More AI tools will not fix this

Businesses don't have a shortage of AI ideas. They have a shortage of good decisions about what to automate.

So they do one of two things. Nothing. Or the wrong thing, expensively.

Here's the filter I use before I touch a single tool. A workflow is worth automating when it's:

• Repeatable. The steps barely change.
• Frequent. It happens often enough to matter.
• Rules-based. You can describe how the decisions get made.
• Slow or error-prone by hand.
• Low-risk if it's occasionally wrong.

Score each one, one to five, on all five.

High across the board? That's your first project.

Low on that last one? Keep a person in the loop. Don't automate anything where a wrong answer is expensive.

Notice what this quietly rules out. The exciting stuff. The judgment-heavy, one-off tasks that make a great demo and a fragile system.

The best first project is almost always boring. The weekly process nobody enjoys, that never changes, that quietly eats hours.

That's not a limitation. That's where the time is hiding.