A pilot on your own matters

A pilot is how a practice group tries this on its own docket instead of our worked examples. It runs for three months against a defined set of matters, with the criteria for success written down before the first simulation runs, so that at the end there is something to judge rather than an impression to argue about.

Reading a worked brief costs nothing, and credits are $50 each in any quantity you like, starting at one, bought with a card on a firm email address and all on the pricing page. A pilot is the step after that, for a practice group that wants several live matters covered at once with criteria agreed before the first run, and it is $2,500 to $5,000 for the three months.

What a pilot is

Three months. Long enough to cover a real expert-disclosure cycle, short enough that a decision is never far away. It does not auto-renew.
A named set of matters. Chosen by you at the start, usually four to eight. Live matters, not samples. If a new matter lands mid-pilot and you want it in, it goes in.
Named people. One partner who owns the outcome, and the associates who will actually run it. A pilot with no named user is a pilot that quietly does not happen. Anyone else at the firm can use it too, and nobody has to be on the list to sign in.
A kickoff and a close. One session to set the matters and the criteria, one at the end to go through what the runs produced against them.

What you provide

The matters. A caption and a docket number is enough to begin; everything else can come from the public record.
Optionally, your documents. The served report, the complaint, the expert disclosure. These improve the output materially, and they are handled as privileged material end to end. You are never required to upload anything.
An hour at each end. Kickoff and close. Between them the work is yours to run whenever you want it.

What you get

The four sections on every matter you run: the opposing expert’s likely opinion, the cross-examination weaknesses, the Rule 702 attack surface, and the settlement-range implications.
Every supporting citation checked verbatim against the public record, with the passage the run relied on readable in place. What cannot be verified is dropped rather than guessed.
Experts shaped around your docket: the expert classes your matters need come first.
An hour of onboarding for every person who will use it, one to one rather than a group session nobody remembers, and run by the founder.
Two hours of founder time a week, held for you for the length of the pilot and yours to book or not. A live matter, a second read, a question about what a run produced.
Priority on your requests. What your matters need comes first, and you see it before anyone else does.

How we agree it worked

The criteria are set at kickoff, in writing, and they are yours to choose. These are the ones we would suggest, because each is answerable without anybody’s opinion:

Did it name what the matter actually turns on? Scored against your own read of each matter, recorded before the run.
Did any citation fail to check out? A fabricated authority is a failed pilot, and we would rather you find one than not look.
Did the cross-examination section produce a question you had not written? The one that is hardest to fake and easiest for a litigator to judge.
Would the team run it on the next matter without being asked? The only measure that predicts anything about renewal.

Your data

Inference runs server-side on the leading general-purpose models, under zero data retention with US inference, so prompts and completions are neither stored nor used to train anything. Matter content is never used to train any model, ours included, privileged content is not written to logs, and a deleted matter is gone after a 30-day recovery window. A signed data processing addendum recording all of that is available with a pilot. The detail is on Security & trust, and the commitments are governed by our Privacy Policy and Terms of Service.

This is internal preparation, which is usually the first thing a firm’s general counsel asks about. Nothing it produces is filed, disclosed under Rule 26, or offered as a testifying opinion. Your retained expert still signs the report and still takes the stand.

Commercials

$2,500 to $5,000 for the three months, which is 50 to 100 credits at the same $50 a credit as everywhere else. Scoped to how many matters you want covered and not to the size of the firm. Invoiced, so it can sit on a matter’s expense schedule rather than on a card.
The credits are at list, and deliberately so. What the pilot buys that a card purchase does not is the onboarding, the founder time and priority on your requests. Volume pricing comes with an annual commitment, not with a trial.
Credits do not expire. Anything unused at the end of three months stays yours and carries into whatever comes next.
The balance is the firm’s. Anyone who signs in on your firm domain draws on it, so a colleague who joins in month two is already covered.
No auto-renewal. The pilot ends on its end date unless you decide otherwise.

How it ends

Three outcomes, and all three are fine. You convert to a firm account and keep going. You extend, because the matters you wanted to test were not the ones that moved. Or you stop, in which case we ask for twenty minutes on why, and your unused credits remain available if you come back.

How to start

Read a worked matter first. The sample briefs are complete and free, and they carry the same output your own matters would. If it holds up, one call is enough to scope the matters and the criteria, and the paperwork is a one-page order form against our standard Terms.

Book a demo, or write to richard@suprememind.ai.