For Lawyers · Published Free
How this firm actually runs — the systems, the workflows, the software, and the mistakes. Written for solos and small firms whose bottleneck is hours, not talent.
Start Here
4 articles·56 min· read in order
Essay
For ten years, firm management was an integration project — assembling a stack of SaaS tools that mostly talked to each other. That era is ending. Five concepts from before I was a lawyer that decide whether you can build your own.
14 min
Essay
Strip away the interface and a practice management system is a relational database, a set of views, and an API. That used to be worth every dollar. Now the database is a weekend, the connectors reach anything, and the honest question is which parts you should still rent.
16 min
Essay
Everything in this section — the wiki, the agent briefs, the checklists, the eval set — lives in a single directory of plain text. That one architectural fact is what lets you change models on a Tuesday afternoon, including to one running on the machine under your desk.
12 min
Essay
AI can draft, research, and summarize. It cannot tell you what you were doing. A working attorney on the mental load of parallel work — and the kanban board built to carry it.
14 min
Latest
Tool
A small firm does not need a database server. Choosing one is a decision to take on operational work that produces nothing for a client. One file on your own machine — with WAL, foreign keys, FTS5, and a real migration path — is more capable than most lawyers realize.
Read the articleOne email when something new is published here. No course, no upsell — the Lab stays free either way.
§01 · Corpus Map
Each article is reduced to a TF-IDF vector over its own vocabulary; every pair is scored by cosine similarity, and each article keeps its two strongest links. The layout is a Fruchterman–Reingold force simulation run to convergence at build time — deterministic, so the same corpus always yields the same map. Clusters are emergent, not assigned.
§02 · Similarity Matrix
The graph is a projection; this is the object it projects. Cell (i, j) is the cosine similarity between articles i and j. Rows and columns are ordered by agglomerative hierarchical clustering with average linkage over 1 − cosine distance, so the dendrogram's leaf order puts related articles adjacent — the dark blocks along the diagonal are the clusters, not an artefact of ordering.
§03 · The Archive
Workflow
Generating a .docx or a filled PDF is easy. Generating one that opens correctly on the reviewer's machine, prints right, and does not embarrass you in front of a clerk is a pipeline with five specific failure modes — all of them invisible to the code that produced the file.
Essay
Everything in this section — the wiki, the agent briefs, the checklists, the eval set — lives in a single directory of plain text. That one architectural fact is what lets you change models on a Tuesday afternoon, including to one running on the machine under your desk.
Workflow
A pre-publication link check written as a shell loop reported 'Safe to publish' without ever having checked anything. The bug, why zsh made it invisible, the fix, and the discipline that catches the whole category in thirty seconds.
Tool
The practical stack for running Qwen on firm hardware — models, runtimes, quantization, what fits in how much memory — and the client-facing tier election it makes possible. If the client picks the tier, the routing has to enforce it.
Essay
A prediction, with the mechanism and the failure conditions: open weights drive the economic value of the model itself to nearly zero, and the durable value migrates to whoever owns compute — and beneath that, power. What a law firm should do differently if that is right.
Essay
Strip away the interface and a practice management system is a relational database, a set of views, and an API. That used to be worth every dollar. Now the database is a weekend, the connectors reach anything, and the honest question is which parts you should still rent.
Essay
For ten years, firm management was an integration project — assembling a stack of SaaS tools that mostly talked to each other. That era is ending. Five concepts from before I was a lawyer that decide whether you can build your own.
Workflow
The most common citation defect in AI-assisted legal writing is not the fabricated case. It is the real quotation that stops one clause too soon — and it is nearly invisible on review. Here is a detector that catches it mechanically.
Essay
Most legal futurism is unfalsifiable, which is why it is always half-right and never useful. Twelve predictions about how AI reshapes legal practice — each with the mechanism driving it and the specific evidence that would prove it wrong.
Tool
If your AI tool reads statutes through a summarizing fetch layer, you do not have a legal research system. You have a paraphrase generator with citations attached. Here is the scraper I replaced it with, and the four bugs it found.
Tool
Nobody in law builds evaluation sets, so nobody can tell a workflow that got better from one that quietly got worse. Ground truth, the metrics that actually matter per task type, and why cost-weighted error is the only honest score in a law office.
Essay
Legal aid can't hire against one-year grants, and volunteer hours consume the supervision they're supposed to add. A forward, specialty-tagged purchase commitment fixes both — and bundles the one input big firms actually have to give.
Tool
Why dumping a matter file into a context window fails, why embedding search underperforms specifically on legal text, and how to build retrieval that recovers the right paragraph with a citation attached.
Essay
Every system that automatically ties work to a matter faces the same temptation: always produce an answer. In a law office that is the wrong design, and the reason is Rule 1.5.
Workflow
Twelve calls, four web inquiries, a portal message and three e-service notices is not a lot of information. It is a lot of interruptions. On classifying inbound automatically, routing it by consequence, and the discipline of suppressing what does not matter.
Tool
The strongest confidentiality posture is client data that never leaves the building. Here is what local models actually cost, what they are genuinely good at, where they fall down, and the hybrid architecture that is usually the right answer.
Workflow
Unbilled time is not a discipline problem, it is a capture problem — the work happened somewhere your billing system could not see. On reconstructing it honestly, why rounding decides whether approved equals billed, and the one rule that must never bend.
Essay
The dangerous bug in a firm's own software is never the crash. It is the button that does nothing, the file saved to a default location, the integration that stopped syncing in March and told no one.
Workflow
The mail arrives as a stack of scanned PDFs and leaves as filed documents, calendared deadlines, and a short list of things somebody has to decide. Six steps, and a clear rule about which ones a human must approve.
Essay
Every AI-assisted workflow generates artifacts — prompts, retrieved documents, intermediate drafts, transcripts. Are they work product? Are they privileged? Does a litigation hold reach them? Nobody knows yet, and most firms are retaining them without having decided to.
Tool
There is an open standard for exactly this problem. My practice management vendor does not implement it. So the connector is four hundred lines of my own Python, and the interesting part is not the code — it is why every firm is writing it separately.
Workflow
The software you build for your own firm will be used by lawyers and legal assistants, not engineers. The single fastest way to make it unusable is to inherit the vocabulary of the tools you built it with.
Workflow
The reason AI gives you generic legal work is that it has no idea where it is. A structured Obsidian knowledge base — plain markdown, real schema, provenance on every fact — turns a general assistant into one that already knows your practice.
Essay
Do you have to tell a client you used AI on their matter? What happens to the fee when a task takes twenty minutes instead of two hours? The disclosure question is genuinely unsettled — the fee question is not, and it cuts against the lawyer.
Workflow
Minnesota's three-day mail rule moved in 2020. Rule 6.05 was abrogated; the text now lives in Rule 6.01(e). If you learned it the old way you have been carrying a wrong rule for years — and so has every AI you ask.
Workflow
A fabricated case is the easy failure — it does not exist, so somebody notices. The dangerous one is a real case, correctly cited, standing for something it never said. Here is the four-pass verification gate that catches it.
Workflow
One general-purpose AI assistant gives you a capable generalist with no limits and no handoffs. A real office is not organized that way, and neither should this be. On writing agent briefs, and why the limits matter more than the capabilities.
Workflow
Every firm's institutional knowledge lives in its form file, and every firm's form file is quietly wrong in places nobody has checked since 2014. On template hygiene, stale-rule propagation, and why generation should populate a template rather than rebuild one.
Essay
Nine of every ten ethics complaints in Minnesota close with no discipline at all — and every one of them was answered by a lawyer. A look at what our regulatory apparatus costs the people who never get a lawyer in the first place.
Essay
AI can draft, research, and summarize. It cannot tell you what you were doing. A working attorney on the mental load of parallel work — and the kanban board built to carry it.
Nothing matches that.
One email when something new is published here. No course, no upsell, no newsletter about the newsletter — the Practice Lab stays free either way.
Prefer a reader? Subscribe by RSS →
The Premise
Most people with a real legal problem never see a lawyer. Not because they have no claim, and not because no lawyer would take it — but because an hour of legal help costs more than they have. Legal aid is chronically underfunded; pro bono is essential and will never scale to the size of the gap.
If we are serious about opening the courthouse door to the people currently standing outside it, there is one lever with enough leverage to matter: the cost of producing legal work has to come down. A lawyer who can carry more files well is a lawyer who can say yes to files the economics would otherwise rule out.
So everything here is written for other lawyers — solos, small firms, anyone whose bottleneck is hours rather than talent — and everything here is free.
There is no signup, no course, no subscription, and nothing for sale. We are not asking for donations.
If something here saves you time, spend some of it on people who could not otherwise afford you.
Take the case at a rate that works for someone on a fixed income. Keep a sliding scale. Say yes to the file you would otherwise have to turn down on economics. That is the entire price.
Nothing in this section discusses a client, a matter, or a real case. Screenshots use invented clients and file numbers. Everything here is general commentary on practice management — it is not legal advice, it is not ethics advice, and reading it does not create an attorney–client relationship. Every lawyer is responsible for their own compliance in their own jurisdiction.