When Precedent Has No Citation: Filevine’s Semantic Citator

The First Legal Productivity Tools

Shepard’s moved online in 1980, but it wasn’t until LexisNexis acquired Shepard’s in 1999 that the research tool’s online platform gained the extensive analytical capabilities today’s users have come to appreciate. Westlaw, with a slightly larger market share, traces it’s roots to 1872, uses a similar tried and true system of legal research that some have found to be more accurate, based on a “Keycite” system. Both Lexis and Westlaw have moved heavily into the AI world with impressive systems. But both still focus on following particular citations through legal systems and legal histories.

These two giants in legal tech now have a new competitor. Filevine’s LOIS, or “legal operating intelligence system,” can find semantic similarities even if a particular case is not cited by another court, focusing on the underlying reasoning rather than on a particular name.

As Ryan Anderson told me, “Westlaw and Lexis are blind to your best arguments.”

Supercharging The Search

LOIS, best known for a slick video starring Sarah Rafferty (famous for her role as Donna Paulsen in Suits) as a senior litigator in a promotional short film, integrates directly into case files, documents, calendars, and communications to execute tasks. Now, the system can contextually shepherdize along with its other productivity processes, adding onto, if not replacing, other search tools.

Kyle Usner, Associate Attorney, Mass Torts, Insurance Claim HQ, said “For medical records review, LOIS is saving some cases hours. My paralegals and I always click back to verify the source. That’s what gives us confidence– it’s not making things up.”

The consequential idea is not semantic legal search by itself. Legal platforms, like Casetext or Paxton, have used natural-language, concept-based, and semantic search for years. Nor is semantic search the opposite of Shepardizing. Semantic search is principally a way to find authority, while Shepard’s and KeyCite are principally ways to validate authority after it has been found.

Filevine’s more ambitious move is to combine the two functions. LOIS starts with the particular passage a lawyer proposes to cite, follows the conventional citation graph, and simultaneously searches for later opinions addressing the same legal proposition even when they use different language and never cite the original case. It then attempts to classify the treatment and produce a source-linked memorandum. Filevine describes this as dual-pathway retrieval: one pathway follows citations; the other follows meaning.

Traditional citatory trace what happened to a case. A semantic citator attempts to determine what happened to the legal proposition inside the case. To do both on one platform can be a meaningful advance, assuming Filevine can establish that it works accurately across jurisdictions, courts and areas of law.

Suits Said It Best

Competing directly with some of the largest and best known legal technology providers is not for the faint of art, but Anderson and Filevine have demonstrated that they play to win, adding another element in their walled garden strategy to try to get ecosystem lock in for clients and customers who want to live in Lois’s world. The citator function has long been one of the principal reasons firms maintain expensive Westlaw and Lexis subscriptions. A research platform can have an attractive interface and good search, but lawyers will hesitate to depend on it if they cannot reliably determine whether the resulting authorities remain good law.

That explains why Paxton and vLex have also invested in citators and why legal-technology observers have described the citator as the “holy grail” for challengers to the major research providers. A viable independent citator could weaken one of the most durable competitive protections in legal publishing.

Or, as one of the most famous lines from Suits reminds us, “I’m Donna. I know everything.”