Skip to main content

I Built 40 Automation Skills. I Use 8. Here's What the Other 32 Taught Me

A cluttered toolbox, a metaphor for automation skills that go unused
Photo by Ono Kosuki on Pexels

I've built around forty automation skills for my own work. The number sounds impressive until you look at how many I actually use: seven or eight, every week. The rest sit in a drawer, and I can't remember the last time I opened most of them.

For a while I explained that away the easy ways, I'd gotten lazy, or the task had simply gone away. But when I actually looked, there was a pattern, and it taught me more about automation than any of the ones that worked.

The Ones That Gather Dust

The skills I never use were almost all built on "this would be handy to have." Not a present need, a potential one. I'll probably want this someday, I thought. And when someday arrived, one of two things had happened: I'd forgotten how the skill worked, or reality had drifted just far enough from my assumptions that it no longer fit.

The ones I use constantly were the opposite. Each was built because I needed it right then. I made it and used it the same hour, fixed it while using it, learned its edges by living with it. The difference wasn't how well I'd engineered them. It was whether a real, repeating need had pulled them into existence, or whether I'd pushed them out ahead of one.

The Wider the Use, the Less It's Used

This one is counterintuitive but it held every time: a "just ask me anything" skill gets called far less than a "take this data and format it exactly this way" skill. The narrower the scope, the clearer the when, and the clearer the when, the faster it becomes a habit.

A too-general tool is always an option, and a thing that is always an option is, in the end, never chosen. For a skill to be the thing you reach for, what it does has to match the shape of the moment exactly. Between a tool you could use broadly and a tool you use only in this one situation, the one you actually pick up when the situation arrives is the second one. Usefulness isn't decided by design. It's decided by habit, and habit needs a repeating situation with a tool shaped precisely to it.

What You Can Automate vs What You Should

Here's the trap I kept falling into. In the middle of some task, a thought arrives: I could automate this too. The thought isn't wrong. The problem is what's actually happening in that moment: I'm not evaluating whether the task can be automated. I'm feeling the urge to automate it. And before I've checked whether the work is truly repetitive, or whether context shifts every time, or whether judgment is involved, the conclusion "I should build this" is already sitting there.

Take a task I do every week: pulling together a similar customer situation. On the surface it's repetition, same format, similar data. But the point to emphasize is different every time, the issue that matters most right now is different, what the reader expects is different. Automate it and the format stays while the context vanishes. The same document comes out every week, and nobody reads it.

That's the line I'd missed. Format can be automated. Context cannot. If you don't feel that difference, you try to automate both at once, and what you get is a tool that produces useless output faster. What you can automate and what you should are different questions, and the gap between them is exactly where the dusty skills come from.

The Thing to Throw Out First

So before building any skill, there's one belief I've learned to discard: that building it will solve something.

A skill is a tool for doing a thing you already do well, faster. It is not a tool for fixing a thing you don't yet do well. Build a skill for work you don't understand, and you've built a tool that does that work wrong, faster. So the question before the build isn't "can I automate this?" It's "am I already doing this well, by hand, without the skill?" If yes, automate it. If no, go learn to do it first. Automation is not a way to learn a job, it's a way to scale a job you've already learned.

The test I use now, before making anything, is three plain questions: Does this task actually repeat? Will it keep repeating? Is the pattern of that repetition stable enough? Only when I can answer "yes" with a straight face do I build, and those are the skills that don't end up in the drawer. The forty taught me that the measure of a good automation isn't how clever it is. It's how honestly I judged, before building it, whether it should exist at all.

This essay is adapted from one of my Korean e-books on automating my own enterprise work as a non-developer. Details are anonymized: no company, system, or tool names, only the shape of what happened.

Related reading: Eight Hours Became Two: What Automation Revealed About the Essence of My Job · The Mobile App I Decided Not to Build: On Overkill and the Maturity of Stopping

Comments

Popular posts from this blog

AWS vs Azure vs GCP in 2026: Which Cloud Platform Should You Choose?

The cloud platform decision is one of the most consequential technology choices an organization makes, and in 2026 it's also one of the most misunderstood. Most of the debate I see in enterprise architecture forums reduces to "we're an AWS shop" or "we go Azure because of Microsoft" — neither of which is a strategy. A platform choice made primarily on inertia or existing vendor relationships is a choice that will cost you for years. I've spent significant time in all three major cloud environments — AWS for scale workloads and data engineering, Azure for enterprise SAP and Microsoft-integrated architectures, and GCP for AI-intensive and analytics-heavy use cases. My goal in this guide is to give you a genuine, nuanced comparison that goes beyond feature lists and into the practical realities of choosing and running a cloud platform in 2026. I'll cover market position, each platform's honest strengths and weaknesses, how to match workloads t...

EU AI Act Compliance in 2026: What Every Enterprise Needs to Do Now

EU AI Act Compliance in 2026: What Every Enterprise Needs to Do Now The EU AI Act entered into force on August 1, 2024. The first provisions took effect six months later, and the full implementation timeline runs through 2027. If you're building, deploying, or using AI systems in or for the European Union, this law applies to you — and the window for being caught unprepared is closing fast. I've spent the past year working with enterprise clients on AI governance programs, and one pattern shows up again and again: organizations badly underestimate how much operational work compliance actually takes. It's not a checkbox exercise. It's a rethink of how you develop, document, deploy, and monitor AI systems. This guide is what I wish someone had handed me when I started — the substance of the law, the practical requirements, the deadlines that matter, and the mistakes I keep watching enterprises make. Photo by Petrit Nikolli on Pexels Photo by Karolina Gra...

I Was Keeping Passwords in My Code: A Non-Developer's First Taste of Technical Debt

While I was hastily building ten automations, I was writing passwords straight into the code. My excuse was that it was fastest. To log into a system you need an ID and a password, and if you don't put them somewhere in the code, the automation can't log in. A human just types them when the login screen appears; an automation has no hands to type. So I took the simplest road: I wrote them on the top line of the code. ID is this, password is that. It worked. And because it worked, I waved it off. What matters is that it runs — who cares where I wrote it down? It's code only I use, sitting only on my laptop. Nobody's going to open my laptop and read this file. That's how I talked myself out of it. Actually, "talked myself out of it" is too generous. I never even registered it as a problem. The question "is this okay?" never once surfaced in my head. The convenience that quietly becomes a liability The trouble with a shortcut like this is th...