Harvey

Harvey is a generative AI platform designed to enhance the workflows of lawyers and law firms across various practice areas. It offers a suite of products that leverage large language models trained on complex legal tasks and fine-tuned by domain experts to provide specialized support for the legal field.

Key features:
- Augments legal workflows with domain-specific models to streamline routine tasks
- Answers complex research questions based on trusted legal sources
- Assists with drafting, analyzing, comparing, and querying legal documents using natural language
- Provides governance tools for tracking productivity and controlling usage and data provenance
- Ensures high security standards through deployment on Microsoft Azure

How it works:
Users interact with Harvey primarily through a web-based platform. They can input natural language queries or documents, and the system utilizes its generative AI capabilities to provide relevant insights or document analysis. This enables lawyers to efficiently draft or review legal documents, conduct in-depth research, and analyze the relevance and implications of specific regulations or filings.

Integrations:
Harvey is deployed on Microsoft Azure, suggesting seamless integration with various Azure services and potentially other Microsoft products commonly used in professional environments, such as Microsoft 365.

Use of AI:
The platform leverages generative artificial intelligence built on large language models (LLMs) that have been specifically trained and fine-tuned for legal applications by domain experts. This ensures that Harvey's capabilities are both broad in scope and highly specialized for legal work.

AI foundation model:
Harvey utilizes large language models trained on complex legal tasks as its foundation.

How to access:
Harvey is primarily accessed through a web-based platform, making it easily accessible to lawyers, law firms, and legal departments within larger organizations looking to enhance their productivity and leverage AI in their legal workflows.