Legal research and legal AI

The lawyer's research desk, now with AI

News reports about court cases often say that a judge relied on an earlier Supreme Court judgment. That is how law works in India, as in Britain and America: the past decisions of courts guide the decisions that follow. So before a lawyer argues a case or advises a client, the lawyer has to research what the law says.

What a lawyer looks for. Legal research draws on three kinds of material:

Statutes

the laws passed by Parliament or a state legislature.

Case law

the judgments of courts, which guide later cases.

Commentary

books and articles in which experts explain the law.

A lawyer also has to check that each judgment is still good law, meaning no later court has overruled it. A tool called a citator shows every later case that cited a judgment, and whether that case followed it or overruled it.

What a lawyer checks: statutes, case law and commentary, and then the citator, which shows every later case that cited a judgment and whether it followed it or overruled it.

Who sells it. A handful of publishers own the great legal databases. In America and Britain, two research services lead: Westlaw, from Thomson Reuters, and Lexis, from LexisNexis. Law firms, companies' legal departments, courts and law schools pay for subscriptions to them. The judgments themselves are public. What the publishers sell is the work around them: summaries, links between cases, and the citator.

What changed. Generative AI has changed this desk faster than any other part of law. An AI assistant can answer a legal question in plain English, draft a research memo or summarise a hundred judgments. The legal publishers and a new wave of AI companies are now racing to build the assistant every lawyer uses. The publishers bring databases they have built over decades. The AI companies bring products built around AI from the start.

The database matters because AI models can invent cases that don't exist. Lawyers in America have been fined for citing such made-up cases in court. An assistant is only as reliable as the legal database behind it.

The race to build the lawyer's AI assistant: the publishers, Wolters Kluwer, LexisNexis and Thomson Reuters, bring databases built over decades, and the AI companies, Harvey, Legora and Eudia, bring products built around AI. An assistant is only as reliable as the legal database behind it.

India's part. The legal publishers run large engineering centres in India, where they build much of their research platforms.

Every company named here has posted software engineering jobs in India. Famous companies that don't actively hire software engineers in India are left out.

Keep reading, free

One more section is on this page: Research databases and AI assistants. Sign in to read it here, in full.

Continue with Google

Research databases and AI assistants

Research databases and AI assistants

The publishers

Wolters Kluwer

the Dutch information company, publishes legal, tax and medical references, and the software built on them. Its Indian centres, in Pune and Chennai, build much of that software. It has by far the most job postings in this Industry Vertical. (More on Wolters Kluwer in Market Segment 22.2, AI medical scribes and clinical documentation, in Industry Vertical 22, Digital health, and in Market Segment 11.3, Financial close, FP&A and tax software, in Industry Vertical 11, Finance and accounting software.)

LexisNexis

part of RELX, with engineers in Delhi and Chennai, runs Lexis+ and Shepard's Citations. Lexis+ is one of the two great legal research services, and Shepard's is its citator.

Thomson Reuters

with engineers in Bengaluru and Hyderabad, runs the other great research service, Westlaw. It also runs CoCounsel, its AI assistant for lawyers. Thomson Reuters is covered mainly in another Market Segment. (More on it in Market Segment 33.1, News and professional information providers, in Industry Vertical 33, News, publishing and social media.)

RELX

with engineers in Mumbai, Chennai and Delhi, is the parent group of LexisNexis. It is covered mainly in another Market Segment. (More on it in Market Segment 33.3, Academic and scientific publishing, in Industry Vertical 33, News, publishing and social media.)

The AI companies

Harvey

with engineers in Bengaluru, builds the generative-AI platform used by many of the world's largest law firms.

Legora

a Swedish company, and Eudia compete with Harvey. Both have engineers in Bengaluru.

One company here works for the public instead

FindLaw

with engineers in Bengaluru, runs a legal information website and a directory of lawyers. People use it to understand a legal problem and to find a lawyer.

The junior lawyer's work done by AI.

In a law firm, junior lawyers are called associates. Their early work is research memos, first drafts of the documents filed in court, called briefs, and reviewing hundreds of documents. That is the work the legal AI companies automate first. Harvey grew to serve many of the largest law firms within two years. The legal publishers are putting their archives behind their own assistants. The open question for law firms is how to train new lawyers when AI does the work they used to learn on.

That is the lawyer's research desk: a few publishers that own the legal databases, a new wave of AI companies, and a race to build the assistant every lawyer uses. Much of it is built in India.

Who they hire

Who these companies hire, and for what, is on What legaltech companies hire for.

← Previous in Legaltech

Contract lifecycle management (CLM) software

Next in Legaltech →

Legal practice and matter management software