Data analysis with AI explained simply

What is it?

AI can help you understand data without being an expert. You can upload a spreadsheet or paste a table and ask questions in plain language, such as which product sold most, or what changed between two months.

The more reliable tools do not just guess numbers: they write and run code or formulas to calculate the answer, and can draw charts. Language models on their own are weak at exact arithmetic, so it matters whether the tool actually computed the result. You still need to check the data is correct and that the conclusions make sense.

Which models use it?

ChatGPT can analyse uploaded files by writing and running code.
Claude and Gemini can read data files and write analysis code or formulas.
Microsoft Copilot in Excel and Gemini in Google Sheets help with formulas and summaries.
Many business intelligence tools let you ask questions about data in plain language.

How to use it

Remove or hide personal and confidential data before uploading.
Explain what each column means and what question you want answered.
Ask the tool to show the formula or code it used, not just the result.
Request a chart and a short plain-language summary.
Spot-check a few numbers yourself in the original data.
Treat findings as a starting point; important decisions need careful review by qualified people.

Benefits

Answers data questions without writing code yourself.
Quickly writes formulas, cleans messy data and builds charts.
Explains statistics and results in plain language.
Speeds up routine reports.

Limits and risks

May calculate wrongly if it guesses instead of running code.
Can find patterns that are coincidences, or confuse correlation with cause.
Results depend on data quality; missing or wrong data gives wrong answers.
Not a substitute for professional advice on financial, medical or legal decisions.

Prompt template

Copy it, replace the parts in [brackets] and paste it into any AI chat.

I have a dataset about [topic]. Columns: [column name - meaning, ...]. About [number] rows.

Question: [what you want to find out].

1. Check the data for missing values or obvious errors and tell me what you found.
2. Calculate the answer using code or formulas, and show them.
3. Make one simple chart that shows the main result.
4. Summarise the finding in 3 sentences and list any limits or assumptions.

Data analysis with AI