Learning to read with Escher

Reading a form can appear a trivial task when given to a human to do. We understand titles, labels, formatting all naturally. Unfortunately tasking a person with reading many thousands of forms without making any mistakes is a different problem all together.

Enter Escher

Escher is a key part of The Compliance Workbook’s software stack. It’s the important component that allows us to support nearly any PDF format and still give accurate compliance results.

How is it different?

Unlike OCR solutions or basic PDF reading software. Escher understands the sort of document it’s reading and intelligently extracts the relevant data from it while keeping it’s context and discarding all unnecessary information like font and styling.

This makes it a perfect match for our advanced compliance calculations and reporting.

How is it Different?

Escher has evolved in requirements and functionality over the last 5 years to support new use cases as we’ve needed.

It’s core flexibility is due to our modular extraction engine that allows Escher to quickly:

  1. Decrypt / Reassemble any damaged PDFs
  2. Identify the format of document present
  3. Process all the fields found in the document
  4. Export the discovered data in JSON/XML to another system


Escher isn’t the same as the numerous Optical Character Recognition solutions available that enable you to “search” your PDFs.

Escher is used to convert your PDFs back into structured, useful data to be fed into The Compliance Workbook or other integrations.

Machine map of the brain to demonstrate machine learning