datons for Law

A rule is never one text: it is a series of versions, amended at different dates, layered across national, regional and EU law, published in gazettes that do not talk to each other. Answering which version was in force on a given day means reconstructing it by hand. Meanwhile the firm’s own knowledge — contracts, opinions, filings, the argument that worked two years ago — sits in a document manager that only finds a file if you already know what it is called. In a large firm the answer exists but lives in another department. In a small one it lives in the head of the person who wrote it, and leaves with them.

Most of that searching no longer needs a person. Finding which clause was in force, checking a deadline across a matter, pulling the precedent that matches the case in front of you: a machine does it today, and the hours go back to the work clients actually pay for. What it does need is ground to stand on. An AI pointed at unverified sources produces a citation that reads perfectly and does not exist — in this profession that is not a wrong answer, it is a filing you have to explain to a judge. Verified sources first, automation on top of them.