BRENNEN

Industries / Finance and business services

Legal and professional services

Legal and professional services sell expert judgment for a fee: lawyers research, draft and argue the law, accountants keep and audit the books, tax practitioners work out and file what is owed, and consultants advise managers.

1292

Edward I of England ordered the judges of the Common Bench to choose "the best lawyers from the counties" to serve the king's court, "and no others", which put both kinds of lawyer, attorneys and pleaders, under the judges' control.

middletemple.org.uk

37

Milestones achieved across 7 domains.

8

To watch for: announced, not yet confirmed outside.

12

Open deliverables. Nobody has shipped them.

In one minute

Where it started

By the 1280s England's royal courts had paid lawyers, and in 1292 Edward I put them under the control of the judges. The method behind accounting reached print in Venice in 1494, in a treatise on double-entry bookkeeping in Luca Pacioli's Summa, and Edinburgh's accountants received a royal charter in 1854. Consulting grew partly out of a Boston testing laboratory that Arthur D. Little co-founded in 1886.

Where it stands

The work is still mostly reading and writing: case law, contracts, ledgers, tax returns and reports. Legal search moved online with LEXIS in 1973, and in 2012 a US judge approved software-assisted document review in what he called the first opinion to do so. AI tools now draft and check much of this work, and outside tests give mixed results: in a 2025 test, AI tools beat a lawyer baseline on research questions, while a public tracker lists 2,306 court decisions, as of 16 November 2026, about AI-made content such as invented citations.

What is open

The common gap is checking. Most of the open jobs below are free tests and answer keys, such as page numbers to check citations against, a blind test for AI document review and a check of every reference in published consulting reports, or public records turned into data, such as city accounts. Each can be started this month with public data, open-source tools and current AI models, and none of them gives advice to a real client.

Domains

AchievedWatch forOpen

Other narrower industries

Court reportingLegalTechE-discovery

Figureheads

2 people

Advances to build with

1 research note

Open in legal and professional services

3 of 12 open deliverables

Open · Legal research

A free pin-cite checker

A free tool that reads a brief, finds each pin cite, and checks the quote or point against that page of the printed law report. A pin cite gives the exact page: "1 F.3d 1, 5" means volume 1 of the Federal Reporter, Third Series, a case that starts on page 1, and a point made on page 5.

Built with what exists

The Caselaw Access Project's free case files, whose web versions mark where each printed page begins (for example "2" and "3" in 1 F.3d 1, checked in this session); eyecite and the CourtListener citation lookup API, both from Free Law Project; a language model such as Claude or GPT-5 to compare each quote or point with the page text.

Credit · unclaimed

Open · Contracts

Two agents, one NDA

A public test in which two AI agents negotiate a mutual non-disclosure agreement (NDA, a contract to keep each other's information secret) for opposite sides. Each agent gets a written playbook of what its side may accept. Lawyers grade each final draft blind, beside drafts negotiated by pairs of people with the same playbooks. No real client is involved.

Built with what exists

Common Paper's Mutual NDA, a free standard form written by a committee of lawyers and licensed CC BY 4.0, whose one-page cover sheet holds the terms that get negotiated, such as purpose, term and governing law; two agents built with a framework such as the Claude Agent SDK; a shared document that records each change. (see the Note "Many agents, one budget")

Credit · unclaimed

Open · Litigation and discovery

Blind test for AI review

A shared test in which any document-review tool, generative or not, is run on the same public collection and the same document requests, and scored by assessors who do not know which tool marked which document.

Built with what exists

the TREC Legal Track collections (scanned tobacco-industry records, and the EDRM email collection used in 2010 and 2011) with their answer keys; new document requests written for this test; review tools built on Claude, Gemini, GPT and open-weight models; random sampling to estimate recall. (see the Note "A thousand questions, one pass")

Credit · unclaimed