5. Garbage In, Garbage Out

This is the simplest problem and the most important.

If your input data is messy, an automation doesn't clean it—it processes it. It processes it faster and more consistently than a human would, which means it creates garbage at scale.

Example: You automate "copy all customer emails into our CRM." But half the emails have the customer name, and half don't. Half have a phone number, and half don't. Your automation dutifully creates records with missing data. You now have a half-complete CRM that's worse than useless.

Where to draw the line: Before you automate anything, look at the data. Is it structured? Is it clean? Does it have what you need? If the answer is "kind of" or "sometimes," you need a data-cleaning step in your workflow, or you need human review before the data goes into the final destination.

The automation that works: Extract → Validate → Clean → Process → Store. Most people skip the validate and clean steps. Don't.