A working picture of a language model
A large language model predicts likely next text from patterns in its training. It is useful for drafts and structure. It is not a database, and it does not know your company’s numbers unless you give them in the prompt or connect a tool that does.
Analyst uses that stay honest
Keep the model on language and structure. Keep the numbers in SQL or your BI tool.
- Ask it to outline a query, then run the query yourself.
- Ask it to turn a result table you paste in into a short brief.
- Ask it to list caveats you might have missed, then check each one.
- Do not ask it to invent a figure you have not calculated.
Where this sits in the Data Science Course
Gen AI is taught after you can evaluate an output. The point is to speed up work you can already verify, which is the standard teams expect from an analyst who uses these tools.

