Invoice scanning + Excel output software

[extraT]

Snappy title, I know…

Any recommend for a very basic invoice scanning software that will then spit out an excel table?

Example:

Drag and drop invoice into software. It scans invoice and in excel I would see:

Company name
Amount owed
Due date

That’s it really, I don’t need much else.

Oh and free, if possible. Free is good.

Thanks

ET

[wyson]

Doesn’t the excel phone app allow you take a photo of tables and then ocr them automatically into an excel spreadsheet?

I saw this demoed on tiktok mind, so might not be the best source.

[extraT]

I’ll check that out! Hopefully won’t need to print the invoice out first! Thanks!

[extraT]

Sadly didn’t work as I envisioned it would

[Planet Claire]

I would try using Excel and 'Get Data' from PDF (from the Data tab), bring the data into the Data Model and analyse it that way.
If you had all your invoices in one folder and they all had the same format then you can just reference the folder and bring all the data in at once.

[paulrockliffe]

Planet Claire said:

I would try using Excel and 'Get Data' from PDF (from the Data tab), bring the data into the Data Model and analyse it that way.
If you had all your invoices in one folder and they all had the same format then you can just reference the folder and bring all the data in at once.

This is a decent starting point, but my experience is that how well it works depends on the source too much. If it's a genuine old-school PDF that's nothing more than a compressed image of the invoice, then it's not very good. If the file has a more modern genesis and has the data aspect encoded into the file, then it's really good. The problem is that if you don't have certainty on that you have to check every invoice as you go and if it's complicated, it's quicker to just type the numbers you need manually.

I started on a similar project a while ago, the idea was that if you scanned a document it would drop it into Sharepoint, Power BI would use Power Query - which is what you're talking about - to extract the PDF data, a series of rules would be used to identify what the document related to and then that would be used to get Power Automate to put it into the correct folder. Then a bespoke to the file type Power Query process would be triggered to extract it again and to extract the relevant data, building it into a Power BI dataset, both to index the files and to make the contents available for analysis.

I couldn't get it to work anywhere near well enough, the problem was inconsistency. It would work for 19 out of 20 documents, but completely screw up the 20th. There are genuine OCR options in Power Automate, tools that take the image and read it rather than hunting around in the file for tables of data in its construction but these are Premium features that cost money, so I wasn't able to test those options.