Applying Artificial Intelligence on Legal Document Review

Today’s market software is powered by artificial intelligence (‘AI’) and already carries numerous legal tasks. This will continue to impact autonomous legal departments and law firms as time progresses. For example, reviewing documents, conducting legal research and electronic discovery.

AI has been adopted to expedite traditional methods and this functionality provide completeness when using predetermined data models to predict case outcomes. AI can look for additional documents and sorting through faster than the human eye.


This whitepaper will focus specifically on document review.


Document Review


During litigation, document review requires searching through mountains of documents containing specific keywords or evidence, including emails, call logs, voice notes and graphics. These are contained across multiple cloud storage drives, software platforms, hard drives, devices (mobile phones and laptops) and archives.


Microsoft AI can achieve this by identifying fraud, topics and keywords in discovery documents, and predictive text patterns. For example: defendants communications. This benefits lawyers, barristers and paralegals from spending too much time on irrelevant documents.


Microsoft Azure Text Analytics


Microsoft Azure Text Analytics resides within Azure Cognitive Services language pillar and contains different features which can aid your legal document review. These are:


Key Phrase Extraction which is useful in identifying topics, relevant words, phrases and text in documents and document collections.


Recognising text from over 90 different languages mentioned within documents.


Moreover, we’ve seen a steady pace in Sentiment Analysis which detects expressed or implied sentiment about a legal matter within the context of a legal setting. This is by mining the text for clues about positive or negative sentiment using sentiment labels based on the highest confidence score found by the service at a sentence and document-level.


Sentiment Analysis is accompanied by opinion mining which is the process of extracting human thoughts and perceptions from unstructured text to tune legal discourse within the legal realm. It is particularly useful in understanding whether an expressed opinion appears to be fact based or might be construed as non-factually based.


Computer Vision


Artificial Intelligence goes beyond the language pillar. Using the Artificial Intelligence Vision pillar to boost content discoverability, automate text extraction and analyse video in real time. Use visual data processing to label content with objects and concepts, extract text, generate image descriptions, moderate content and understand people’s movement in physical spaces.

Specifically, to document review, you will be able to extract printed and handwritten text from images and documents with mixed languages and writing styles.



Follow us on LinkedIn for more interesting content on document review as part of this series or call us on 0208 057 7558 for any corporate or legal IT challenge.