I think about this a lot when people talk about AI and productivity. The instinct is often to take an existing task and ask how AI can make it faster.
That can be useful.
But it can also be a trap.
If the process is messy, unclear or unnecessary, automation doesn't fix it. It just makes the mess move faster.
Before automating anything, I think there are a few better questions to ask.
Why does this process exist?
Who uses the output?
What decision does it support?
What would happen if we stopped doing it?
That last question is usually the interesting one.
Speed is not always progress
There is a strange comfort in making a bad process more efficient. It feels productive. You can point to the time saved. You can show the workflow. You can say the team is using AI.
But doing the wrong thing faster is still the wrong thing.
I've seen reports automated that nobody really used. I've seen manual checks turned into workflows when the real problem was unclear ownership. I've seen people build clever systems around work that probably should have been challenged first.
Automation is powerful, but it is not neutral. It reinforces whatever process you attach it to.
Start with the problem
The best automation starts with understanding the problem properly.
Not the task.
The problem.
If a team is spending three hours every week pulling data into a spreadsheet, the task is data collection. The problem might be that the data lives in too many places, or that nobody trusts the dashboard, or that the weekly meeting is asking for the wrong thing.
Those are different problems.
They need different solutions.
Where my thinking is today
I still love automation. I probably like building tiny tools more than is sensible.
But I'm trying to get better at pausing before I build.
Because sometimes the most useful automation is not the one that makes the work faster.
It's the one that makes the work disappear.