How does this affect me?
This feature enables Copilot Studio agents to generate and execute Python code to respond to user prompts. Code interpreter supports tasks such as data preparation, statistical analysis, table joins, forecasting, and chart generation. These tasks work with structured files like CSV and Excel. Code interpreter supports the GPT-4.1 model.
Using code interpreter for agents:
- Unlocks trustworthy analysis inside agents and prompts by employing deterministic, reproducible computation instead of best-effort large language model (LLM) reasoning for data questions.
- Lowers the skill barrier for advanced analytics by letting users ask analytical questions in natural language.
- Generates tables and visualizations of results programmatically that users can see, download, and reuse.
- As an uploaded file in an agent chat.
- As a referenced knowledge source in a SharePoint documents library.
What action do I need to take?
This message is for awareness, and no action is required.
If you would like more information on this feature, please visit Use code interpreter on SharePoint sources in agent conversations.