Google Sheets.
On that routine route, after Google Sheets, you add the OpenAI (ChatGPT, Sora, Whisper) module Generate a response. You connect it to your own OpenAI account with an API key, pick a model, and write a prompt that tells ChatGPT who it is writing as, what to include and what to avoid. You map the patient's name and reason for visit from the form into the prompt. Then you add Gmail Send an email, send it to the email address the patient typed into the form, and map the module's Result, the text ChatGPT wrote, into the body.
Who this Make.com tutorial is for
This module is for small business owners, office managers and front-desk teams who already have a simple Make scenario and want it to write, not just move data. No coding is needed. It is the third module in the Make.com Automation Basics course. You need an OpenAI account with API credit to do it for real; the practice copy uses a fake key.
Writing a prompt that works in an automation
In a scenario nobody reads the reply before it goes out, so the prompt has to do the job a careful person would. Say who the model is writing as, what the email must include, and what it must never do. For a dental office that means greeting the patient by first name, responding to their reason for visit, promising a call within one business day, and giving no medical advice, prices or treatment promises.
Put the instructions first and the mapped form answers after them, so every run gets the same rules with different details. Ask for the exact format you need. Because Gmail sends the body as HTML, the lesson asks for simple paragraphs in p tags, so the email does not arrive as one long block of text.
Costs, models and tokens
The Generate a response module uses your own OpenAI API key, and OpenAI bills that use separately from any ChatGPT subscription. A short email costs a small fraction of a cent on a mini model, but you need credit on your OpenAI account before the module will run. The model list comes from your account, so it may show newer models than the lesson.
Max Output Tokens caps how long a response can get. It includes the model's hidden reasoning tokens as well as the visible text, so setting it too low can cut a reply short or leave it empty. The lesson uses 2000, which leaves room for a short email while keeping each run's cost predictable.
How to use this in your business
The pattern is the same for any reply your team writes over and over: map the customer's details in, tell the model exactly what to write, and map its Result into the next module. Before you switch the scenario on for real, run it several times with realistic test responses and read every reply. If you would rather check each email first, send the Result to Gmail Create a draft email instead of Send an email.
Try it yourself, free
Reading the steps is a start. Doing them is how it sticks. The interactive module walks you through every click in a practice copy of Make, checks each step, and never touches your real accounts.
Start “Add AI to Make.com: ChatGPT Writes a Personal Reply, Gmail Sends It”Step by step: Add AI to Make.com: ChatGPT Writes a Personal Reply, Gmail Sends It
All 30 steps of the interactive module, in order. In the module you watch each one, then do it yourself in a practice copy that checks your work.
1. Add the AI module
- Add a module after Google Sheets. This is the scenario from the last module. Routine requests land in Google Sheets. Click the + on the right of Google Sheets.
This is the scenario from the last module. Urgent requests alert the front desk, and routine ones land in the sheet. Now every routine patient should also get a personal email reply, written by ChatGPT.
- Search for OpenAI. Type openai.
- OpenAI. Click OpenAI (ChatGPT, Sora, Whisper).
That's the app for ChatGPT.
- Generate a response. Under Responses And Conversations, click Generate a response.
At the top there's simple text prompt. It runs on Make's own credits, with no OpenAI account. You'll use your own OpenAI key instead, so under responses and conversations, click generate a response.
2. Connect OpenAI
- Create a connection. Click Create a connection.
The module needs a connection to your OpenAI account.
- Paste your API key. Keep the connection name. In API Key, paste your secret key from your OpenAI account. (Here, use the practice key: sk-proj-EXAMPLE-not-a-real-key-0000)
Keep the name. In the API key field, paste a secret key from your OpenAI account. You make one on the OpenAI platform, under API keys. OpenAI bills API use separately from a ChatGPT subscription, so you need a little credit there, as the tip says. Here, paste the practice key.
- Save the connection. Leave Organization ID empty and click Save.
Organization I D is optional, so leave it empty.
3. Write the prompt
- Model. Click the Model field.
Now pick the model that writes the reply.
- Pick a small model. Click gpt-5.4-mini (system).
A short email doesn't need the biggest model. A mini model is fast and costs very little. Your list may show newer models.
- The instructions. Keep Prompt Type as Text prompt. In Prompt, paste the instructions.
Keep the prompt type as text prompt. The prompt is where you tell ChatGPT who it is, what to write and what not to do. Near the end, they ask for simple H T M L paragraphs, because the email body is sent as H T M L.
You are the front desk at Brookside Dental, a friendly family dental practice. Write a short email reply to a new patient who just asked for a first visit. - Greet them by first name and thank them. - Respond to their reason for visit in one or two warm sentences. - Say we will call within one business day to book a time. - No medical advice, no prices, no promises about treatment. - Sign off as: The Brookside Dental team Write only the email body, as simple HTML with each paragraph in <p> tags. The patient's name and reason for visit:
- Map the name. The mapping panel shows the form answers. Under Full name, click Value.
Now add the patient's own answers, so every reply is personal. Under full name, click value.
- Map the reason for visit. Under Reason for visit, click Value.
And under reason for visit, click value.
- Max Output Tokens. In Max Output Tokens, type 2000.
Max output tokens caps how long the answer can get, and so what one run can cost. It counts the model's hidden reasoning too, so don't set it too low. Two thousand leaves plenty of room for a short email.
- Save the module. Click Save.
4. Send it with Gmail
- Add the email. Click the + on the right of the OpenAI module.
Now send what ChatGPT writes.
- Search for Gmail. Type gmail.
- Gmail. Click Gmail.
- Send an email. Under Actions, click Send an email.
- To. Check that Connection shows your Gmail connection from the last module (pick it if it doesn't). Under To, click Recipient email address 1.
Check that the connection is the Gmail connection you made in the last module. If it's empty, pick it from the list. The reply goes to the patient, so under To, click recipient email address one.
- The patient's email. Under Email, click Value.
Under email, click value. That's the address the patient typed into the form.
- Subject. In Subject, type Your visit request at Brookside Dental
- Body type. Click the Body type field.
- Raw HTML. Click Raw HTML.
The body is one field, and the paragraph tags from ChatGPT keep the email readable.
- Body contents. Click the Body contents field.
- Map ChatGPT's reply. The mapping panel now shows the OpenAI module's output. Under 5. OpenAI (ChatGPT, Sora, Whisper) - Generate a response, click Result.
The mapping panel now shows the OpenAI module too. Result holds the text ChatGPT writes.
- Save the module. Click Save.
5. Test it
- Save the scenario. Click the Save icon in the toolbar.
Save the scenario with the save icon in the toolbar.
- Run once. A new routine request from Daniel Brooks is waiting in the form. Click Run once.
A new request is waiting. Daniel's daughter needs her first checkup, and she's nervous.
- Look at the output. Every module on the routine route shows a bubble with 1. Click the bubble on the OpenAI module.
Every module on the routine route ran once.
- Check the sent email. Under Output, Result holds the reply ChatGPT wrote. Now open the Gmail tab.
Under output, result holds the reply, written for Daniel and Emma.
Frequently asked questions
- Which OpenAI module should I use in Make.com to generate text?
- Generate a response, under Responses And Conversations in the OpenAI (ChatGPT, Sora, Whisper) app, uses your own OpenAI connection. Simple text prompt, under Get Started, needs no connection and uses Make's credits instead. The Chat Completions module, under Chat And Assistant, is called Generate a completion.
- Do I need a ChatGPT Plus subscription?
- No. The module uses the OpenAI API, which is billed separately by usage. You create an API key in your OpenAI account and add a little credit there. A ChatGPT subscription does not include API credit.
- Where is the AI's reply in the output?
- In the Generate a response output bundle, Result holds the text the model wrote. Map Result into the next module, such as the Body contents of Gmail Send an email.
- Why does my AI email arrive as one long paragraph?
- Gmail's Body contents is HTML, so plain line breaks are ignored. Ask the model to put each paragraph in p tags, as the lesson's prompt does, or set up the body with HTML yourself.
- How does the interactive lesson work?
- Each step is shown to you first on the left (Watch), with a short voice-over. Then you do it yourself on a practice copy of Make on the right (Your turn), and every click is checked. If you get stuck, press Do it for me. Nothing touches a real Make, OpenAI or Gmail account, and the API key is a fake placeholder.
More in Make.com Automation Basics
Build your first Make.com scenario: every new Google Forms response becomes a new row in Google Sheets. Trigger, connection, mapping, Run once and a schedule.
Free · Live ~12 min
Webhooks: Connect a Website Form to Make.comStart a Make scenario with a Custom webhook, paste its URL into your website form's webhook setting, send a test so Make learns the fields, and save every request to…
Free · Live ~12 min
All 5 modules in Make.com Automation Basics
Terms in this lesson
Plain-English definitions, each with an example from a free lesson.