engagement · 04/pilot

Thirty days. One product. In production.

Every Danielle engagement starts with a 30-day pilot on one of our named products — Margin, Basin, Ledger, or Chart — or the Harness itself. We run it on your real work, in your environment. Then you decide whether to roll out.

Connect with the teamRead the harness brief →
01/how it runs

Four weeks. Four checkpoints.

WEEK 0
STAGE 01

Calibrate

You send us 20 redlined contracts (or, for a matter-based product, a closed matter we can replay). We extract your positions into a versioned playbook.

playbook v0.1
WEEK 1
STAGE 02

Shadow

Danielle runs in parallel with your team. Every output is compared, side-by-side, to what your reviewer did.

shadow run
WEEK 2
STAGE 03

Tune

We refine the playbook against your feedback. Rules get sharper, exceptions get encoded.

tuning
WEEKS 3–4
STAGE 04

Decide

Danielle handles real work end-to-end. You measure speed, accuracy, accept rate. You decide whether to roll out.

production
02/scope
we commit

We commit

  • A calibrated playbook against your last 20 redlined contracts (or one closed matter)
  • End-to-end runs for 30 days, on your real workflow
  • A side-by-side accuracy & turnaround report at end of pilot
  • Zero data retention — your documents do not train any model
you commit

You commit

  • One named workflow owner on your side
  • 20 redlined sample contracts (or one closed matter) at week 0
  • 30 minutes a week for a checkpoint call
  • A reviewer's hour to score the side-by-side report
03/measurement

We score the pilot before it starts.

≥ 90%
TARGET · first-pass accept rate · vs human
≤ 10 min
TARGET · median turnaround · per artifact
100%
TARGET · outputs with source citation
≥ 4×
TARGET · speedup vs current process
04/security

Your work is your work.

Pilots run under NDA. Documents are isolated to your tenancy. Nothing trains a model. Outputs are signed against a versioned playbook, so the same input always produces the same output. Audit logs are exportable.

zero retention

Zero retention

Documents and outputs not used to train models.

tenancy isolation

Tenancy isolation

Per-customer encryption keys. No shared corpora.

audit-exportable

Audit-exportable

Every decision journaled with source citation.

soc 2 in progress

SOC 2 in progress

Type II audit underway. Reports under NDA.

05/how pricing works

Three ways to pay. Real numbers on the first call.

01 · PILOT
Flat-fee 30-day pilot

One scoped engagement on one workflow. Calibrate the playbook, run on real work, ship the side-by-side report. No conversion obligation.

02 · PRODUCT
Seat or matter license

For ongoing document work (Margin): seat-based. For matter- and portfolio-based products (Basin, Ledger, Chart): per-matter or per-portfolio.

03 · PLATFORM
Harness license

License the Harness directly. Run our playbooks, or build your own. Per-instance, with a deployed-in-your-tenancy option for regulated buyers.

We share numbers once we understand the workflow and the volume. Most buyers land somewhere between "less than one experienced reviewer's loaded cost" and "less than one outsourced engagement."

06/common questions

Before you ask.

What does it cost?

Pilots are scoped flat-fee. We share pricing on the first call once we understand the workflow.

What if I don't have 20 redlined contracts?

We can work with as few as 8, or build a starting playbook from your defaults and tune in week 1.

Where does our data live?

Inside our tenancy on us-west-2 by default. Other regions or your tenancy on request.

Can we keep our outside reviewers in the loop?

Yes. Outputs are tracked-changes documents; outside reviewers see Danielle's work the same way they see a junior reviewer's.

What languages?

English today. Spanish and French on request for pilots starting Q3.

What if the pilot fails?

You walk. No conversion obligation. We learn something either way.

Start with one real workflow.

team@danielle.legal

Danielle Legal, Inc.