Courts have held that putting client material into public AI tools can waive privilege. Not that it might feel risky — that it can waive it, because the material was handed to a third party under that company’s terms. Every lawyer in your firm already knows this, which is why the real question was never whether the tools are useful. It is on what terms the client file is handled. Here the terms are written for you: the firm’s matters live in an encrypted space of their own, in our private cloud on Amazon Web Services, scoped to the firm alone and handled by us only under a written scope — and your assistant reads and drafts from that space, citing the file every answer came from.
We will not pretend the firm’s matters stay within its own walls — they live with us, and this page says exactly how. What we offer is narrower and checkable: an engagement in writing, for your ethics counsel to read, rather than a consumer policy you are asked to trust.
What changed
For two years the profession has been caught between two facts: the tools are genuinely useful, and the confidential material is exactly what makes them useful. Public AI tools resolve that tension badly — the client file goes to the tool, under consumer terms written for nobody in particular. We resolve it with terms written for the firm: your matters go to exactly one place, an encrypted space of your own in our private cloud, loaded under a written scope, never pooled with anyone else’s, never used to train anyone’s model, and deleted — confirmed in writing — when the service ends. Whether any particular use of an outside tool affects privilege in your jurisdiction is a question for your ethics counsel, not for us. Our part is to make sure that when your counsel asks how the material is handled, every answer is already in writing.
What the firm hands it
“Find the indemnity clause, Reyes file.”
Section 9, page 14 — and it differs from the version you sent in June.
Every matter, one question away.
“Build the binder for Thursday.”
Pleadings, exhibits, and correspondence in order. Two exhibits are missing signatures.
The binder, assembled overnight.
“What’s due in the next ten days?”
Four filings. The Halvorsen response is first, on Tuesday the 11th.
Deadlines watched, not remembered.
“Summarize this 60-page deposition.”
Five pages, with the three answers that changed. Page numbers on every point.
Read what matters first.
“Have we seen this party before?”
Twice — a 2019 lease dispute and a matter you declined in 2022. Both files are flagged for your review.
Twenty years of matters, remembered.
“Draft the engagement letter.”
From your own template, with the scope you described. It waits for a partner’s yes before anything goes out.
Drafted in the firm’s voice.
A day at the firm
Five moments from a normal Tuesday.
A brief with your coffee: today’s matters, what’s due, whose file needs your eyes.
Thursday’s binder was assembled overnight. Two missing signatures are flagged.
It drafted the request to the clerk to move the hearing, and a note to everyone on the matter — both waiting for your yes.
A 60-page deposition comes back as five pages you can use in the meeting — with the page numbers.
Unbilled hours surface. Tomorrow’s filings are queued for your review.
And every step of it happened in one place — the firm’s own encrypted space with us, not a lawyer’s personal account with a consumer tool. Where that place is and how it is kept is set out plainly on where it runs.
The math, per firm instead of per lawyer
The consumer tools bill per seat. A serious user runs past $100 a month in AI subscriptions before any specialist legal tool — and the specialist tools bill per lawyer too, every month, forever. Eight people at that rate is roughly $10,000 a year just for the general apps, with every lawyer’s account a separate door and every question handled under those companies’ own terms.
This is one flat price for the whole firm — no per-person charge, ever. One home under one roof, or one practice in one office, is the first year — however many of you there are. Several homes, several offices, or staff spread across sites: by conversation, scoped in writing. Not sure which you are? Order the first year; if it needs more, our letter says so before anything is final. Your assistant is built on the firm’s own matters and serves the firm alone while the service runs — never pooled, never shared, never used to train anyone else’s model.
Privilege, in plain words
The everyday shape of the service is the answer, so here it is without decoration. The firm’s matters live in your own encrypted space in our private cloud on Amazon Web Services — encrypted at rest with keys held in AWS KMS, encrypted whenever they move, in stores with no public endpoints. When your assistant reads and drafts, it reaches its models through Amazon Bedrock, inside the same boundary, over private networking rather than the public internet — and nothing the models see is used to train them or shared with the companies that make them. The firm’s archive is loaded by hand under a written scope. Nothing is sent or filed without a yes from the firm; approvals expire; everything is logged. And when the service ends, the space and everything in it is deleted, confirmed in writing. We will not give you legal advice about privilege; we will give you the handling in writing, so your ethics counsel is reading terms rather than weighing promises.
Your clients’ business stays the firm’s business. Nothing is sent or booked without a yes from the firm; approvals expire; everything is logged. Never pooled, never shared, never used to improve anyone else’s model — and no ads, no selling information, ever. Every promise, in plain words, is on the Trust page.
Can using public AI tools really affect privilege?
Courts have held that putting client material into public AI tools can waive privilege, because the material is disclosed to a third party under that provider’s terms. Whether a particular use does so in your jurisdiction is a question for your ethics counsel, not for us. What we change is the terms: the firm’s matters live in an encrypted space of their own in our private cloud, handled only under a written scope, never pooled and never used to train anyone’s model, and deleted — confirmed in writing — when the service ends. Your counsel decides whether that meets your obligations; our part is making sure every answer they need is already in writing.
Does the firm’s material ever reach you?
Yes — we run the service, so this deserves a straight answer. The firm’s matters live in the firm’s own encrypted space in our private cloud on Amazon Web Services: encrypted at rest with keys held in AWS KMS, encrypted in transit, in stores with no public endpoints. Model traffic travels private networking inside the same boundary, not the public internet, and what the models see is not used to train them. The archive is loaded by hand under a written scope, and when the service ends the space and everything in it is deleted, confirmed to you in writing.
Weighing it against the apps? Your own assistant vs the AI apps — the honest comparison, including what the consumer tools do well and what their terms cost you.
What it does for a practice — every one of these is in the first year, and none of it is bought separately.
Your practice
What it does for a practice
What the business knows — the institutional memory, searchable, including the parts that left with someone
Every client, remembered — prior conversations, what was agreed, what is next
Where a project actually stands — assembled from the mail, the documents and the meetings, not from a status report
Suppliers, agreements and renewals — who, what, when it ends, what is unresolved
The briefing before you meet someone — role, history, and what you last promised them
The week in review — exceptions, decisions, and the work that has stopped moving
Everything on this list is in the first year. Nothing on it is bought separately, and nothing on it is a plan for later — if it is written here, it works on the day your assistant is ready.
Your own AI. Your own material. One assistant, yours alone.
One price, and one way to begin.
Everything above is in the first year — there is nothing else to buy and nothing to configure. Begin with the first year; if you like, add the Archive Assessment before it.
The first year costs more because it contains the preparation — your archive prepared and loaded by hand under a written scope, and your assistant built on it — as well as the first year of the service; every year after is the service running. You order and pay at checkout; we write back within days — a decline returns every dollar, and until our letter confirms the year you may withdraw in writing. From that letter the year is final, and your assistant is ready on the date the letter names.
We answer in writing, within one business day — and if we cannot serve you well we say so before our letter confirms a year.
Prefer to write to us first?
