Automated Document Review

Automated Document Review

Bjoern Meyer

Natural Language


Published on July 25, 2019 by Bjoern Meyer

Natural Language

Automated Document Review

The professional services network Deloitte developed an automated document review platform that uses natural language processing to recognize relevant information from documents.

Natural Language Processing

Natural language processing (NLP) is used in many AI-powered applications including virtual assistants such as Alexa, Siri and Cortana that can understand voice commands and complete tasks for a user. For documents, natural language understanding is a challenging task that helps analyzing and understanding documents in an automated way.

In order to understand the context of a document, several tasks are required:

  • Syntax recognition
    The document gets parsed and segmented into a structure of words, sentences and potential terminologies that can be automatically extracted.

  • Semantics
    This task handles the meaning of individual words or sentences and what they stand for in this specific closed or open context.

Automated Document Review Platform

Deloitte, the professional services network, developed an automated document review platform using NLP based on cognitive technologies to automatically identify semantic information.

According to the case study, Deloitte developed this tool to accelerate the process of reviewing documents which is a typical task in delivering client professional services. For an audit, one of the major businesses for Deloitte, humans have to spend thousands of hours reading contracts, meeting minutes, invoices and other relevant paperwork that help to understand the situation in order to audit a process or result.

The developed platform uses cognitive technologies that imitates human reasoning by applying pattern recognition on natural language segments. Machine learning is used to identify relevant information and key terms in documents (such as customer names, specific project names or other contextual phrases). According to Deloitte, this system is improving its accuracy over time by using learning algorithms.

The platform integrates this automated document reviewing process into existing manual processes to help reviewers implementing more effective workflows.

One team was able to increase the scope of their contract review effort by multiple orders of magnitude—processing more than 150,000 documents—using the application.

Automated document review

Document Automation Helps to Gain Efficiency

This development shows a trend in businesses such as legal services and professional services: A significant amount of work can be automated or semi-automated. The management consulting firm McKinsey estimates that 22 percent of a lawyer’s job and 35 percent of a law clerk’s job can be automated.

Dana Remus from the University of North Carolina School of Law and Frank S. Levy researched the arguments that automation will replace work performed by lawyers. A key conclusion is that "if all the technology [above] were implemented at one time, it would result in an estimated 13 percent reduction in hours".

The above numbers include all document automation processes including document creation, automated document generation processes and document management strategies.


Document automation technologies including automated document reviewing, document collaboration tools and online document editing help professionals to work more efficient and to streamline document processes.