Harvey: GPT-6 Astra for More Comprehensive Legal Documents

Harvey claims that the integration of GPT-6 Astra enhances the quality of legal projects produced by its platform. The company highlights improvements in formatting, better contextual understanding, and a memory panel to incorporate lawyers' preferences directly into the workflow.

Harvey claims more comprehensive and better-formatted documents

Harvey reports having observed substantial improvements in document formatting and contextual consideration with GPT-6 Astra compared to other models. According to the company, these advancements result in more comprehensive projects that are truer to the source material. Harvey also asserts that GPT-6 Astra's ability to process more context allows for the production of higher-quality legal documents, which, it claims, enables clients to spend more time on strategy.

GPT-6 Astra applied to the analysis and synthesis of legal sources

Harvey indicates that it uses GPT-6 Astra to assist lawyers in analyzing, synthesizing, and drafting from the elements that make up a case. Among the sources considered are judicial data, files from law firms, research on case law, and other legal context databases. The company states that GPT-6 Astra provides more context during drafting and generates more structured outputs.

Writing preferences integrated into the production workflow

The memory panel offered by Harvey, which fits into the drafting process, allows for the configuration of personal preferences such as the use of numbered lists, highlighting EDGAR as the primary source, or using color codes to categorize points by priority. These settings are displayed alongside reference documents and the memorandum draft to provide more precise guidance on the final output. Harvey specifies that this approach is part of the secure deployment of AI in complex legal workflows, from litigation to mergers, and that its clients use it to transform large volumes of information into complex legal documents.