On this page
Google Document AI: document processing features and setup
A Google Cloud service for building document processors. It combines extraction, classification, and splitting to turn unstructured documents into information that can feed a business workflow.
Tool by the Otomatzia editorial team · source content checked on
Overview
Document AI uses processors selected for specific tasks, so it is useful to distinguish document type recognition, packet splitting, and field extraction. A packet may contain a form, an invoice, and an attachment, and it is not advisable to send them all to the same output schema without checking for a fit. Processor selection also includes a region and version, which should be documented alongside extraction test results.
Features
- Processors for data extraction
- Combined document classification and splitting
- Create processors through an interface or API
Business uses
- Sort a document packet before an operations team handles it
- Extract custom fields from a business document
Getting started
- Choose a suitable processor and prepare a set of documents with expected answers.
- Run a packet containing two document types and check that classification or splitting keeps the right pages together before they move to field extraction.
- Compare results on documents that were not used to configure the processor, and record the language, version, and region so you can retest the same configuration after a change.
What to check
- Check languages, regions, and versions for each processor individually, rather than for the service in general.
- Compare processing costs and extraction quality against documents that were not included in the initial configuration.
Sources to check next
Vendor features and prices change. Confirm details on the linked official pages before you rely on them.