How to Build Your First Automation Without Coding: A Real Example That Saves Me an Hour a Week

Every guide to automation opens with someone claiming they automated their entire life. Mine opens smaller: I automated one annoying task, it took under two hours, and it has quietly saved me about an hour every week since. That modest version is the honest on-ramp, and picking a project that small is the actual secret to your first automation surviving.

Picking the right first project, which is most of the battle

Before the walkthrough, the part every automation guide skips: most first automations die not because the tool was hard but because the project was wrong.

A good first automation has three properties. It happens on a trigger you can name precisely (“an email arrives from X”), it involves moving or transforming information rather than making decisions, and it is annoying enough that you will notice the relief. Mine was newsletter triage: I subscribe to a handful of weekly newsletters, I want the links out of them without the emails colonising my inbox, and I was doing this by hand every Sunday like it was 2015.

Bad first projects, from experience and from watching others: anything involving judgement (“reply to important emails”), anything with more than three steps, and anything where failure has consequences. Your first automation should be one you could abandon without pain, because the goal is learning the shape of the thing.

The build

Step 1: Say the automation out loud before touching any tool.

Mine: “When an email arrives from these three senders, pull out the links, add them to a list, and once a week send me the list.”

If you cannot say your automation in one sentence with a clear WHEN and a clear DO, it is not ready to build. This sounds like advice from a fortune cookie. It is the single thing that most predicts whether the build takes one hour or four.

Step 2: Set up the trigger in Zapier.

I used Zapier because it is the gentlest of the automation tools for a first project, with the largest library of pre-built connections. Alternatives exist and I will compare them properly in another article. Make, for instance, is more powerful and noticeably steeper.

In Zapier you create a “Zap”: a trigger plus one or more actions. My trigger: “New email in Gmail matching a search.” The search string took a couple of attempts, which is where my lost half hour went. My first filter matched too broadly, and the automation cheerfully processed a receipt, a shipping notice, and a message from my mother. The fix was tightening the Gmail search to specific sender addresses rather than keywords. Lesson: test your trigger filter with your real inbox before connecting anything downstream.

Step 3: Add the action.

Second half of the Zap: “Create item in Notion database.” Zapier walks you through connecting your Notion account and mapping fields: the email subject becomes the item title, the links land in a text field, the date fills automatically. This part is genuinely just filling in a form. No moment of it requires anything resembling code.

Step 4: Test with one real email.

Zapier has a test button that runs your automation once on real data. Mine worked on the second try, after the filter fix. When the test item appeared in my reading list, unprompted, built by a thing I had made: I will not pretend there was no small fist pump. There was.

Step 5: The weekly digest.

A second, even simpler Zap: every Sunday evening, send me an email listing what landed in the reading list that week. Trigger is a schedule, action is an email. Ten minutes.

Three weeks later: the honest report

It has run for three weeks without attention. Roughly 60 of Zapier’s 100 free monthly tasks consumed. Two newsletters changed their sending address at some point, quietly stopped being captured, and I only noticed a week later. That is the real shape of automation maintenance: not dramatic breakage, but silent drift you discover late. A once-a-month glance at the Zap’s history log is enough.

The honest accounting: about an hour saved per week, minus perhaps five minutes a month of checking. The maths works. More than the maths, though, the mental shift is the actual product. Once one automation exists, you start seeing your week differently. Every repeated task acquires a small question mark over it.

Where to go from here

Resist the urge to automate everything at once. Run one Zap for a month. Learn its failure modes. Then add a second. The people whose automation systems survive are the ones who built them one boring piece at a time.

And when you hit the ceiling of what Zapier’s free tier can do, that is a genuine milestone: it means you have automated enough to have opinions. The paid tiers and the more powerful tools will still be there, and by then you will know exactly what you need from them.