Bill against the artefact
An Action Package is scoped work: the defect, the finished copy, the owner, the duration, and how it will be verified. That is a line on an invoice your client can read. A dashboard screenshot is not.
White-label
Fifteen client brands. One flat price. Your name on every page.
$129 a month, whether you run three clients or fifteen. Your logo, your colour, your support email on the PDF — and a public methodology page your client can read without asking you to defend the numbers.
$18.43 per client brand per month at 7 brands. Unlimited seats. Cancel in the dashboard.
The five-second screen
An account manager opens the client list and sees one row per brand. Nothing else competes for attention.
Northgate Dental
REAL CHANGE
Bellwether Coffee Co
NO CHANGE
Harlow & Vance
NOT ENOUGH DATA
Illustrative names. The third row is the one that matters: it is not sorted to the bottom, not coloured red, and not hidden behind a filter. An account manager needs to see “we cannot tell yet” before a client asks.
Why this shape
Per-brand pricing punishes you for winning
A meter that ticks up with every client turns your growth into a monthly cost review. We charge flat so nobody at your agency ever has to ask whether a particular client is “worth” a workspace this month.
Defensible in the room
When a client asks “how do you know?”, you send them our methodology page. It states the thresholds, the test, and what the method cannot tell them. You do not have to become the expert witness.
The series is the deliverable
Measurements accumulate into a longitudinal record per client, and it exports whole — CSV and JSON, raw runs included, during the subscription and after it. If you leave, the history leaves with you.
White-label
What you may change, and what you may not
The second list is the important one. Every vendor lets you swap a logo. If an agency could also delete the uncertainty, our promise would be a per-tenant setting — so it is not one.
Yours to set
- Your logo, on every page of every PDF
- Your accent colour, applied to the report’s single signal colour
- Your agency name and support email in the footer
- A footer note of your own wording
- Removal of all CitedProof branding — the report can carry no trace of us
Not negotiable
- The three verdict strings. REAL CHANGE, NO CHANGE, NOT ENOUGH DATA are not editable copy
- The hatch. Missing data is drawn as missing, never as a zero or an empty bar
- Evidence grades. A grade-D finding prints as grade D in your client’s copy
- The printed sample size. Every rate carries the runs it came from
- The methodology link. It stays in the footer even when our name does not
This protects you more than it protects us.
The day a client checks a claim you made on their behalf and finds the sample size printed next to it, you become the
agency that does not overstate. That is worth more than a removable footer.
Published prices, dated
What white-label costs elsewhere
Vendor list prices as published on their own pricing pages and read on 30 July 2026. Prices change; check them before you quote us on this. We have picked the tier at which each vendor first allows white-label or multi-client work.
| Vendor | Tier | Price / month | Note |
|---|---|---|---|
| CitedProof | Agency | $129 | 7 brands, white-label included, unlimited seats |
| Rank Prompt | Agency | $119.17 | White-label included; 1,000 credits, where 1 credit = 1 prompt analysis |
| Rankscale | Growth | $385 | White-labelled dashboards, 50 dashboards, 5,500 credits |
| Profound | Growth | $399 | 3 engines, 100 prompts, 3 seats; no API on this tier |
| Trakkr | Scale | $500 | 10 brands × 50 prompts, white-label portals, unlimited seats |
| Writesonic | Agency Starter | $200 | White-label not included at this tier; requires Agency Enterprise |
The number worth doing arithmetic on is the credit model. A 1,000-credit agency plan where one credit buys one prompt analysis, spread across the “500 brands” the tier advertises, resolves to two prompt-runs per brand per month. Two runs cannot separate a change from noise on a surface where three parses of the same query agree on 9.2% of URLs. That is not a cheap version of measurement; it is a different thing wearing its clothes.
We are new and we have no customers to point at. Comparing on published prices and stated limits is the only honest comparison available to us, and it is the only one we will make.
Start here
Run it on one client first
Pick the client you would least like to be surprised about. We run their AI Access Report free, and you decide from the finding, not from a demo.