Zapier & Make: Connecting Your Tools
Lesson 7.2 — Zapier and Make: Automation Without Code
Automation used to require a developer. You needed someone who could write scripts, connect APIs, manage servers, and handle the infrastructure that keeps things running reliably. That changed when visual automation platforms arrived. Zapier (founded 2011) and Make (formerly Integromat) turned complex multi-app workflows into drag-and-drop diagrams that anyone with logical thinking skills can build.
Adding AI into these workflows is what makes them genuinely powerful. The automation handles the repetitive mechanical steps; the AI handles the parts that require language understanding, judgment, or generation. Together, they can replace significant amounts of manual work.
The Core Concept: Triggers and Actions
Every automation built in Zapier or Make follows the same fundamental structure:
Trigger → Action(s)
A trigger is an event in one app that starts the automation:
- A new email arrives in Gmail
- A form is submitted on your website
- A row is added to a Google Sheet
- A new file appears in Dropbox
- A Slack message is sent to a specific channel
An action is something the automation does in response:
- Send an email
- Create a calendar event
- Add a row to a spreadsheet
- Send a Slack message
- Call an AI model and generate a response
A multi-step automation chains several actions together:
- New form submission → AI generates personalised reply → Draft saved in Gmail → Notification sent in Slack
The "AI" step is one action in a sequence — you send text to a model (via an OpenAI or Anthropic connector), receive a response, and pass it to the next step.
Key takeaway: AI in automation is just one step in a chain. The power comes from connecting AI's language capabilities to real systems — email, databases, CRMs, calendars — that do things in the world.
Zapier vs Make: Which Should You Use?
| Feature | Zapier | Make |
|---|---|---|
| Ease of use | Very easy — linear, step-by-step | More complex — visual flow diagram |
| Free tier | 100 tasks/month, 5 Zaps | 1,000 operations/month, unlimited scenarios |
| Number of app integrations | 6,000+ | 1,000+ |
| Visual interface | Simple, text-based steps | Visual drag-and-drop canvas |
| Complex logic (loops, conditionals) | Limited on lower plans | Excellent |
| AI integrations | Native OpenAI, Anthropic connectors | Good but sometimes requires configuration |
| Best for | Beginners, simple automations | Power users, complex workflows |
| Paid plans from | ~$20/month | ~$9/month |
Recommendation: Start with Zapier. Its interface is the most intuitive for people who have never built an automation before. Once you outgrow it — or if you need complex branching logic — move to Make.
Free Tiers: What You Can Actually Do Without Paying
Zapier free tier:
- 100 tasks per month (a "task" is one successful action step)
- Maximum of 5 active Zaps (automations)
- Two-step Zaps only (one trigger + one action)
- 15-minute minimum polling interval (your Zap checks for new data every 15 minutes)
This is enough to experiment and build simple useful automations. Meaningful professional use typically requires a paid plan.
Make free tier:
- 1,000 operations per month
- Unlimited scenarios (automations)
- Up to 2 active scenarios at once
- 15-minute minimum scheduling interval
Make's free tier is more generous in terms of operations per month, which makes it better for low-volume but multi-step automations.
Walkthrough: Gmail → AI → Notion
Here is a step-by-step walkthrough of a genuinely useful automation: summarising important emails and saving them to a Notion database.
What it does: When an email arrives in a specific Gmail label, the automation sends the email body to an AI model, which summarises it and extracts key action items. The summary is saved as a new page in your Notion database.
Why it's useful: Instead of maintaining a manual summary log of important client emails, project updates, or reports, the system does it automatically.
Step 1: Set up your Gmail label
In Gmail, create a label called "Auto-Summarise" (or whatever you prefer). Apply it manually to emails you want to capture, or set up a Gmail filter to apply it automatically.
Step 2: Create your Notion database
In Notion, create a new database with these properties:
- Name (title)
- Summary (text)
- Action Items (text)
- Original Subject (text)
- Date Received (date)
- From (text)
Step 3: Build the automation in Zapier
Log into Zapier and click "Create Zap"
Step 1 — Trigger: Gmail
- App: Gmail
- Event: New Labelled Email
- Label: Auto-Summarise
- Test the trigger by applying the label to a real email
Step 2 — Action: OpenAI (or Anthropic)
- App: OpenAI
- Action: Send prompt
- Prompt field (paste this, then map in the email data):
You are a helpful assistant. Summarise the following email in 3-5 bullet points.
Then extract any action items, each starting with a verb.
Email subject: [map Gmail Subject field here]
Email body: [map Gmail Body Plain field here]
Format your response as:
SUMMARY:
- [bullet 1]
- [bullet 2]
ACTION ITEMS:
- [action 1]
- [action 2]
- Model: GPT-4o or GPT-3.5-turbo (3.5 is faster and cheaper for this task)
- Test this step to see the AI output
Step 3 — Action: Notion
- App: Notion
- Action: Create Database Item
- Database: select your Notion database
- Map the fields:
- Name → Gmail Subject
- Summary → parse the "SUMMARY:" section from the AI response
- Action Items → parse the "ACTION ITEMS:" section
- Original Subject → Gmail Subject
- Date Received → Gmail Date
- From → Gmail From
Publish the Zap.
What Happens in Practice
From now on, any email you label "Auto-Summarise" triggers the automation. Within a minute or two, a new page appears in your Notion database with a clean summary and extracted action items — no manual work required.
You can extend this:
- Add a Slack notification as a Step 4: "New summary added: [subject]"
- Add a Gmail action to mark the email as read automatically
- Filter by sender to only process emails from specific clients
Common Beginner Mistakes
Mistake 1: Mapping the wrong field When connecting AI output to the next step, you need to map the AI's text response correctly. Test each step before connecting the next.
Mistake 2: Forgetting to handle failures What happens if the AI call fails or returns an empty response? Set up error notifications (Zapier can email you on failure) so you know when something breaks.
Mistake 3: Running up task counts accidentally An automation that triggers on every email could use your monthly task allowance quickly. Always add filters to restrict what triggers the automation.
Mistake 4: Storing sensitive data in automation logs Zapier and Make log your automation runs, including the data passed through. Do not run sensitive personal or confidential business data through free-tier accounts.
Practice Task
Sign up for a free Zapier account. Browse the "Zap templates" section and find a template that connects two apps you already use. Deploy it, test it, and let it run for a week. This is the fastest way to understand what automations actually feel like to use before you build your own.