Part of Automation + No-Code Hub

Building Your First AI Workflow: A Complete Beginner’s Guide

You’ve heard about AI automation. You know you should be using it. But every tutorial assumes you’re a developer.

This guide doesn’t. By the end, you’ll have a working automation that runs without you.

What You’ll Build

A “Daily Brief” automation that:

1. Pulls your calendar events for today
2. Checks your unread emails
3. Uses AI to generate a morning summary
4. Sends it to your phone via Slack or email

Total build time: 45 minutes

Cost: $0 (uses free tiers)

Prerequisites: A computer, internet, and 45 minutes

Step 1: Choose Your Platform (5 minutes)

Recommendation for beginners: Make.com. The free tier is generous (1,000 operations/month), the visual builder is intuitive, and you can build multi-step workflows without paying.

Go to make.com and create a free account.

Step 2: Plan Your Workflow (10 minutes)

Before building anything, sketch it on paper:

[Trigger: Every morning at 7 AM]
    ↓
[Action 1: Get today's Google Calendar events]
    ↓
[Action 2: Get unread Gmail count + subjects]
    ↓
[Action 3: Send both to ChatGPT with prompt]
    ↓
[Action 4: Format AI response into message]
    ↓
[Action 5: Send to Slack or email]

Why plan first? Because building without a plan leads to tangled workflows that break. Spend 10 minutes planning, save 2 hours debugging.

Step 3: Connect Your Accounts (10 minutes)

In Make.com:

1. Click “Create a new scenario”
2. Click the + icon to add your first module
3. Search for “Google Calendar” → Connect your account
4. Search for “Gmail” → Connect your account
5. Search for “OpenAI (ChatGPT)” → Add your API key

Getting a ChatGPT API key:

• Go to platform.openai.com
• Create an account (or log in)
• Navigate to API Keys → Create new key
• Copy it (you won’t see it again)
• Add $5 credit (minimum) for API access

Cost note: The ChatGPT API costs ~$0.01-0.03 per workflow run. At 30 runs/month, that’s $0.30-0.90/month. Practically free.

Step 4: Build the Workflow (15 minutes)

Module 1: Trigger

• Type: Schedule
• Set: Every day at 7:00 AM
• Timezone: Your local timezone

Module 2: Google Calendar

• Action: Get events
• Time range: Today
• Calendar: Primary

Module 3: Gmail

• Action: Get emails
• Filter: Unread
• Limit: 10 emails (don’t overload the AI)

Module 4: ChatGPT

• Action: Create a chat completion
• Model: gpt-4o-mini (cheapest, fast, good enough)
• Prompt:

You are a morning briefing assistant. Generate a concise daily brief using this information:

CALENDAR EVENTS:
{{Google Calendar output}}

RECENT UNREAD EMAILS:
{{Gmail output}}

Format the brief as:

📅 Today's Schedule (list events with times)
📧 Priority Emails (flag any that seem urgent)
🎯 Focus Recommendation (suggest the #1 priority for today based on calendar + emails)

Keep it under 200 words. Be direct and actionable.

Module 5: Send Output

• Choose Slack or Email
• For Slack: Send to a specific channel (#daily-brief)
• For Email: Send to your personal email

Step 5: Test Everything (5 minutes)

Test each module individually:

1. Run Module 1 (Schedule) → Does it trigger?
2. Run Module 2 (Calendar) → Does it pull today’s events?
3. Run Module 3 (Gmail) → Does it fetch unread emails?
4. Run Module 4 (ChatGPT) → Does it generate a readable brief?
5. Run Module 5 (Send) → Do you receive the message?

Common issues:

• “Module 2 returned empty” → Check your calendar has events today
• “ChatGPT output is garbage” → Check your prompt formatting
• “Slack message didn’t send” → Check channel permissions

Step 6: Activate and Monitor

Once testing passes:

1. Toggle the scenario to ON
2. Set error notifications (Make.com can email you if it fails)
3. Check it runs correctly for 3 consecutive days
4. After 3 days of success, stop manually checking

Troubleshooting Table

Scaling This Workflow

Once your daily brief works, add:

• Weather data (OpenWeatherMap integration)
• News summary (RSS feed → AI summary)
• Task list (Todoist or Notion integration)
• Team updates (Slack messages from yesterday)
• Health metrics (Apple Health / Fitbit data)

Each addition takes 10-15 minutes. Build incrementally.

Cost Breakdown

Less than a dollar per month for a personalized daily brief. That’s cheaper than the coffee you drink while reading it.

Getting Started Right Now

1. Open make.com → Create account
2. Follow Steps 2-5 above
3. Test once before activating
4. Let it run for 3 days without changes
5. Adjust based on results

The hardest part isn’t building the workflow. It’s starting. Open make.com right now.

Sources: Make.com pricing (make.com), Zapier pricing (zapier.com), n8n docs (docs.n8n.io), OpenAI API pricing (openai.com/pricing)

Related Posts

• n8n AI Agents: Self-Hosted Automation Guide (2026)
• Private AI Is Becoming the New Developer Edge: Why Venice AI Fits the 2026 Shift
• MCP Explained: The USB-C of AI Connections

Frequently Asked Questions

Do I need an OpenAI API key, or can I use my ChatGPT Plus subscription?

You need an API key. ChatGPT Plus is for chatting in the browser; the API is what other tools call programmatically — different product, different billing. The API is pay-as-you-go and a workflow like this one runs for cents per month, not dollars.

What happens if the workflow breaks at 7 AM and I don’t notice?

Make.com and Zapier both email you when a scenario errors out. Turn that on before you ship. For anything mission-critical, also add a “healthcheck” — a second workflow that pings you if the first one didn’t run by 8 AM.

Can I build this with only free tools and no API costs at all?

Mostly yes. Skip the ChatGPT step and have the workflow just send you raw events + emails in a formatted message. You lose the AI summary, but the time savings of “one digest instead of three apps” is most of the value anyway.

Why does the post recommend Make.com over Zapier for the first build?

Two reasons: Make’s free tier is 10x larger (1,000 operations vs 100 tasks), and the visual builder shows you the whole flow on one canvas, which is better for learning. Zapier is fine — it’s just a less generous starting point.