Google LLC announced today that its Document AI service, which helps extract information contained in digital and printed documents using machine learning, is now generally available on Google Cloud.

Announced in preview in November, the DocAI platform is powered by underlying artificial intelligence technologies such as computer vision and natural language processing. It also leverages Google’s Knowledge Graph service to validate and enhance fields in documents, and it can be trained to create customized document models while accommodating human interaction to ensure greater accuracy where needed.

In a blog post today, Sudheera Vanguri, head of product for Document AI, said the service would be useful for millions of companies that still rely on manually entering data and use guesswork to make sense of it all.

One of the main tools in DocAI is Lending DocAI, a service that can be used by banks, mortgage brokers and other lenders to process loan applications more quickly. It works by processing borrowers’ income and asset documents using a set of specialized machine learning models. It also automates routine document reviews so that mortgage providers can focus on more important work.

Vanguri said Lending DocAI is now generally available and comes with more specialized parsers than before that can process critical loan documents including paystubs, bank statements and more. “Our goal is to provide the right tools to help borrowers and lenders have a better experience and close home loans faster,” Vanguri said.

To help borrowers adopt Lending DocAI faster, Google said it’s partnering with the mortgage servicing firm Mr. Cooper to help it deploy the service and get started.

“Over the last few years, we have made substantial investments in our proprietary servicing technology and core mortgage platform,” said Mr. Cooper Chief Executive Jay Bray. “By joining forces with Google Cloud AI, we are able to build on those advances and help make these technologies available for the mortgage industry to deploy through Google Cloud.”

Also generally available today is Procurement DocAI, a service that makes it possible to automate procurement data capture at scale. The idea is that companies can accelerate the processing of documents such as invoices, receipts and more, helping to reduce their processing costs by as much as 60%, Google claims. Procurement DocAI has already proved to be a big hit at the human capital management software provider Workday Inc., which uses the service at a massive scale in order to automate the processing of receipts with multilanguage support.

“Automating data capture is helping our customers increase accuracy and also lower their procure-to-pay processing costs,” Vanguri said. “We are continually expanding the types of documents Procurement DocAI can process — the latest is a utility parser for electric, water and other bills.”

Google said the new specialized parsers in both Lending DocAI and Procurement DocAI can be used alongside its existing AutoML Text & Document Classification and AutoML Document Extraction services that are widely used by customers to create new document models.

Finally, Vanguri announced a new feature in DocAI called Human-in-the-Loop AI, which is meant to help achieve more accuracy with document processing by enabling humans to review those processes manually. By adding human reviews into the mix, Vanguri said this will boost accuracy and help customers better interpret business predictions using purpose-built tools.

Human-in-the-Loop AI, which is also generally available from today, provides a workflow to manage human review tasks, producing a percentage confidence score based on how sure the reviewer is that the document was ingested and processed correctly.

“Document AI extracts data from documents with machine learning, and when paired with Human-in-the-Loop AI, human reviewers are able to verify the data captured,” Vanguri explained. “This system is customizable, providing the flexibility to set different thresholds and assign individual groups of reviewers to various stages of the workflow.”

Analyst Holger Mueller of Constellation Research Inc. told SiliconANGLE that AI is a driving force aimed at eliminating many of the most routine and tedious tasks that humans must perform, and that document processing is a prime candidate for that.

“Google has been innovating in this space for a while, recently with its Lending DocAI and Procurement DocAI offerings that are just going into general availability,” Mueller said. “The most interesting thing we see today is that Google wants a human feedback loop on the quality of data ingestion, which is where the new Human-In-The-Loop feature comes into play. This is key to improving the accuracy of the Document AI, and we will be looking to see how well the service can learn from human feedback.”

Image: Google