AI for Data Analysis and Spreadsheets: Ask Questions in Plain English

You no longer need to be a formula wizard to get answers from your data.

• Write formulas ("sum column C where column B is 'Paid'").
• Clean data: fix dates, remove duplicates, standardise names.
• Explain what a messy spreadsheet or formula does.
• Analyse: spot trends, outliers and correlations.
• Create charts and summaries from a table.
• Write queries (SQL) from plain-English questions.
• Automate repetitive spreadsheet work with scripts.

1. Describe the data (columns and what they mean) and your question.
2. Ask the AI to show its method: formula, query or code.
3. Test on a small sample where you know the answer.
4. Check totals against the original.
5. Document the steps so others can repeat them.

• Wrong logic that looks right: plausible formulas can still be wrong.
• Hidden assumptions: how are blanks, duplicates and dates handled?
• Sensitive data: don't paste personal or financial data into tools you haven't vetted.
• Causation: "A rises with B" doesn't mean A caused B.
• Small samples can mislead.

• "What are the three most surprising things in this data?"
• "What data quality problems do you see?"
• "What would you check before trusting this result?"

AI makes analysis accessible, but you are still responsible for the conclusion. Make it show its work, verify the numbers, and be honest about uncertainty.