Evisort, calling itself the leader in artificial intelligence (AI) for contract lifecycle management (CLM) and analysis, announced its successful development of the industry’s first large language model (LLM) specifically for legal agreements. Evisort’s bespoke LLM will be a key component of the company’s new, multi-faceted generative AI engine, combining third-party and proprietary technology to bring unparalleled innovations to contract management.
The rise of general LLMs such as OpenAI’s ChatGPT, Anthropic’s Claude, and Meta’s LLaMA 2 has sparked the imagination of users and ushered in a new wave of AI investment. However, the complexity and unique terminology associated with contracts necessitate a domain-specific model to optimize the benefits of generative AI for legal and business stakeholders.
Created by Evisort AI Labs, which has developed proprietary AI technology for legal use cases since 2016, Evisort’s LLM is a Transformer-based language model trained and fine-tuned on a dataset of over two trillion tokens, combining generic and domain-specific knowledge culled from millions of legal contracts. It will enable Evisort to continue delivering AI-powered innovations with greater control over accuracy and responsiveness than is possible when relying only on generic, third-party LLMs.
“Evisort strives to help organizations know everything about their contacts and contracting processes – without limitations or dependencies,” said Jerry Ting, Evisort Founder and CEO. “Generative AI is helping us drive the next phase of our mission.”
Evisort’s LLM is the latest milestone in the company’s generative AI development. Earlier this year, Evisort launched Automated Redlining and Clause Creator to help users draft, edit, and negotiate contracts faster and with less risk. Evisort also recently introduced the ability for users to ask questions about their contracts in simple, natural language — such as “What happens if we terminate early?” — and receive clear, reasoned answers in a conversational experience.
In addition to helping power these features, Evisort will also leverage its new LLM to enhance its existing capabilities for users to easily create custom AI models. Released in 2021, Evisort’s Automation Hub™ has already been used by premier corporations like Microsoft, BNY Mellon, and NetApp to track tens of millions of unique contract clauses of interest via custom AI models they trained themselves with simple, drag-and-drop controls.
Soon, Evisort users will be able to track virtually any desired data point from their contracts (including handwritten text and data from tables) and from related documents like order forms, SOWs, and invoices. Such data points empower legal and business teams to proactively monitor their contractual rights and obligations across the organization — via search, dashboards, and reports in Evisort or via any enterprise system the organization connects to Evisort, such as ServiceNow, Salesforce, SAP Ariba, or hundreds of others.
“Having a contract-specific LLM in our ‘best-tool-for-the-task’ AI engine lets Evisort build tailored innovations on a strong foundation,” said Evisort Founder and CTO Amine Anoun. “Not only can we deliver highly accurate models at the outset, we have the ability to continually refine their accuracy over time, something most CLM providers cannot do.”