Document to text: Split out multiple documents from within the document

Hi. Our org receives scanned copies of paper surveys completed by email. Rather than enter them in manually, we’re keen to use a document to text action to take an upload of the scanned copy of the surveys and convert the responses into rows of data organised with individual questions answers in the fight field etc.

The document to text AI action is great. Our issue is that the document to upload actually contains multiple surveys, so at some point each survey needs to split out, to then run the workflow. So for instance we get emailed a scanned document containing 12 surveys, each about 4 pages long.

Does anyone know a decent way to achieve this, so that then the document to text extraction workflow run creates a row per survey.

(They have to be completed by paper and it might be a little arduous to request a document per survey).

Thanks in advance :slight_smile:

Have a look at this

I would suggest telling AI the breakpoints between the surveys (e.g: a survey title).

Then, extract the text and convert it to a JSON array with same rules as above.

Loop through that array in a workflow and add rows to the destination table.

1 Like

Thanks Eric!

I’d used Bob’s other video on the Document to text for the main workflow…I’ll be making use of email trigger now for sure, that will save downloading and uploading the survey.

1 Like

Excellent! I’ve updated my instructions to include the breakpoint and seem to have arrays now to loop through.

Thanks @ThinhDinh

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.