How much should AI search optimization cost?
There is survey data on what local SEO actually sells for, and it puts most "AI optimization" pricing in context. Here are the bands, and what should make you walk away.
"AI search optimization" is new enough that pricing is all over the place — the same scope has been quoted at $300 and at $4,000 a month. The underlying work overlaps heavily with local SEO, which has actual survey data, so that is the sensible place to anchor.
What the market charges
Ahrefs surveyed 439 SEO providers on what they charge for local SEO:
- the average monthly retainer came out around $1,557
- the most commonly reported band was $501–$1,000/month, cited by about 20.4% of providers
- one-off audits in the US entry tier typically run $500–$1,000
Some useful conclusions:
Under about $300/month, the arithmetic does not work. Not as an insult to cheap providers — as arithmetic. Weekly tracking across four engines, a report somebody read before sending, and any human judgement at all does not fit inside that number unless something is being skipped.
Above roughly $2,000/month for tracking alone, you should be getting execution too — content produced, citations placed, things actually changed — not just a dashboard and a call.
A one-time audit is worth $500–$1,000 and is a reasonable way to start, with one caveat that matters more here than in traditional SEO.
The caveat that is specific to AI search
AI answers are not stable. Orbit Media re-ran 72 identical prompts two days apart and found Claude kept only 30% of its previously-cited URLs, Gemini 38%, Perplexity 65%. Across 1,792 query and domain combinations, all four engines agreed on the same source 1.7% of the time.
So a one-time audit of AI visibility is measuring something that will have moved by the time you read it. That does not make it worthless — a snapshot of citation sources and entity problems is still useful — but it does mean a single number with a date on it is not a measurement of anything durable. Price it as a starting point, not as an answer.
What should make you walk away
A package built around llms.txt. Two studies — one across 300,000 domains, one across 137,000 sites — found no measurable effect on citation frequency, and Google has said it is not needed. It is the cheapest deliverable in this field to produce and it demonstrably does not work.
A single "AI visibility score" with no tracking behind it. See the churn data above. You can disprove it yourself by asking the same question three days running.
A report on one engine. They agree 1.7% of the time. A report on ChatGPT is a report on ChatGPT.
No transcripts. If you cannot see the actual queries and the actual answers with dates attached, you cannot check any of it, and the whole value of measurement is that it is checkable.
A price that changes after the discovery call. That is a pricing strategy, not a scope.
What we charge, for comparison
We publish ours, which is itself a position: $750 for a diagnostic delivered over 21 days with three tracked runs across four engines, and $550/month for ongoing weekly tracking with a three-month minimum.
The $550 sits deliberately inside the most common band in that survey rather than under it. It works at that price because the measurement machine costs us the same whether it is running for one client or thirty — so we sell the measurement and the instructions cheaply, and price the execution work separately rather than pretending it is free.
Sources: Ahrefs SEO pricing survey (439 providers); Orbit Media, "AI Citation Sources" (13,184 citations); the 300,000-domain llms.txt analysis and Ahrefs' study of 137,000 sites.