Robotic Document Separator

Advanced AI that transforms a large file into multiple documents.

Without separator sheets.

Business Challenge

Scanned files often contain multiple documents within. For example, a mortgage loan file package contains up to 100 documents. An insurance claims file may contain forms, invoices, receipts, ID documents. A medical file could contain dozens of documents from patient intake forms to X-rays. A box of records might contain hundreds of diverse documents.

For a digital process to make sense of the information contained within that file, each document within it must also be separated. This is a precondition of document classification and data extraction.

Business Challenge

Scanned files often contain multiple documents within. For example, a mortgage loan file package contains up to 100 documents. An insurance
claims file may contain forms, invoices, receipts, ID documents. A medical file could contain dozens of documents from patient intake forms to X-rays. A box of records might contain hundreds of diverse documents.

Stuck in the Last Century?

Surprisingly, this document separation step is still done by humans, the same slow, expensive and mistake-prone way as it has been for over 30 years. A worker manually separates the file into documents, then inserts a separator sheet between documents so the scanning software knows the beginning and end. After the batch is scanned, the sheets are removed and re-used for the next batch. It's an incredibly tedious manual process that is slow and prone to errors. Humans get bored and miss a document, and as they tire their performance degrades. And when it comes to deciding where a document begins and ends, humans are not consistent.

Transform Your Process with AI!

The Haystac Robotic Document Separator (RDS) completely replaces the need for separator sheets, and the human activity associated with them.

RDS reads the file and determines where individual documents begin and end.
RDS separates the documents and saves them as individual files.
The results are sent to the capture workflow.
Exceptions are fed back into the RDS model so it can continue to learn.

Under The Hood
There is some serious AI at work here. RDS uses deep learning AI models that are pre-trained on tens of thousands of diverse documents. This allows RDS to capture document experience and to accumulate knowledge, just like humans do. The pre-trained models use transfer learning so they can work with your documents.

Good news: you don’t have to know a thing about AI or data science to operate RDS. We’ve hidden the magic beneath the simple user interface.

RDS is a no-code solution and you can run it in the cloud or on-premise. Easily integrated into your process with our REST APIs.