I’ve been thinking about the gap between noticing a repetitive task and actually fixing it — and how long that gap tends to be. For me it’s usually embarrassingly long. I’ll do the same fiddly thing three times a week for months before it occurs to me that I could just… not.
What finally moved me recently was a small filing-and-renaming routine that kept interrupting my focus. It wasn’t even the minutes that bugged me most. It was the context-switch every single time. Once I sat down and set it up properly, I couldn’t believe I’d tolerated it as long as I had.
That’s had me wondering how the folks here think about it, so a few open questions for anyone who wants to swap notes:
What’s your personal signal that a task has crossed the line from “annoying” to “okay, I’m automating this”? Is it a time threshold, a frequency thing, or just hitting a breaking point?
When you do fix something, what’s your usual first reach: Shortcuts, Keyboard Maestro, Hazel, AppleScript, or a purpose-built app?
Have you tried using AI (Claude/ChatGPT) to draft a Shortcut or script for one of these? I’ve seen a couple of threads here about it and I’m curious how reliable it’s been in real use versus needing a lot of cleanup.
On the other side: is there something you choose to keep manual on purpose, for privacy, reliability, or just because you don’t trust the automation not to silently break?
I’ll happily share more about my own setup if it’s helpful. Mostly I’m just curious how this crowd reasons about the tradeoff, since I suspect a lot of you are far more disciplined about it than I am.
Asian Efficiency has the 3x rule. If something annoys you 3 separate times, it’s time to fix it or automate it.
When you do fix something, what’s your usual first reach: Shortcuts, Keyboard Maestro, Hazel, AppleScript, or a purpose-built app?
All of the above. Hazel I’ve struggled to find a use case for on my home machine, except converting webloc files to URL on my desktop instantly.
Have you tried using AI (Claude/ChatGPT) to draft a Shortcut or script for one of these? I’ve seen a couple of threads here about it and I’m curious how reliable it’s been in real use versus needing a lot of cleanup.
Do this all the time for BAT/CMD files on Windows and Applescripts on Mac. AI is incredibly helpful with automating stuff…,
Less about annoyance than RoI for time saved versus time invested. A process that demands three steps but is used once a year has a low RoI to automate. A process that demands three steps but is used once or more a day has a higher RoI.
Some guideline or another exists for making such a choice on RoI. At some point, it becomes your balance. Sometimes also, the headache to automate a seemingly low RoI process may even be worth the investment for future learning.
Also, don’t automate a process that has if-then-else or switch cases. If anything, automate each if-then or switch case as separate processes themselves. For example, I have a Finder service to crop a PDF to end with 1in margins and a Finder service to crop a PDF to end with margins that I must input via a dialog box (default 1 in). I do not have one Finder service that asks me to choose one or the other (and then, if not 1in, asks me to choose a specific size margin). And I have the 1in case specifically because I use that process most often and because it runs directly without any input dialog (unlike the open-choice option).
Because I worked in I.T. a major factor was the importance of the task being completed correctly and on time. So backups, for example, were always automated.
It seems to more often be a battle between having time (now) to fix it vs just get it over with, after which the pressure to fix it is deflated. Plus the amount of time and effort to automate it.
I’ve dabbled with most of those options before. The use cases were rare and sporadic. I’ve dropped any/all subscriptions for those tools.
I would now just go to AI and find a fast, simple automation using whatever tools were easiest/simplest. Most likely not Shortcuts or AppleScript.
@WayneG — the “completed correctly and on time” lens from your IT background really resonates. It’s a different bar than pure time-saved, and I think it’s underrated in these conversations. Thanks for adding it.
I find this sort of response, from a new user, in a discussion thread very odd. The entire point of discussion threads is to share views. Why would anyone want a “no-pitch 20-min chat”? Forgive me for suspecting there is something a little odd going on here.
The purpose of automation is first and foremost to ensure accuracy and consistency in repetitive tasks. Therefore, when I notice a pattern of inconsistency and/or inaccuracy in my output, it’s time to automate it. The ROI will usually be worth it in those cases, even if the time to set up the automation far outweighs the time saved completing the task. I say “usually” because on occasion the repetitive task can be eliminated (or reduced) by reevaluating the workflow, etc.
Am I putting off doing something I need to do simply because the process involved is boring and repetitive?
Will there be more than a few edge cases that a given automation just won’t solve?
If I have to master a new tool to automate something, how many problems will doing so actually solve? Do I have to invest blood and treasure in a MacSparky Field Guide just to get one thing automated?
Is the automation tool itself a pain to use?
Would I really feel like I’d reached Boss level if I mastered python and regular expressions, or am I just kidding myself?
Questions AC (After Claude)
Me: Hello, Claude! There’s too much friction in this workflow. Help me fix it.
Claude: Sure thing! What do you need to do and what tools do you have at your disposal? Oh, and would you like me to teach you how to fish?
To be fair, it’s still worth asking a number of your questions - even with Claude. The time/friction of developing a solution has been drastically reduced with AI, but not to zero.
But the the time required to achieve MVP on a personal automation has been reduced so much that in many cases it may be worth it to simply give an idea a quick try.
Another benefit of automating a task instead of simply ignoring it is that every automation you create increases your familiarity with the available tools. As your experience grows, you become better at spotting tasks that are worth automating—and perhaps more importantly, tasks that you hadn’t realized could be automated in the first place.
Usually I go in bursts. I just forget about it entirely for a while, then suddenly I get inspired and try and automate 10 new things in a row. That said I’ve been automating like crazy with Claude recently.
