Pricing

Two numbers. Published. Fixed.

Fixed fees, agreed before we start. No hourly billing, no success fee, no scope surprises. Everything below is the whole price list — there is no enterprise tier behind a form.

Onboarding · mandatory · once

Baseline Assurance Sprint

$15,000fixed fee
4 weeks · 75–100 invariants and adversarial challenge queries

The mandatory first four weeks of every engagement. Sets the pre-frozen schema denominator, calibrates severity and produces the first signed bundle.

  • Metadata intake: dbt manifest.json and INFORMATION_SCHEMA, read-only, or your own local extractor
  • Automated invariant synthesis for the standard relational families (target: ≥70% of the suite in under 15 minutes)
  • Targeted human authoring of the proprietary high-consequence logic (Type C fixtures)
  • Consequence-tier threat model and pre-frozen assurance coverage denominator, signed by you and our lead methodologist
  • First in-perimeter run, adjudication and 3-tier readout to the CDAO
  • Incremental Consequential Finding Yield (ICFY) measured against your native evaluation
How the sprint runs
Year-1 ACV
$63,000

Year two and onward: $48,000 — which on our published model observable hard value alone very nearly covers, before any risk-adjusted value is counted. Year one carries the one-off sprint on top. Run the model with your own numbers →

Terms that matter

What the contract says, in plain words

Platform subscription billed annually upfront; onboarding sprint billed on card at signature.

Enterprise procurement typically runs 60–120 days on Net 30/60 terms; the sprint is structured to start before that cycle completes.

No success fee of any kind. Whether the agent passes or fails changes nothing we are paid.

Customer grants a limited licence to process metadata solely to deliver the contracted service. No derivative learning on customer data unless you switch it on.

Scoped additions

Priced before you need them

AdditionPriceUnit
Additional agent on the same warehouse$12,000per year
Second warehouse dialect (Databricks, BigQuery, Redshift)$18,000per year
Benchmark intake validation of customer-supplied golden truth$600per engagement, capped
Mode 3 restricted VDI review session$2,400per expert day

Benchmark intake validation exists because validating customer-supplied golden truth is not zero cost; it is capped and disclosed rather than buried in the sprint fee.

Where the published price stops

The published price covers up to three agents on one warehouse, against a pre-frozen surface of up to about 500 dbt models. Past that the engagement is scoped — more invariants, more adjudication, more reviewer time — and we publish the agreed figure in your order form rather than negotiating it privately.

There is no hidden enterprise tier: a scoped engagement is still a fixed fee agreed before work starts, still carries no success fee, and the figure is written into the order form.

Design partner cohort

3 slots. A test, not a discount.

The first three engagements run under the 90-day falsification programme: two paid sprints, then a binary conversion decision to the platform. Card-billed, upfront, on the Common Paper Design Partner Agreement v1.3 so your counsel has nothing new to read. In exchange we ask to be named as a design partner and to publish a case study you approve line by line.

The kill rule

If neither of the first two design partners converts to the recurring platform because one-time audits are sufficient, we halt the platform build. We say this here so you know the offer is a test, not a pitch.

The 90-day programme in full →

Why these numbers

The sprint fee is anchored to the work it takes: metadata intake and validation, invariant synthesis, targeted human authoring of the fixtures only a person can write, one in-perimeter execution with risk-adaptive trials, adjudication of every exception by a qualified reviewer, and the signed bundle at the end — around 50 hours of senior assurance engineering in Sprint 1, which we report to you as an actual number. The platform fee is anchored to keeping that suite alive: re-synthesis on every schema change, execution on every pull request and every week in production, Tier-3 review on exceptions, and a reviewer roster paid at published rates.

We publish this because the alternative — a “contact sales” button and a number that depends on how the call went — is precisely the kind of unfalsifiable claim the rest of this site refuses to make. The price is the price; what changes is whether the findings are worth it, and the sprint exists to answer that in four weeks.

Compensation on the evaluator side of the market is published too: see the Evaluator Guild.

Questions

About the money

Because the platform is only as good as its denominator. The sprint freezes the schema surface, calibrates severity for your consequence tier and produces the first signed bundle. Skipping it would mean charging you for continuous measurement of a surface nobody agreed.

No. A per-call rate for a system whose judge–human agreement and latency we have not yet published would be exactly the unbacked number this site argues against. The platform fee covers the suite, the runs and the reviewers.

Never. Financial independence is the first clause of the Independence Charter: whether the agent passes or fails changes nothing we are paid.

Scoped additions are published below: $12,000 per year for another agent on the same warehouse, $18,000 per year for a second dialect.

The platform is billed annually upfront and recognised ratably. The sprint is billed on card at signature so procurement is not on the critical path for the first four weeks.

Start with the sprint. Decide with the number.

Four weeks, $15,000, one signed bundle and an ICFY figure. Then a binary conversation.

Book a readiness review Open the ROI model