Generative AI in Legal: Customer Stories

In this article, I am sharing 5 customer stories of how both established enterprises and startups used Generative AI for various legal use cases. This is the 3rd article in the Generative AI in Legal series. The previous two articles are here: 1/ Generative AI in Legal: Solution Providers 2/ Generative AI in Legal: Use Cases.

Story 1: Lexis Nexis - Anthropic - AWS: Multiple use cases

LexisNexis is an information and analytics company that provides solutions to legal professionals to help them win cases and increase their business. They looked at four different use cases for their customers - 1/ Helping their customers with a conversational search to find the information that they are looking for 2/ Summarizing the information so that their customers can quickly get to the answers 3/ Drafting an initial draft of a memo, an email to a client, a brief that's going to be used in court. 4/ Analyzing documents, summarizing those and allowing its customers to ask questions about the information in the document that they provided. Lexis Nexis fine-tuned their model using a process called Reinforcement Learning with Human Feedback (RLHF) on Amazon Bedrock. Within Amazon Bedrock, they use four different models from Anthropic Claude because their solution is based on a multi-model approach. They select the best model for the specific use case that they have at the best price.

Over the course of a year, about 300,000 documents come out and Lexis Nexis would review about 20% of those manually through legal editing process. With these fine-tuned models on Amazon Bedrock, they processed 300,000 summaries in an automated way in about six weeks. Their customers are very enthusiastic about these use cases because they see the potential power in these type of solutions, the ability to make them more productive and the ability to provide high-quality work product to their customers.

Story 2: Robin AI - Anthropic: Contract Copilot

Robin AI launched Contract Copilot to help legal professionals find answers and information faster with its dedicated AI assistant. The Contract Copilot can be used across different stages of the contract lifecycle from drafting to reviewing and searching. Robin AI operates a unique hybrid model that combines Anthropic's Claude LLMs with its own proprietary contract data (from over 2 million contracts) and machine learning techniques to read and understand contracts. An in-house team of legal professionals, or lawyers-in-the-loop, helps to further refine the model to ensure the highest levels of quality and accuracy. Robin AI has fully integrated Anthropic’s Claude 3 Haiku model to power the chat functionality of its contract copilot. Leveraging Anthropic Claude's advanced capabilities, Robin AI have successfully reduced the time it takes to complete crucial task of contract finalization and signing process by up to 10 times.

Story 3: Tekmir - Innovative Solutions - AWS: AI-powered case management

Tekmir, a legal tech startup focused on AI-powered case management platform, integrated AWS partner Innovative Solutions’ Tailwinds solution powered by Amazon Bedrock and Anthropic Claude into their lawsuit management platform. Tekmir is looking to cut case management time by 60%, streamlining inbound logistics and allowing attorneys to spend their time on core legal work. Tekmir also aims to reduce prosecution costs by 50% and potentially saving multi-million $ for large law firms. With this integration, Tekmir will specialize in automated validation of class action lawsuits, enhanced decision-making, and empowering legal professionals to focus on higher-value tasks. Tekmir customers will observe a remarkable improvement in staff productivity through automation of repetitive tasks, reduction in case prosecution costs resulting in financial efficiency and accelerated deployment timelines.

Story 4: Nebraska Judicial Branch - ScaleCapacity - AWS: Information retrieval for case related questions

Nebraska Judicial Branch is a vital component of the state’s government for interpreting and applying the law to resolve disputes and uphold justice. More than 180 courts compose the Nebraska Judicial Branch, including district courts, county courts, separate juvenile courts, the Nebraska Workers’ Compensation Court, the Nebraska Court of Appeals, and the Nebraska Supreme Court. The Nebraska Judicial Branch operates under the administrative direction of the state’s Supreme Court. Nebraska trial courts handle more than 285,000 cases annually.

The Nebraska Judicial Branch worked with AWS partner ScaleCapacity on a proof-of-concept (POC) to add semantic search-based query capabilities. ScaleCapacity leveraged Amazon Bedrock and Amazon OpenSearch Service to build the semantic search capabilities to allow users to quickly search for content that may be spread across several documents, videos, and audio files. The application uses large language models (LLMs), OpenSearch Service, and the LangChain open source framework to provide semantic search functionality. Previously, users (including attorneys and court personnel) would have to review all case exhibits to determine answers to case-related questions, a process that could take several hours to a few days. With the generative AI based semantic search, users can ask a question and receive answers within seconds.

Story 5: Utility Company - Sia Partners - AWS: Service Agreement Generation from Contracts

AWS partner Sia Partners worked with a utility company to develop a solution to automate the retrieval of information from this client’s contracts to establish a service agreement. This was based on semantic search through documents and with the use of their SiaGPT tool to extract information within documents using services such as Amazon Bedrock, Amazon SageMaker and Amazon Textract. The solution consisted of the implementation of a user interface where the user can upload files representing a contract, to generate the associated service agreement. The solution was successful in automating extraction for all file types (scan or pdf contract) and contract types.

Source: 1/ LexisNexis Leverages Generative AI amp; Amazon Bedrock to Boost Productivity amp; Transform Legal Services 2/ Robin AI 3/ Tekmir Success Story - Innovative Solutions 4/ Nebraska Judicial Branch modernizes its Electronic Exhibits System using AWS 5/ SiaGPT: Generation of Service Agreements from Contracts

Note: If there is a success story that I am missing, feel free to reach out and I will be glad to add.