Start from your delivery cost
Whatever you charge, you need to know the floor. On pay-per-use tooling the cost of a full pass over a client site is concrete:
| Piece of work | Credits | What the client gets |
|---|---|---|
| Site Check | Free | Crawler-by-crawler access. Whether engines can read them at all. |
| AI Visibility run, 10 prompts × 6 engines | 15 | Where they stand on the questions their buyers ask. |
| URL Audit | 5 per page | Per-engine score, the evidence, a ranked fix list. |
| Gap Analysis | 5 per page | Their page against the pages that won. |
| Page Rewrite | 10 per page | A publish-ready draft from their own content. |
| Audit + Rewrite together | 12 per page | The pair, below the separate prices. |
A seven-page pass at 12 credits a page is 84 credits — about $85 at bundle rates, since $1 = 1 credit. Add a ten-prompt tracker run and the tooling cost of a substantial first engagement is roughly $100.
We are not going to tell you what to charge your client. That depends on your market, your existing rates and what else you do for them. We will tell you what it costs you to deliver, because that is the number you actually need and the one most people are guessing at.
Three models that work
Fixed-scope audit
A defined deliverable: access check, visibility run across a set number of questions, audits on the pages that matter, and a prioritised fix list. Easy to sell because the client knows exactly what arrives. Easy to price because you know the cost.
Audit plus implementation
The audit, then the rewrites, then a re-run to show movement. Higher value because you are selling the outcome rather than the diagnosis — and the re-run is what makes the result visible.
Retained monitoring
Monthly re-runs with a short report on what moved. Lower ticket, recurring, and it works best once you have already done the first engagement and have a baseline to compare against.
Most agencies find the second model easiest to sell first, because the client can see the before and after rather than buying a report.
What to include so the number is defensible
The strongest thing you can do for the price conversation is make the deliverable concrete. Four artefacts do that:
- An access verdict — can AI engines reach and read the site at all.
- A visibility baseline — where they stand, question by question, engine by engine.
- A gap comparison — their page beside the pages actually being cited.
- A prioritised fix list, and the drafts — not recommendations to act on later.
The fifth thing renews the engagement: a re-run thirty to sixty days later showing what moved.
What not to promise
Do not promise citations, rankings or traffic. Nobody can deliver those reliably, and a client who has been promised them will hold you to it.
Promise the diagnosis, the fix and the re-measurement, all of which you control. Clients tend to trust a service more when it is clear about what it cannot do — and it makes the rest of your claims credible.
Common questions
Is there a standard rate for this yet?
No. The service is too new for a going rate, which is why pricing from your delivery cost and your existing rates is more defensible than copying someone else's number.
What does it cost me to deliver?
On pay-per-use, a seven-page site pass is about 84 credits - roughly $85 - and a ten-prompt visibility run is 15. So the tooling cost of a substantial first engagement is around $100.
Should I charge for the Site Check?
It is free to run, so it works better as a pitch opener than a line item. Running it live on a prospect's site during a call is a strong way to start the conversation.
How do I price ongoing monitoring?
Match it to your existing reporting cycle rather than inventing a new one. Monthly re-runs of a stable question set are a small, predictable credit cost you can build a retainer line on.