Most teams I talk to want AI for their data problems. They think it will magically understand their spreadsheets.

Meet Tom, a gym manager tracking member trends.

His monthly routine:

• Pull attendance data from the door scanner

• Match it with membership payments

• Spot who stopped showing up but still pays

• Find patterns in cancellations

• Write notes on what's working or failing

Tom fed all this into ChatGPT, thinking it would save him some time. It spit out fancy charts and insights that sounded smart but felt off.

The AI said attendance was "seasonally declining due to weather patterns." But Tom knew it was because the parking lot flooded twice and members couldn't get in.

When the owner asked why attendance went down, the AI (and now Tom) had the wrong answer. The AI made assumptions he couldn't track.

Here's what worked instead:

• Simple Airtable linked members to their check-ins

• Formulas flagged anyone missing for two weeks

• Rules marked payment issues automatically

• Notes connected directly to specific events

Now Tom sees why numbers change. Low attendance in March wasn't weather. It was the flooded parking lot and broken heater.

The impact: Three hours became one hour. More importantly, Tom could explain every trend instead of guessing.

AI works great for SOME things, but it’s a black box. For data where assumptions and tracking matters, you want simple rules you understand.

Still keeping it simple,

Jamie