Legal Case Law Model Developed for Generative AI Startup

Challenge

An OpenAI-backed startup is building a generative AI model focused on the legal profession. They are building a platform to support contract review, due diligence, and regulatory compliance. To effectively train a large language model (LLM), however, the client needs a partner with an unprecedented mix of attributes– scale, legal experience, affordability, and diversity.

Approach

To build a comprehensive and accurate LLM, the startup needs the best data on the market. They begin Phase 1 with a 12-person team to review initial questions for the model. In Phase 2, Axiom ramps up to 100 lawyers—diverse in skillset, location, experience, etc.—to engage directly with and evaluate the model. Simultaneously, the startup checks and compares Axiom’s answers with legal case law to improve the model in real time. In the final phase, a new contingent of Axiom lawyers (~750 total) deep dives on the trained model to optimize for niche subsets of the law and specific regulations.

Impact 

• Enables client’s legal team to focus on core business matters by leveraging Axiom’s unique blend of size, experience, and diverse lawyer base to train the LLM
• Highly qualified Axiom lawyers provide detailed data on specific regulations to ensure the AI model accurately handles complex contract review and regulatory compliance requirements.
• Nearly 1000 total Axiom lawyers evaluating the LLM’s answers in real-time rapidly accelerates AI response reliability and accuracy

*Cost savings estimates are based on actual Axiom rates compared to equivalent law firm rate averages as listed in Wolters Kluwer’s Real Rate report.