What AI Can (and Can't) Do With Your Data

Spreadsheets intimidate a lot of capable people. Rows, formulas, pivot tables — it feels like a specialist skill. AI quietly lowers that barrier: you can now ask questions about data in plain English. But you have to know exactly where the tool is strong and where it will quietly lie to you.

Where AI shines with data:

• Explaining what you're looking at. Paste a confusing table or chart description and ask "what is this actually telling me?"
• Summarizing trends in words. "Here's our monthly sales for the year — what's the story?" It's genuinely good at turning numbers into narrative.
• Suggesting what to look at. "What questions should I ask of this data?" / "What might explain this dip?"
• Writing the formula for you. "I need a spreadsheet formula that adds up column B only where column A says 'paid'." Describe the goal; it writes the syntax.

Where AI is dangerous with data:

• Doing the actual arithmetic in its head. When AI "calculates" by predicting text, it can get sums, averages, and percentages wrong — confidently. This is its single biggest data weakness.
• Anything that must foot exactly. Financials, payroll, invoices, totals a decision rides on.

The dividing line: trust AI to interpret and explain data. Don't trust it to compute the numbers. Let the spreadsheet do the math; let AI tell you what the math means.

▶️ Try this

Paste a small set of numbers you already know the total of (say, five expenses) and ask AI to add them. Check its answer against the real total. Whether it's right or wrong this time, you'll never again blindly trust AI arithmetic — which is exactly the instinct that keeps you safe.