pricing-fact both

What the numbers on a quote mean

Two totals that add: cash at order, and year one all-in. Planning bands until a paid recount. The monthly fee is stewardship, not rent. We do not mash those totals into one number. Public figures stay labeled PLANNING until the metadata recount after you sign.

This node What the numbers on a quote mean required-by How the planning band is computed required-by Day-14 terms

Cash at order is what you fund to start. Year one all-in is start plus the first year of keep-good. Both are printed. Hiding the run-rate is the math we attack.

On a worked Standard example (hybrid, mid-market): $30,000 install + about $12,000 of hardware you buy ≈ $42,000 cash at order. Plus about $3,000 a month stewardship ≈ $78,000 year one. Design-partner install is half ($15,000) while that cohort is open. Hardware is yours, not a marked-up line.

The install fee is the burst — reading history into the graph. The monthly fee is the work of keeping an owned brain good: updates, model stewardship, connector upkeep, graph health, classification review, support. Cancel and the brain stays.

The worked example was sized at one mid-market corpus point, so it is published as a ceiling, not a slope. It applies up to the mid-market envelope — about 25B census planning stock, roughly 255 knowledge workers with several years of history. Above that the published band stops applying and the install is quoted from your census. The gate is attested planning stock, not headcount: 600 people with two years of history still fits the worked band; the same team with five years does not.

The pricing page draws an illustrative chart by company size, $50M to $250M. Only the $50M column shows the published worked band. Above it, each column is a planning estimate: the $30k install scaled pro-rata from the ~17B corpus it was sized at (about $51k at $100M, about $129k at $250M), plus about $12k of hybrid-default hardware and about $36k of first-year stewardship. Those numbers are a picture, never the assessment's output — a company with lighter history prices lower, and the measured census sets every real quote. Beyond $250M is always custom.

Greenfield has no public price card on this site. Its planning bands print on the assessment when the numbers fit.

Claims

What this page stands behind.

#claim-two-totals

Quotes show two totals that add: cash at order, and year one all-in.

Applies to
Standard and Greenfield
Evidence
Neither number is a typo. They answer different questions.
Since
2026-08-14
#claim-standard-bands

Worked Standard example: about $42k cash at order, about $78k year one. Planning bands, never a single fixed price.

Applies to
Standard and Greenfield
Evidence
$30k install + ~$12k hybrid hardware; plus ~$3k/mo × 12. Cloud-native hardware is $0.
Since
2026-08-14
#claim-ceiling

The worked Standard example applies up to the mid-market envelope (about 25B census planning stock). Above that the published band stops applying and the install is quoted from the census. The gate is attested planning stock, not headcount.

Applies to
Standard
Evidence
The $30k/$42k/$78k figures were derived at one mid-market corpus point (~150 knowledge workers, a few billion tokens). 600 people × 2 years of history = ~19B still fits the worked band; × 5 years = ~49B is quoted. The chart and the assessment gate on the same imported ceiling constant.
Since
2026-08-26
#claim-size-illustration

The company-size chart is illustrative. Only the $50M column prints the published worked band; larger columns are planning estimates that scale the $30k install pro-rata from the ~17B corpus it was sized at, plus about $12k hybrid-default hardware and about $36k first-year stewardship. They are tagged planning estimates, never the assessment's output — the measured census sets every real quote.

Applies to
Standard
Evidence
Bracket assumptions: ~150 knowledge workers per $50M of revenue and 6 years of history — labelled, not the gate. Rule: install = $30k while the implied corpus is within the worked envelope, then $30k scaled pro-rata from the ~17.01B corpus the $30k was sized at (≈$51k at $100M, ≈$129k at $250M), each rounded to $1k; hardware shown at the ~$12k hybrid default; ~$36k first-year stewardship unchanged. These figures never enter the assessment — a leaner-history company reads lower by tokens, and the measured census sets the quote.
Since
2026-08-26
#claim-install

The install fee is the burst — reading history into the graph.

Applies to
Standard and Greenfield
Evidence
One-time, for history already on disk. Greenfield skips the install fee when the day-zero census is under the planning cap — existing history is measured at sale, not assumed empty.
Since
2026-08-14
#claim-stewardship

The monthly fee is stewardship of a brain you already own, not rent on the loop.

Applies to
Standard and Greenfield
Evidence
You own the install from day one. The fee buys continued vendor work. Cancel, keep the brain.
Since
2026-08-14
#claim-planning

All public figures are planning bands until a paid metadata recount.

Applies to
Standard and Greenfield
Evidence
The assessment prints a band from numbers you supply. Measurement comes after you sign.
Since
2026-08-14
#claim-own-asset

Verified, receipted work is designed to write back into a map you keep, even if you cancel. No amortization schedule and no ROI dollar figure is claimed.

Applies to
Standard and Greenfield
Evidence
Write-back is a design property. The compounding curve is not a measured benchmark, so none is claimed.
Since
2026-08-20
#claim-category-scan

A two-round competitive scan found no named vendor offering all five of: runs in your environment, you own the knowledge graph, facts carry provenance, products operate on top, and mid-market pricing.

Applies to
Standard and Greenfield
Evidence
The cleanest checkable gap is pricing: the only sovereign alternative we found runs one to two orders of magnitude over mid-market, per its own public disclosures.
Since
2026-08-20

Machine-readable → index.json · /ai-info · llms.txt

browse The rest of the brain