Skip to main content
All CollectionsWorkflows
Importing unstructured data for bulk runs (smart import)
Importing unstructured data for bulk runs (smart import)

Extract structured data from various file types

Jake Rosenthal avatar
Written by Jake Rosenthal
Updated over a week ago

How to import unstructured data for bulk runs

To run a workflow multiple times with different inputs, you can use the bulk run feature.

One way to populate the bulk run view with data is by using smart import, which can take extract data from unstructured documents, spreadsheets, PDFs, images, and messy files in general. Here's how:

  1. Navigate to the bulk run tab: In a workflow, go to the bulk run tab.

  2. Click the "Smart Import" button: At the bottom of the page, click the "Smart Import" button.

  3. Upload or drag and drop your file: Select the file you want to import from your computer, or simply drag and drop it into the designated area.

    *Below is a sample of the types of files that are accepted

  4. Click "Extract Data": After the file is successfully uploaded, click the "Extract Data" button to initiate the data extraction process.

  5. Review the extracted rows of data: Once the extraction is complete, review the resulting rows of structured data.

    • To delete a row, hover over it and click the trash icon.

    • To make edits, click into the cell and modify the content as needed.

  6. Refine the extraction (if needed): If the results are not as expected, click the "Try Again" button and provide extra instructions or context that would help the AI better detect what needs to be extracted. Then, click "Extract Data" to re-run the extraction process.

  7. Save and import the data: If the results are satisfactory, click the "Save & Import" button to import the extracted data into the bulk run view.

  8. Review the imported rows: After the import is complete, take a moment to review the data in the bulk run view and ensure everything looks correct.

  9. Run the workflow: To execute the workflow on the imported data, click the "Run All" button. Cassidy will process each row and generate results based on the workflow's configuration.

  10. Review outputs: Once the workflow has finished running, review the generated outputs to ensure they meet your expectations.

By following these steps, you can easily extract structured data from unstructured files and import it into the bulk run view, allowing you to run your workflow at scale with varying inputs derived from diverse file types.

Did this answer your question?