UniCourt’s New Product Combines LLMs with APIs to Let You Go DEEP Into Customizing Court Docket Data and Delivering It Where You Need It

UniCourt, an organization that gives entry to litigation information and analytics, right this moment launched into beta a first-of-its-kind product that mixes LLMs and APIs to permit its prospects – which embody legislation corporations, insurers, authorized tech corporations, and others – to extract any information factors they want from UniCourt’s assortment of over 1 billion dockets and paperwork and ship the information to wherever they need, similar to to an information lake or to software program similar to Basis.

Known as UniCourt DEEP – for Docket Extraction and Enrichment Platform – the product permits customers to uniquely leverage what the corporate says is the biggest normalized and structured database of litigation information in america, protecting over 40 states and greater than 3,000 state and federal courts.

By combining this complete docket database with superior generative AI options, UniCourt DEEP permits customers to create custom-made docket and doc views, permitting corporations to extract particular information factors and insights tailor-made to their distinctive wants and use circumstances.

Docket information extracted by way of DEEP may be instantly built-in into common purposes similar to Litera Basis or pushed to information warehouses similar to Snowflake, Salesforce DataCloud, Azure Synapse, and Microsoft Cloth.

This permits corporations to include litigation information instantly into their present workflows and purposes, UniCourt says, pushing the particular docket information they should the situation the place they want it.

UniCourt says that one in every of DEEP’s defining options is unparalleled real-time entry to courtroom information, with 100% uptime. “This reliability is essential to be used circumstances the place up-to-date info is important and units a brand new customary within the authorized tech trade,” the corporate says.

DEEP offers this customization by enabling using LLM immediate engineering strategies on-platform so customers can discover and extract precisely what they’re searching for from dockets or paperwork.

“Our purpose is to permit the subject material specialists to leverage their information in our AI instruments to simply discover and construction precisely what they’re searching for,” mentioned UniCourt founder and CEO Josh Blandi. “It’s now tremendous easy for corporations, insurance coverage carriers, and authorized tech corporations to construct purposes and their very own customized AI fashions on this information.”

A Want for Customized Information

Forward of right this moment’s launch, I used to be briefed on the product and given an indication by Blandi, Akshay Kumar, vp of product administration, and Rob Lynch, chief working officer and former head of product.

In a nutshell, UniCourt DEEP takes unstructured information from courts and different sources and converts it right into a extremely structured litigation information graph constructed round dockets, paperwork, attorneys, legislation corporations, events, judges, and courts.

The product developed, Lynch mentioned, from the conclusion that some legislation corporations wished extra particular information factors than the usual information being supplied by way of UniCourt’s APIs, which incorporates classes similar to case, legislation agency, lawyer, case sort, decide, get together and courtroom.

“So what DEEP does is it places the ability of their palms without having an enormous engineering group to get the entire energy out of our information set, without having to place a group of engineers on it, or for us having to construct a product,” mentioned Lynch. “It simply places the ability of their palms to say, discover what you need, extract it, and push it the place you want it.”

“For instance,” Blandi added, “you might establish, ‘I would like this lawyer throughout all of the jurisdictions he’s ever filed circumstances in. I need to load a set of paperwork with a sure decide.’ Proper. And establish sure paperwork and establish issues inside these paperwork.”

Earlier than the event of DEEP, prospects would typically come to UniCourt with extremely particular information wants that UniCourt would then construct for them. For instance, UniCourt’s information would have already got tagged all private harm circumstances, however a person would possibly need to know if, throughout these circumstances, accidents from burns in lodge kitchens are trending upwards.

“Now when of us ask that query,” Lynch defined, “we are able to say, ‘Look, you could have a platform right here that you need to use and you may extract that info proper down into the doc stage, and you may reply these questions utilizing it, construction it, after which push it the place you need it.’”

What they realized, Blandi mentioned, is that having the shopper’s material knowledgeable on the platform, utilizing UniCourt’s AI instruments, and in a position to develop and iterate and validate prompts, is rather more environment friendly and exact than having it achieved by UniCourt’s personal engineers, who usually are not SMEs.

A Video Instance

This video exhibits how DEEP permits customers to customise their views of courtroom information. On this instance, the person is searching for all circumstances involving chest accidents the place greater than $500,000 was awarded.

They begin utilizing the structured fields of Case Sort to pick Private Damage circumstances and Court docket to seek for New York circumstances after which additional refine the search to seek out simply complaints and judgments.

That produces an inventory of circumstances, from which the person can then use AI to floor particulars from inside these paperwork and dockets.

The person prompts, “Discover circumstances with chest accidents. Present judgment awarded.” DEEP then exhibits the accidents and award quantities. These outcomes can then be piped into different purposes, both on a one-time or ongoing foundation.

Coming Out of Alpha

The product up to now has been in a personal alpha improvement part that permits the information to be custom-made and considered inside Snowflake. The product is now shifting right into a public beta stage, and new pipelines will enable the information to attach into extra platforms similar to Basis and Salesforce.

At this level in its improvement, the product is restricted to verdict information from private harm circumstances, which UniCourt extracts and layers on high of its structured information graph.

Throughout this subsequent part, UniCourt hopes to get extra giant legislation corporations to signal on to check the product. The corporate hopes to launch the manufacturing model within the fall.

Kumar mentioned the ability of the product can be a mixture of two issues. One is the information UniCourt already has. “What we have now, actually on the earth proper now, I’d very frankly say [is] the most effective structured information graph round a litigation information set.”

However the second piece is the AI ingredient, “which may dig deeper into phrases and items of information than a straight-up tagging might ever do.”

For these attending ILTACON this week, UniCourt will likely be demonstrating DEEP there of their sales space within the exhibition corridor.