There was one task in my week I genuinely dreaded. Not hard, not important — just soul-crushingly tedious. The kind of thing that sat on my to-do list for days because every part of me resisted starting it.
You probably have one too. The report nobody reads but everyone expects. The data cleanup. The status update. The thing that makes you feel like a human spreadsheet.
I finally automated mine with AI. And the strange part wasn't the hours I got back — it was how much lighter the whole week felt once that one dreaded thing stopped looming. Here's exactly how I did it, and how you can find and kill your own version.
To automate the worst part of your job: first, pin down the exact task you dread and break it into literal steps. Most "creative" or "judgment" work is actually 80% mechanical once you look. Then hand the mechanical steps to automation and the one genuinely judgment-heavy step to an AI model. You stay in the loop only for the final approval. The relief comes not just from saved time but from removing a recurring dread from your week.
The task to automate isn't necessarily the one that takes the longest. It's the one you avoid. The one that generates dread out of proportion to its difficulty.
For me it was a recurring report. Pull data from a few places, clean it up, summarize what happened, write it in a readable way, send it. Maybe two hours of actual work. But it cost me far more than two hours, because I'd dread it for days, procrastinate, and do three lower-value things to avoid starting it.
That dread tax is the real cost. The task isn't just the minutes it takes. It's the mental weight it carries all week. Research summarized by the American Psychological Association on procrastination and unfinished tasks points the same way: the open loop drains far more attention than the work itself ever would.
So pick the task by feel, not by stopwatch. What do you put off? What makes you sigh when it appears on the calendar? That's your target.
Here's the insight that made it possible. I always told myself this report needed me — my judgment, my framing, my interpretation. That's why I assumed it couldn't be automated.
Then I actually broke it into steps. And almost none of them needed me.
Out of six steps, one truly needed human judgment, and even that one an AI could draft. I'd been treating the whole thing as sacred creative work when it was 80% data shuffling I could fully automate. This is the realistic version of what these tools do well, which is exactly the case I make in the honest truth about AI productivity tools — they're brilliant at the mechanical middle and weak at the judgment you'd never want to hand off anyway.
The tasks we dread feel like they require us. Break them into steps and most of the dread turns out to be mechanical work wearing a creative costume.
Here's roughly how I assembled it. None of it was hard once I'd broken the task down.
The dreaded two-hour, all-week-looming task became a two-minute review. And critically, I still control the final word, so nothing wrong goes out under my name. If you want to see this same three-part shape stitched into a full overnight pipeline, I broke it down in the AI workflow I built that runs while I sleep.
I expected to save time. I did. But three things surprised me.
The dread vanished, and that mattered more than the minutes. I hadn't realized how much mental space that recurring task occupied. Killing it didn't just free two hours — it freed a low hum of avoidance that ran all week. My whole week felt lighter.
The quality went up, not down. I assumed automation meant "good enough." Instead, because the AI draft was consistent and I was reviewing fresh instead of grinding through it exhausted, the reports got better. No more rushed, resentful versions sent at the last minute.
I started seeing other tasks differently. Once I'd broken one dreaded task into mechanical steps, I couldn't stop. I started noticing how much of my week was secretly automatable. That first win rewired how I look at my own job.
A few hard rules, learned partly the hard way.
Killing one dreaded task did something I didn't expect: it changed how I see my entire week. Once you've watched a thing you assumed "needs you" turn out to be 80% mechanical, you can't unsee that pattern anywhere.
I started keeping a running list. Every time I caught myself doing something tedious and repetitive, I'd jot it down — not to automate immediately, but to notice it. Within a couple of weeks the list was longer than I'd have guessed. The status update I write every Monday. The way I reformat the same kind of document over and over. The little summary I send after every meeting. None of these felt automatable until I'd proven, with the first task, that they probably were.
I didn't automate them all at once — that's the mistake that builds a fragile mess. I picked them off one at a time, in order of how much dread each one carried, and only moved to the next after the previous one ran clean for a week. Slow and boring, but each one stuck.
The compounding part is what surprised me most. Each task I removed didn't just save its own time — it freed attention I'd been spending on avoidance, and that attention went into the work that actually matters. By the third or fourth automation, my week had a fundamentally different texture. Less grinding through things I hated, more time on the parts of the job I'm actually good at and paid for. That shift started with one report I didn't want to write.
Pick the one task you keep putting off, break it into literal steps this week, and try automating just that — the dread is the signal, and following it once tends to change how you see the whole job.
Q: What if my dreaded task really does need judgment? Some judgment, sure. But break it down — you'll find most of it is mechanical setup around a small core of real judgment. Automate the setup, let AI draft the judgment, and keep final approval. The "it needs me" feeling is usually 80% wrong.
Q: I'm not technical. Can I still do this? Yes. Modern automation tools are mostly point-and-click, and the AI step is just a plain-English instruction. The hard part is breaking down the task, which is thinking, not coding. Anyone can do that.
Q: What if the automation makes a mistake? That's exactly why you keep the final approval step. The automation drafts; you catch errors before anything goes out. As you build trust over weeks, you can loosen the leash on the low-risk parts.
Q: How long did it take to set up? A few hours spread over a couple of evenings, most of it spent breaking the task into steps and testing. It paid for itself within the first two cycles and has been running quietly ever since.
I automated the one task I dreaded most, and the biggest gift wasn't the time — it was waking up to a week that no longer had that dread sitting in it.
The work you avoid is rarely as human as it feels. Break it down, and most of it is just mechanics waiting to be handed off.
The task you dread most is usually the one most worth automating. The dread is the signal. Follow it.
What's the one thing on your list right now that you keep putting off? Break it into steps this week. I'd bet most of those steps don't need you at all.
One person, output that looks like five. It isn't about working more hours — it's about a kind of leverage teams rarely have.
One idea a week to a published issue in under an hour. The boring system behind a newsletter I never dread sending.
Behind a lot of lean, profitable companies is the same small stack of AI tools. Here's what's actually running the show.
Comments
Sign in to join the conversation
No comments yet. Be the first to share your thoughts!