Document AI, also known as Document Intelligence, is a field of technology that employs Machine learning techniques, such as natural language processing, to develop computer models capable of analyzing documents in a manner akin to human review. Through natural language processing, computer systems understand relationships and contextual nuances in document contents, facilitating the extraction of information and insights. This technology also enables the categorization and organization of documents themselves. Applications extend to processing and parsing a variety of semi-structured documents, including forms, tables, receipts, invoices, tax forms, contracts, loan agreements, and financial reports.
Document AI combines Artificial intelligence methods with document-specific data structures. It leverages Deep learning architectures to handle both the textual content and the spatial layout of documents, enabling tasks that range from simple data extraction to complex document understanding and generation.
Key Features
Machine learning is utilized in Document AI to extract information from both printed and digital documents. The technology recognizes images, text, and characters in various languages, aiding in the extraction of insights from unstructured documents. Use of this technology can improve the speed and quality of decision-making in document analysis. Automation of data extraction and validation contributes to increased efficiency in document analysis processes. Since the early 2020s, integration of Large language models has extended Document AI beyond extraction toward generative tasks, including automated drafting of forms, contracts, and document summaries.
Data Dimensions and ML Architecture
Data is typically distinguished into spatial data and time-series data. Spatial data includes images, maps, and graphs, while time-series data includes signals such as stock prices or voice recordings. Document AI combines text data, which has a time dimension, with other types of data, such as the position of an address in a business letter, which is spatial.
Historically, spatial data was analyzed using a Neural network architecture like a convolutional neural network, and temporal data using a recurrent neural network. With the advent of the dimension-type agnostic Transformer (architecture) architecture, these two different types of dimensions can be more easily combined. Document AI is an example of this integration, as transformers process sequences of tokens while also incorporating positional encodings that capture layout information.
Example: Business Letter Analysis
A business letter contains information in the form of text, as well as other types of information, such as the position of the text. For instance, a typical letter contains two addresses before the body of the text. The address at the very top, sometimes aligned to the right, is the sender address. This is normally followed by the date of the letter, with the place of writing. After this, the receiver address is listed.
The distinction between the sender address and the receiver address is conveyed solely by the position of the address on the page; there is no textual indication like "Sender:" in front of the addresses. Document AI systems must learn to use this spatial information to correctly classify and extract these fields, demonstrating the importance of combining textual and layout data.
Benchmarks
Several public datasets are used to evaluate Document AI systems. FUNSD (Form Understanding in Noisy Scanned Documents) contains 199 annotated forms with token- and block-level labels for form understanding tasks. CORD (Consolidated Receipt Dataset) supports key information extraction from receipts. DocVQA contains approximately 50,000 questions over 12,000 document images for layout-aware visual question answering. These benchmarks help researchers compare the performance of different models on tasks such as key information extraction, form understanding, and document-based question answering.
Common Uses
Document AI systems automate document processing and information extraction in business and financial workflows, including invoice and receipt processing, data entry automation, anomaly detection, mortgage processing, loan portfolio monitoring, credit risk management, and fraud detection such as counterfeit currency and fraudulent checks. They are also applied in regulatory compliance and contract analysis, including assessing changes in legal and regulatory documents. In real estate, Document AI supports document classification and structured information extraction for standardized processing and analytics. With the adoption of Generative AI, Document AI systems can also generate and pre-fill structured documents such as contracts or business forms from natural language prompts.