Ontra is now running OpenAI’s GPT-4 model within its Legal Operating System, a software program that automates legal tasks for sponsors.
The external model is integrated into Ontra Synapse, which is the contract automation provider’s AI-powered engine that is part of the operating system and supports its components. Ontra Synapse is using GPT-4 in addition to Ontra’s proprietary AI models, said Nat Kunes, the firm’s chief product officer.
The front-end experience remains the same, but Ontra Synapse will use a different model depending on the relevant features that are part of the service provider’s software.
GPT-4 and the proprietary AI models are called large language models, or LLMs, Kunes explained. These are specific kinds of models that incorporate data to churn out insights for their users.
And in OpenAI’s case, GPT-4 powers its widely publicized ChatGPT program, which is a front-end search tool.
Ontra is willing to include outside LLMs in its software if they work the best for customers, Kunes noted.
And the service provider has been on the lookout for other models, comparing outside ones to its own.
“All along the way we’ve been testing all the different large language models out there for specific use cases in our products and saying, ‘okay, is our proprietary model better for this particular feature?,’” he said.
Ontra pointed out that GPT-4 is also included in Insight, a component of the operating system that helps GPs with fundraising and side letter compliance. The model’s addition comes after the firm updated Insight with a digitized most-favored nation feature.
LLM updates aren’t the only AI changes that sponsors should expect.
Kunes said that Ontra plans to add a search capability for users to run queries across broader sets of data. There isn’t a firm release timetable for this upgrade, but he doesn’t expect it to be far out into the future.
Users can currently get insights that are derived from AI scouring specific documents that they upload, Kunes said.