How Generative AI Helps Legal Pros Dig Out of Document Purgatory

DataStax
3 min readMay 28, 2024

By Huseyin Evecen

Founded in 2022 by a team of seasoned Norwegian legal tech professionals, Concide delivers diverse legal-tech services tailored to customer needs to streamline and enhance legal and business operations. The Concide team focuses on contract lifecycle management, legal project management, and integration hub services.

Concide’s Chat for Legal platform is designed for legal users, offering a user-friendly generative AI document Q&A chat solution and a seamless document upload feature. This feature is a game-changer for lawyers who often find themselves buried under a mountain of documents related to contracts that require review and consultation. Once a document is uploaded, users can ask questions directly related to the document’s content. The chatbot provides context-specific answers, creating an interactive and conversational user interface. This feature enables natural language processing to extract insights from contract documents and generate summaries, making the platform a valuable and easy-to-use tool for legal practitioners.

Concide built its Chat for Legal platform using the Microsoft Azure-hosted OpenAI LLM and embedding engine. To enhance chatbot responses and ensure compliance with internal security policies, Concide required control over data storage locations and access to the data. It chose Azure Document Intelligence to store uploaded documents and DataStax Astra DB to store content embeddings. Using Astra DB for its chat platform resulted in a 2X increase in search speed and a 30% reduction in costs.

Contextual understanding through similarity searches

Using LangChain Javascript and CassIO, the chatbot conducts straightforward similarity searches using end-user input to retrieve the top-10 ANN similarity search results from the stored embeddings for the respective chat thread. This process enhances the accuracy and relevance of the chatbot’s responses. It instills confidence in finding pertinent information and context within the vector store derived from uploaded documents.Azure Document Intelligence identifies structured data and text within files, storing them in Astra DB with properties such as user ID, document ID, page ID, and chat thread ID.

What’s next

Concide’s next project is a SaaS solution offering for the Chat for Legal platform. This platform is designed to streamline communication and document handling for legal professionals. Aiming to target a wider user base, Concide plans to leverage Astra DB to ensure the platform can handle large volumes of data and high user traffic without compromising performance.

In addition to the core chat functionality, Concide plans to integrate various products to enhance the platform’s capabilities. They will utilize an agentic chatbot, an advanced AI-driven assistant designed to perform complex tasks. This chatbot will employ multiple APIs to execute functions such as sending documents to a review tool, retrieving legal information, and more.

By automating these tasks, the platform will provide a smarter, multi-tool solution that increases lawyers’ efficiency and productivity. Integrating these tools aims to reduce manual effort, minimize errors, and streamline legal workflows, ultimately offering a comprehensive solution tailored to the needs of legal professionals.

--

--

DataStax

DataStax provides the real-time vector data tools that generative AI apps need, with seamless integration with developers' stacks of choice.