You may know the feeling of thinking a problem is solved while it keeps costing you.
Sometimes the fix at hand is genuinely the right one.
Other times it arrives in a stretch of rapid-fire incoming, and you implement whatever works because you need to move on.
The cost can be money, energy, or the low drag of knowing you settled. You pay it, and you move on, because that is what the day requires.
You suspect something better exists. You never check, because the thing in front of you is working, and checking takes the one resource you don’t have.
I paid Zapier thirty dollars a month for about a year. Not for one thing. For a room full of automations I had built when there was no other way to do any of it.
The room emptied slowly.
Tools got better, I learned what they could do by using them, and one by one the automations came out and got rebuilt somewhere else for free.
That went on for months.
In the end, all that was left was a single workflow moving items from a YouTube playlist into my ideas inbox in Notion.
One tenant in an empty room, and I kept paying the rent on the whole thing.
It worked, and every renewal notice annoyed me for about four seconds a month. Because it was the last thing standing between the old way and whatever came next, and I had stopped looking at it.
Then I built the agents. I gave them roles, access, and a clear job description.
And it still took me longer than it should have to have the epiphany: if you don’t know, why don’t you ask what you just created?
So I described the situation to Roger.
What the workflow did, what it cost, and what I wanted to be true at the end.
I did not tell him how.
He worked it with Conrad, who already had the way into Notion. The two of them found a route I hadn’t considered, ran it, and came back with the result.
The playlist still lands in the inbox. The thirty dollars is gone.
That’s the whole story. One pass, in minutes.
What surprised me was that I wasn’t surprised.
I had built these agents believing they could do this. That was the whole premise.
And still, when the result came back, and the thing was solved, something in me sat up.
Believing a thing can work and watching it work are not the same experience.
The distance between them is the gap I write about every week, the one between knowing and doing. I know which side I’ve been standing on.
This time I watched it close. I had a theory, a desired outcome, and something took action.
Then a pause. Not celebration.
Just standing there considering what I had handed to something outside my own mind.
I hadn’t given them a task. I’d given them a problem and let them find the answer.
That’s different from automation.
Automation is me knowing the steps and making something else perform them.
This was me knowing the outcome and nothing else, and being willing to live with a route I didn’t pick.
What I have is one small result.
Handing off the tedium gave me back the room to ask a question I had stopped asking.
Something else happened this week that I can’t prove is connected. But I suspect there may be a connection.
I have had a camera at home that stopped working, an electrical problem, small enough that I had rescheduled looking at it four or five times.
Two days after the Zapier thing, the answer showed up while I was doing something else.
I wasn’t troubleshooting.
I wasn’t thinking about the camera.
It just arrived.
I don’t know that one caused the other.
I do know I stopped holding one unresolved thing, and a different unresolved thing worked itself out.
The answer may have always been there.
It may have needed the room.
That’s the part worth watching.
Not what the agents can do, but what gets freed up in me when they do it.
Whether that scales, I don’t know. One simple workflow is not a body of evidence.
But there’s a pathway now that wasn’t there a month ago, and I’m standing at the front of it wondering what else I’ve stopped asking about.
What else is hiding in the noise if I can just let it come through?