Highlights:
- With the extra funds, the company intended to buy vLEX, a business that offers a searchable database of laws, court decisions, and other legal documents.
- Harvey’s software uses a customized large language model from OpenAI as part of its internal engine.
Harvey is reportedly raising USD 100 million in recent funding. It develops AI tools for lawyers.
According to the sources, Alphabet Inc.’s GV fund is anticipated to lead the financing. Harvey’s estimated value due to the investment is USD 1.5 billion, roughly twice what it was worth after a USD 80 million boost in December last year. The OpenAI Startup Fund was a part of that latter agreement.
Harvey’s initial goal was to raise USD 600 million, not the USD 100 million that it is currently trying to achieve. With the extra funds, the company intends to buy vLEX, a business that offers a searchable database of laws, court decisions, and other legal documents. The idea was to improve Harvey’s AI models using the vLEX dataset.
According to reports, the company reduced its fundraising goal to USD 100 million following the acquisition’s collapse. On the other hand, Harvey’s requested USD 1.5 billion valuation is reportedly USD 500 million less than what it had initially hoped for.
Harvey, officially known as Counsel AI Corp., was established in 2022 by lawyer Winston Weinberg and former DeepMind researcher Gabriel Pereyra. It provides a suite of AI technologies to enhance productivity of legal teams.
From trial transcripts, one of the products can spontaneously retrieve crucial data. By producing preliminary drafts, a different technology called Assistant guides legal teams in responding to clients’ legal inquiries more swiftly. Harvey claims that the responses made by the machine can include data from several dozen court documents.
The company’s third product helps litigation teams save time by automatically locating court-supporting legal papers. The company claims that the AI also consults other documents and regulatory filings in addition to historical cases. The program makes it easy for lawyers to locate the source papers by including citations in each response.
Harvey’s software uses a customized large language model from OpenAI as part of its internal engine. The firms claim that compared to GPT-4, AI can produce lengthier, more thorough responses to legal queries and experience fewer hallucinations.
How the business intends to use the funds is yet to be revealed. Harvey might increase the scope of its collaboration with OpenAI.
The creator of ChatGPT introduced a program last year that enables businesses to order unique GPT-4 versions. In response, it disclosed in May last year that its engineers had started preparing the GPT-4’s replacement. Harvey may enhance the proprietary OpenAI model that powers its software with a more sophisticated version based on the ChatGPT developer’s next-generation LLM using the funding round that is presently being developed.