Extracting specific text from pdf file

[DNeil]

Hi all

I'm a doctor, our secretary spends countless hours sending our letters to referring dentists. 

So I'm trying to create a workflow where when a pdf file is added to a folder, certain text is selected.

This text is the name of the referring dentist.

These names are limited in number (we work with say 40 dentists who refer regularly).

This name corresponds to a certain email adress.

I could make an excel file with these names and the email next to it.

The workflow should look for the name of the dentist, and should compose a mail to the corresponding email adres, adding the pdf as an attachment.

Idealy it would also be able to extract the name and birthdate of the patient.

The name of the patient is always located in the beginning, the name of the dentist at the end, see pics.

Is this at all possible? I have just started to use Power Automate so i'm quit the noob.

Thanks in advance for any help.

Kind regards, ND

[Sundeep_Malik]

Hey @DNeil 

This is very much possible in Power Automate. You can use the combination of Power Automate and AI Builder to do this. 

For AI Builder you need atleast 5 PDFs, and then just make fields that you require to extract from that. After that tag those fields in those 5pdfs. Then train the model. From the 6th pdf onwards it will automatically detect what ever fields you need. Just there is one thing to note, all the PDFs should be of same layout/structure.

Below is a link how to use AI Builder:

https://m.youtube.com/watch?v=J3d6bx3i4l0

https://m.youtube.com/watch?v=GW5iUV2rRGQ

[DNeil]

Thanks a lot @SN, 

i was able to build the model, train and test it. It seems to work fine!

could you point me in the direction for the configuration of the following steps?

- comparing the name of the referring dentist to names in a column of an excel file

- selecting the email adres in the row next to the name

- composing an email to this emailadres, adding this file as an attachement (this i think i know how to do 🙂 )

Thanks so much for your help already.

greets, N

[Sundeep_Malik]

@DNeil 

After you have use the action of AI Builder, use a compose action, in that save the name of the person from the dynamic content you will be getting from the above AI Builder action.

After that make sure your excel sheet data is saved as a table. If not, just format the excel as table.

After this use list rows in a table action. Here you can do one of the 2 things:

1) Use filter query to just get the row where you find the name of the person. Only try this when your column name has no spaces in between or if you can change it then it would be good. In the filter query, write something like:

Yourcolumnname eq 'Output expression of above compose'

Make sure to write braces beside the expression

Or

2) If you dont want to use the above, you can use a condition action.

In condition match the column that has name and compare it to the output above.

If they match just save the email in another compose and send mail. If they dont match do nothing.

Make sure you have unique names there, otherwise, you will be sending multiple mails. Or match it to something like an id so that if two people have same name, then only 1 mail is sent.

Benefit of 1st technique:

It will increase performance of the flow. Only by 1-5 seconds, but its better to use filter query than a condition.

[DNeil]

Thanks again @Sundeep_Malik !

I'm sorry for the delay, i looked at this further with our IT department, and we seem to have figured it out.

I opted for the flow with the condition, because sometimes we send the same letter to multiple referring dentists, like an orthodontist and a prostodontist for example.

As a control, i would like the file deleted when at least 1 mail has been sent. So our office managers know they have to manually check the files that were not sent automatically. I have tried this with a condition again, but can't get this to work.

Thanks again for any assistance!

Kind regards, Neil