The Journal · Written from the studio
Short essays on building the next legal practice.
Notes from active transformations, thesis fragments, and short arguments too specific for a manifesto. Published irregularly and never syndicated.
The long-form doctrine lives separately, under the founder's own byline: see the essays.
J/01
May 2026 · Method
The order is fixed.
Why 'name, then systematise, then automate' is not a preference. It is the only sequence that produces a working practice.
7 min readJ/02
April 2026 · Thesis
Against the innovation committee.
A short brief on why every AI-transformation program at a large firm looks the same - and produces the same nothing.
5 min readJ/03
March 2026 · Method
Evidence is the answer to hallucination.
The right response to AI making things up is not better AI. It is a firm-wide rule that unanchored claims do not ship.
6 min readJ/04
March 2026 · Thesis
The associate math.
Why the numbers underneath a leverage model make AI adoption a threat rather than an upgrade - stated plainly, without euphemism.
6 min readJ/05
February 2026 · Method
What a matter actually is.
Most firms run three competing definitions of their most basic unit without noticing. Every downstream system inherits the confusion.
5 min readJ/06
February 2026 · Method
Fixed fee is a systems problem.
Firms that treat alternative pricing as a marketing decision retreat within a year. The ones that treat it as a measurement problem do not.
5 min readJ/07
January 2026 · Sector
In-house is not a smaller law firm.
The economics run in the opposite direction, which means the playbook does too. Importing firm practices into a department reliably misfires.
5 min readJ/08
January 2026 · Sector
Procurement is a structural constraint, not an excuse.
The right response to a nine-month approval cycle is not to complain about it. It is to sequence the work that needs no approval at all.
5 min readJ/09
December 2025 · Studio
The tool is not the transformation.
A note on why we do not recommend products, and why the question 'which AI should we buy' is almost always premature.
4 min readJ/10
December 2025 · Studio
Why we say no.
The engagements we decline, and the specific reasons - including the one where the honest answer is that a full rebuild is not available to you.
5 min readJ/11
November 2025 · Method
Measuring a practice that never measured itself.
Improvement you cannot demonstrate did not happen, as far as the organisation is concerned. A short note on baselines.
5 min read
Subscribe · No newsletter
We don't run a mailing list. New pieces show up here first, occasionally on the founder's site.