Petit test
Guess first, no penalty. A manager wants Q3 revenue by region in an hour. Which part do you hand to AI first?
SQL is the language for asking a database questions. AI tools now draft SQL, spreadsheet formulas and Python in seconds. Google says Gemini in Sheets can generate formulas, tables and charts, and Microsoft's Copilot for Power BI page lists writing DAX queries (the formula language behind Power BI) and summarizing reports.
Drafting is the easy part. Getting it right is harder. In a Microsoft Research study of 200,000 Bing Copilot conversations (Working with AI), the common work activities with the lowest share of positive feedback included five data analysis activities.
Employers still expect the job to grow. The World Economic Forum's Future of Jobs Report 2025 lists Data Analysts and Scientists among the fastest-growing jobs for 2025 to 2030.
So split the work. AI drafts and explains. You own what each metric means, whether the number is true, and what gets presented.
A quick first win: the 80-line query you inherited and never fully understood. Explaining code is something AI does well, and you can check every claim against the query in front of you.
A worked example
A prompt (the message you send to the AI):
Explain the SQL query below for an analyst who didn't write it.
List: the tables it reads, every filter in the WHERE clause, every
join and what it joins on, and what one row of the result means.
Then list anything that could make the totals too high or too low.
[paste the query here]Complétez les trous
Fill the gaps.
Prompt
Explain this query for an analyst who didn't write it. List , say what , and list anything that could make the totals .
Try it
Vos étapes, environ 2 minutes
- 1.Run the prompt on one old query you use. Check two of its claims against the query itself. Start a note called "Data-check checklist": each lesson in this module adds one check.2 min