Overview of Document AI

Document organizing and understanding platform by Google

Documents are everywhere. Emails, blogs, reports, receipts, bills, etc. But can you use them as is to build some processes, calculations, applications? For example, you have a directory of purchase-receipts, can you build an app to query all those receipts with total less than $100. You can’t. Because those receipts are just scanned images, right? Such data is called ‘unstructured’ data, as it is without any query-able structure or schema. Opposite to that, ie ‘structured’ data, is arranged as, say, key-value pair or in a tabular form. Once data becomes ‘structured’ it becomes usable in downstream applications, calculations, predictions, etc.

[Are you looking to learn more about Google Cloud Document AI? Check out Processing W2 & Payslips is now even simpler with Document AI.]

Document AI by Google is one such platform, that stores documents and helps understand them as well, ie. it makes ‘unstructured’ data usable by making it ‘structured’. Following sketchnote gives an overview of Document AI.

With platforms like Document AI, it is possible to unearth the hidden information, the gold. Once data becomes structured, then its application can be numerous, sky is the limit.