I’ve seen a few people tripping over this lately, so I wanted to share a quick breakdown of why this happens and how to fix it.
If you’ve been running into issues where your Excel files upload successfully to a Copilot Studio agent, but the agent claims it “can’t see” any files or you end up with a generic SystemError, you’re not doing anything wrong. This is a fairly common hurdle and comes down to how Copilot Studio agents handle structured data compared to “standard” Copilot.
Why this happens
Even though .xlsx is listed as a supported format, Copilot Studio’s default runtime is optimized for unstructured text (like PDFs and plain documents). When it encounters an Excel file, it often fails to properly attach the workbook’s contents to the model context.
When that happens, the model genuinely thinks no file was provided, which is why you see messages like “nothing is uploaded”. After a few retries, the backend may time out and surface a generic SystemError.
How to resolve it
Primary fix: Enable Code Interpreter
In your agent configuration, make sure Code Interpreter is turned ON.
Despite the name, this isn’t just for running Python. Enabling it allows the agent runtime to correctly open, parse, and reason over Excel workbooks. Without it, Excel files often resolve as “empty” to the model.
Most reliable workaround: Use CSV
If your data doesn’t require multiple sheets or complex formatting, exporting the file as CSV is the easiest workaround. CSVs are treated as plain text and are almost always ingested immediately without issues.
Optimizing .xlsx files (if you must use Excel)
If Excel is required, keep the workbook as simple and “clean” as possible:
- Format the data range as a Table (Ctrl + T)
- Use a single worksheet
- Ensure the first row contains clear headers
- Remove macros, pivots, and complex formulas
Bottom line
If your Copilot Studio agent feels “blind” to Excel uploads, turning on Code Interpreter is usually the magic switch that fixes it. Otherwise, CSV is still the most dependable option.
✅ If this answer helped resolve your issue, please mark it as Accepted so it can help others with the same problem.
👍 Feel free to Like the post if you found it useful.
