Using New Outlook on Windows for personal email, calendar, and contact management
Hello Mekala,
Thank you for reaching out regarding your issue with the AI Builder model and extracted entities.
It can be frustrating when the data you expect to see isn’t appearing. Here are some steps and considerations to help you troubleshoot and resolve the issue:
Troubleshooting Steps:
Verify Model Output:
- Ensure that your AI Builder model is indeed extracting entities correctly. You can do this by testing the model with sample inputs and checking if it returns the expected output.
Check the AI Model Configuration:
- Ensure the Model is Trained: Confirm that the AI Builder model is fully trained and published. Sometimes, if the model is still in a training state, it may not return results.
- Validate the Training Data: Check that the documents used for training included all expected entities and that they were accurately labeled. If not, retrain the model with more representative data.
Verify Extraction Steps:
- Review the Extraction Process: Double-check that you’re using the correct action in Power Automate to extract entities from the documents. Make sure you’re pointing to the right model and providing the necessary inputs (e.g., the document text).
- Output of the Action: After the extraction step, add a Compose action or a Send an Email action to output the extracted data. This way, you can check if the entities are being recognized but not displayed in the expected format.
Inspect the Data Format:
- Sometimes, the data might be present but not in the format expected for the Excel export. Check the output structure from your AI Builder extraction to ensure the entities are correctly formatted (e.g., as an array or a string).
Debugging the Flow:
- Use the Flow Run History feature in Power Automate to see the outputs at each step of your flow. This can help identify where the breakdown occurs.
- Look for the specific output of the AI Builder action to verify whether the entities are being extracted successfully.
Output the Data:
- If you confirm that the data is being extracted but not making it to Excel:
- Ensure that you’re correctly mapping the outputs to the columns in your Excel table. The entity names should match the table headers in Excel.
- If the entities are nested (like JSON objects), make sure you’re extracting the specific fields you want to map into your Excel sheet.
Permissions Check:
- If your Excel file is stored on SharePoint or OneDrive, ensure that you have appropriate permissions to access and modify it.
- Sometimes flows may fail due to permission issues when accessing files created by other users
Additional Suggestions:
- Testing with Sample Data: Try running your model against a known sample document where you are certain the entities exist. This can help verify if the issue lies with the document content or the model itself.
- Consulting Logs or Documentation: Review the logs or the documentation for the AI Builder for any specific requirements or known issues related to entity extraction.
- Community Resources: Don’t hesitate to check the Microsoft Community forums or documentation for similar cases; you might find additional insights from others who faced similar challenges.
Additionally, you can refer to these Microsoft articles;
Use flows with Excel for the web - Power Automate | Microsoft Learn
Extract information in Excel using Power Automate - Azure AI services | Microsoft Learn
Use the ID reader prebuilt model in Power Automate | Microsoft Learn
If you've tried these steps and still face issues, please provide more specific details about your flow setup or any error messages you might be encountering. This will help us narrow down the problem further.
Please understand that our initial response may not always resolve the issue immediately. However, with your help and more detailed information, we can work together to find a solution.
Appreciate your patience and understanding and thank you for your time and cooperation. Have a great day!
Sincerely,
Microsoft Community Moderator