My biggest mistake learning AI automation

When I started learning automation, I focused too much on tools.
“Should I learn n8n?”
“Should I learn Make?”
“Which AI model is best?”
But the bigger question is:
“What problem am I solving?”
A simple automation that saves a company 5 hours/week is more valuable than a complicated AI agent nobody uses.
My current approach:
Find a repetitive problem
Design the workflow
Add AI where it creates value
Improve based on feedback
Tools change quickly.
Problem-solving skills stay.
What was your biggest lesson when learning automation?