Jack Stoneman
Head of AI Legal Practice
To make the law scale exponentially.
About
I work at the intersection of legal, software and ai as a lawyer / product manager / software engineer (Head of AI Legal Practice, M&A and Capital Markets @ MinterEllison) and as a teacher (Honorary Senior Fellow @ Melbourne Law School). Society faces a structural issue, where the legal system scales linearly but society scales exponentially - meaning that over time the legal system reaches fewer and fewer people. Applying software and ai to the legal system we can turn it into an exponential system and ensure that it serves everyone.
1 Project
Template Builder
template-builder turns a firm's precedent contracts into a deterministic document generator a lawyer can actually sign off on. Feed it 20–30 agreed examples of a document type and an LLM decompiles them into a clause graph — variants, conditions, defined terms, semantic dependencies — but everything downstream of that boundary is pure code: rendering, validation sweeps, content-hash approval and constrained edits are deterministic, so "approved" means exactly what was reviewed. A matters layer runs live negotiations: counterparty redlines are diffed clause-by-clause against the known render, routed through an escalation inbox, and every human decision is journaled with its rationale. Those journals distil into a citation-checked playbook, and an autonomy gate lets the assistant act alone only on clauses where it has demonstrably reproduced the lawyers' past decisions — automation earned per clause, never switched on. Python 3.14, Pydantic, FastAPI, a single-file web UI, and 268 offline tests.