Why Definely chose LangGraph for building their multi-agent AI system

Why Definely chose LangGraph for building their multi-agent AI system

See how Definely used LangGraph to design a multi-agent system to help lawyers speed up their workflows.

3 min read

Definely is revolutionizing the legal industry by integrating AI-driven solutions into lawyers' daily workflows. Focused on enhancing efficiency and accuracy, Definely is building specialized agents that operate within Microsoft Word, allowing lawyers to streamline contract drafting, negotiating and review. With a goal of building the most effective infrastructure for legal workflows, Definely chose LangGraph to flexibly design a multi-agent system that could help lawyers solve complex legal tasks.

Key Problem: Time-consuming workflows for lawyers

Legal professionals often deal with high volumes of contracts, legal research, and document organization — tasks that are time-consuming and prone to human error. Previously, Definely utilized a retrieval-augmented generation (RAG) chatbot built on LangChain, a solution that primarily answered questions about contract content and was limited in the range of tasks it could solve.

To address challenges in contract drafting, review and negotiation workflows, Definely is building an advanced agentic system composed of multiple specialised agents. Engineered to tackle a broader range of complex legal tasks that lawyers need to complete, this architecture is designed to:

  • Plan: Break down intricate legal challenges into smaller, manageable subtasks.
  • Solve: Deploy dedicated agents that work together to complete each subtask.
  • Adapt: Dynamically revise the approach as tasks evolve.
  • Interact: Incorporate human input and approval at key moments to guide decisions.

Using LangGraph to build a customized multi-agent system

Definely adopted LangGraph to build a sophisticated multi-agent system that enables lawyers to collaborate seamlessly with AI agents for various legal tasks. Definely’s decision to adopt LangGraph was driven by its graph-based architecture, which enabled the creation of an adaptable, highly customizable multi-agent framework that mirrors the way lawyers handle daily tasks. Its non-opinionated and flexible design makes it especially effective for complex fields like law, where integrating extensive domain knowledge and bespoke information is crucial. This approach empowers legal professionals to delegate specialized tasks to AI agents while retaining full oversight and control over the final outputs.

With LangGraph, Definely’s AI system can:

  • Extract key contract information: Automatically pull relevant details from contracts, including definitions, clauses and other kinds of contract data, accurately and reliably. This allows the lawyer to better understand the terms, obligations and commitments in a contract.
  • Analyse consequence of changes to contract: Analyse the consequence of each change made to a contract. This allows the lawyer to understand whether any changes to a contract have flow-on implications for other parts of the agreement.
  • Answer complex legal queries: Answer questions a user has across multiple contracts, ensuring that every pertinent aspect is considered. This allows the lawyer to better understand their contracts across multiple documents.
  • Draft and enhance contract language: Assist in drafting, reviewing, summarizing, modifying and enhancing contract language without losing context - including drafting new clauses tailored to your needs and based on the lawyer’s prior work. This allows the lawyer to draft and review their contracts with speed, accuracy and high quality.
  • Seamless system integration & document management: Integrate with your law firm’s document management system to access gold standard clauses, unique precedents, deal data, and historical knowledge enhancing decision-making, clause drafting, and responses to legal queries. This allows the lawyer to draft and review contracts in line with the firm’s practices and prior deals, and maintain the firm’s high quality standard.
  • Contract mark-up: Automatically mark-up legal teams’ contracts according to their guidelines, preferences, and prior deals. This allows the lawyer to draft and review contracts with high consistency, efficiency, and the relevant context.
“LangGraph’s flexibility has allowed us to build AI agents that mirror how legal professionals approach their work. This is a game-changer for lawyers and the industry.” –  Sigurjon Isaksson, Head of AI at Definely

Conclusion 

Definely is poised to redefine how legal professionals interact with AI to augment workflows where accuracy is paramount. The next phase of development will focus on expanding the system’s capabilities, refining agent performance, and ensuring seamless integration with additional legal databases and tools.

As Definely continues to innovate, the company remains committed to building AI-powered legal infrastructure that enhances productivity while maintaining high standards of accuracy.