Be part of our each day and weekly newsletters for the most recent updates and unique content material on industry-leading AI protection. Be taught Extra
When authorized analysis firm LexisNexis created its AI assistant Protégé, it wished to determine one of the simplest ways to leverage its experience with out deploying a big mannequin.
Protégé goals to assist attorneys, associates and paralegals write and proof authorized paperwork and be sure that something they cite in complaints and briefs is correct. Nonetheless, LexisNexis didn’t need a normal authorized AI assistant; they wished to construct one which learns a agency’s workflow and is extra customizable.
LexisNexis noticed the chance to carry the facility of enormous language fashions (LLMs) from Anthropic and Mistral and discover the most effective fashions that reply consumer questions the most effective, Jeff Riehl, CTO of LexisNexis Authorized and Skilled, informed VentureBeat.
“We use the most effective mannequin for the particular use case as a part of our multi-model strategy. We use the mannequin that gives the most effective end result with the quickest response time,” Riehl stated. “For some use circumstances, that can be a small language mannequin like Mistral or we carry out distillation to enhance efficiency and cut back value.”
Whereas LLMs nonetheless present worth in constructing AI purposes, some organizations flip to utilizing small language fashions (SLMs) or distilling LLMs to develop into small variations of the identical mannequin.
Distillation, the place an LLM “teaches” a smaller mannequin, has develop into a preferred methodology for a lot of organizations.
Small fashions typically work finest for apps like chatbots or easy code completion, which is what LexisNexis wished to make use of for Protégé.
This isn’t the primary time LexisNexis constructed AI purposes, even earlier than launching its authorized analysis hub LexisNexis + AI in July 2024.
“We’ve used a number of AI up to now, which was extra round pure language processing, some deep studying and machine studying,” Riehl stated. “That basically modified in November 2022 when ChatGPT was launched, as a result of previous to that, a number of the AI capabilities had been form of behind the scenes. However as soon as ChatGPT got here out, the generative capabilities, the conversational capabilities of it was very, very intriguing to us.”
Small, fine-tuned fashions and mannequin routing
Riehl stated LexisNexis makes use of completely different fashions from a lot of the main mannequin suppliers when constructing its AI platforms. LexisNexis + AI used Claude fashions from Anthropic, OpenAI’s GPT fashions and a mannequin from Mistral.
This multimodal strategy helped break down every activity customers wished to carry out on the platform. To do that, LexisNexis needed to architect its platform to swap between fashions.
“We might break down no matter activity was being carried out into particular person parts, after which we might determine the most effective giant language mannequin to help that part. One instance of that’s we are going to use Mistral to evaluate the question that the consumer entered in,” Riehl stated.
For Protégé, the corporate wished quicker response occasions and fashions extra fine-tuned for authorized use circumstances. So it turned to what Riehl calls “fine-tuned” variations of fashions, primarily smaller weight variations of LLMs or distilled fashions.
“You don’t want GPT-4o to do the evaluation of a question, so we use it for extra subtle work, and we swap fashions out,” he stated.
When a consumer asks Protégé a query a couple of particular case, the primary mannequin it pings is a fine-tuned Mistral “for assessing the question, then figuring out what the aim and intent of that question is” earlier than switching to the mannequin finest suited to finish the duty. Riehl stated the subsequent mannequin may very well be an LLM that generates new queries for the search engine or one other mannequin that summarizes outcomes.
Proper now, LexisNexis principally depends on a fine-tuned Mistral mannequin although Riehl stated it used a fine-tuned model of Claude “when it first got here out; we aren’t utilizing it within the product at the moment however in different methods.” LexisNexis can also be interested by utilizing different OpenAI fashions particularly for the reason that firm got here out with new reinforcement fine-tuning capabilities final yr. LexisNexis is within the means of evaluating OpenAI’s reasoning fashions together with o3 for its platforms.
Riehl added that it might additionally take a look at utilizing Gemini fashions from Google.
LexisNexis backs all of its AI platforms with its personal data graph to carry out retrieval augmented technology (RAG) capabilities, particularly as Protégé may assist launch agentic processes later.
The AI authorized suite
Even earlier than the arrival of generative AI, LexisNexis examined the opportunity of placing chatbots to work within the authorized {industry}. In 2017, the firm examined an AI assistant that will compete with IBM’s Watson-powered Ross and Protégé sits within the firm’s LexisNexis + AI platform, which brings collectively the AI companies of LexisNexis.
Protégé helps regulation companies with duties that paralegals or associates are likely to do. It helps write authorized briefs and complaints which can be grounded in companies’ paperwork and knowledge, recommend authorized workflow subsequent steps, recommend new prompts to refine searches, draft questions for depositions and discovery, hyperlink quotes in filings for accuracy, generate timelines and, after all, summarize complicated authorized paperwork.
“We see Protégé because the preliminary step in personalization and agentic capabilities,” Riehl stated. “Take into consideration the several types of attorneys: M&A, litigators, actual property. It’s going to proceed to get increasingly personalised primarily based on the particular activity you do. Our imaginative and prescient is that each authorized skilled could have a private assistant to assist them do their job primarily based on what they do, not what different attorneys do.”
Protégé now competes towards different authorized analysis and know-how platforms. Thomson Reuters personalized OpenAI’s o1-mini-model for its CoCounsel authorized assistant. Harvey, which raised $300 million from buyers together with LexisNexis, additionally has a authorized AI assistant.
