Google Document AI: machine learning to extract valuable information from your documents

Google Document AI

Document AI is a Natural Language Processing (NLP) solution from the Google Cloud Platform based on Vision’s optical character recognition (OCR) technology. This new Google tool can identify text, images and characters in over 200 languages.

What can we do with Google Document AI?

According to official Google documentation, with Google Document AI we can:

  • Make better decisions based on document data and make them available to applications and users.
  • Automate and validate data to make workflows more efficient
  • Use the data in the documents to obtain new and valuable information about our products and to meet the expectations of the customers.
  • All backed up by Google’s security model and global infrastructure to keep our organisation safe.

Main features

The main features of this software are the reading of documents and their interpretation.

Google Cloud’s Vision optical character recognition (OCR) and form analysis technology uses deep-learning neural network algorithms to recognise text, characters and images.

The Google Cloud Natural Language products employ the same deep learning technology on which Google’s Search and Wizard are based. Thanks to this, you can extract valuable information from unstructured documents.

Advantages of Google Document AI

Google Document AI aims to use the data extracted from our documents to enable us to make quicker decisions, extracting structured data from unstructured documents so that it is easier for our company’s applications and users to process it and thus improve operational efficiency.

Thanks to this we can ensure the accuracy of our data and that it complies with regulations, as well as automating and validating all our documents, which optimises our workflow.

Finally, Google tells us that we can use the valuable information we obtain to meet customer expectations and increase their spending, as well as improve their satisfaction, support and lifetime value.

Use cases

Speed up the processing of mortgage documentation

Document AI’s Lending is a specific solution for the mortgage market, which processes property and income declarations to speed up loan applications. Document AI lending uses specialised models to analyse the documents requested when processing mortgage loans, in addition to automating many of the routine documentation checks.

Automate the capture of procurement data at scale

Document AI Procurement allows companies to automate the procurement cycle, one of their highest volume and value processes. Google Cloud uses Artificial Intelligence to make an analysis. We can apply this analysis to different formats, such as invoices, receipts, etc. to obtain structured and clean data.

How much does Google Document AI cost?

Document AI is intended to be used with other Google Cloud products. Google recommends integration with other services, such as Cloud Vision, the Cloud Natural Language API or AutoML Natural Language.

However, general information on the prices of the forms analyser and the OCR document processor is included. The price is 65 USD per 1000 pages for the form analyser. For the OCR document processor, the price is 1.50 USD per 1000 pages. Google offers discounts if we exceed 1 million pages per month.

Do you want to try?

We can test this new tool on the official website with our own documents, such as an invoice, and consult the structured data obtained by the application. You can also use the ones offered by Google as an example, if you don’t have one at hand.

Click here to try it out.