An owner sat across from me a few weeks ago with a proposal from another shop open on his laptop — $4,200 a month for "AI visibility management." He wasn't objecting to the number. He was asking a much better question: "What am I actually paying for? What gets built with that money that I couldn't build myself?" I didn't have a slick answer ready, because most agencies don't make you ask that question — they just show you a dashboard and a case study logo wall. He deserved a real answer, so I gave him one, line by line, the same way I'm going to lay it out here.
That conversation isn't unique to him. It's the same conversation I've had, in some form, with almost every owner who calls us after getting a GEO or "AI SEO" quote for the first time. The number varies. The confusion about what it buys does not.
There is no toll booth you pay an AI engine to get cited. Citation is earned by making your facts extractable, structurally sound, and consistent everywhere they appear — and that work has a real, calculable labor cost. Agencies charging $1,500 to $25,000+ a month are mostly selling the discipline to do that work continuously, not a secret input you can't access yourself.
What "being the source" actually means
When ChatGPT, Perplexity, or Google's AI Overviews answer a question and cite you, they're not rewarding a payment — they're pulling a passage of your content that was clear, factual, and structured well enough to quote directly. Google's own guidance on structured data and its documentation on writing for people-first, helpful content both point at the same underlying requirement: content has to be legible to a machine as well as a human before it can be surfaced as an answer. Anthropic and OpenAI's crawler documentation — ClaudeBot and GPTBot — describe fetching and indexing raw HTML, not evaluating who paid for placement. "Being the source" is a content and structure problem, not a media-buying problem.
That's the gap most pricing conversations skip. The owner across the table from me wasn't buying access. He was trying to figure out whether he was buying labor he could do himself, or labor he genuinely couldn't.
What actually has to get built
Strip away the jargon and the deliverable list for real AI-visibility work comes down to five categories of work, every one of them things a business owner can inspect and, in most cases, do:
- Passage-level rewrites. Turning a page's key facts — price, process, differentiator — into short, self-contained passages an engine can lift without misquoting you. This is closer to technical writing than to marketing copy.
- Schema markup. Structured data per the Schema.org vocabulary that tells machines explicitly what an entity, price, or answer is, instead of leaving it to be inferred from layout.
- Fact consistency across the web. Your hours, pricing, service area, and founding story matching across your site, Google Business Profile, and any third-party listing an engine might cross-reference. Inconsistency is one of the most common reasons AI engines hedge or omit a brand entirely.
- Crawler accessibility. Confirming GPTBot, PerplexityBot, ClaudeBot, and Google-Extended can actually fetch your pages — see our 20-minute crawler audit for the exact checks.
- Monitoring and re-testing. Running the same prompts across engines on a schedule to see whether the fixes moved anything, and repeating the whole cycle as engines retrain.
None of those five require an agency by definition. All five require time, and most require a level of consistency that a solo owner running a business rarely sustains past the first month.
The DIY cost, priced honestly
If you did all five categories yourself, the cash cost is close to zero — maybe a schema validator or a rank-tracking tool at $50 to $200 a month. The real cost is time, and it is not small. Rewriting a single page's key passages into extractable, quotable form takes a focused hour or two per page done properly, not the five minutes it looks like from the outside. Fixing schema on a dozen key pages is a weekend if you already know JSON-LD, a lot longer if you don't. Chasing down fact inconsistencies across your Google Business Profile, your site, and every directory that has your old address is the kind of tedious work that gets postponed indefinitely once a slow week ends.
Multiply that across a real site — twenty, fifty, a hundred pages — and the DIY path is real work measured in weeks, not hours, before you've even started the monitoring cycle that has to repeat every time an engine updates. That's not a knock on doing it yourself. It's the honest cost, priced in your own hours instead of an invoice.
What the retainer price is actually buying
Once you see the five categories above, the retainer pricing stops being mysterious. A $1,500 to $5,000 a month engagement for a small business typically buys someone else doing the passage rewrites, the schema, and a monthly re-check across a handful of key pages — the same labor you could do yourself, minus your own time. A $2,000 to $25,000+ a month full program adds ongoing monitoring across many more pages, competitor tracking, and the discipline to keep testing prompts every month whether or not anything obviously broke.
What you are not paying for, no matter the price tier, is a backdoor into an AI engine's index. There isn't one. Every legitimate practitioner is running the same five categories of work against the same public crawler documentation you can read yourself. The retainer price is a labor and consistency premium, not a secret-handshake fee.
| What you're comparing | DIY | Small retainer | Full program |
|---|---|---|---|
| Cash cost | Near $0 (tools only) | $1,500–$5,000/mo | $2,000–$25,000+/mo |
| Time cost | High — your own hours | Low — you review, they build | Near zero — full delegation |
| Pages typically covered | As many as you get to | A handful of key pages | Site-wide + competitor tracking |
| Monitoring cadence | Whenever you remember | Monthly | Continuous |
When DIY is the right call
If you're a personal brand or a small local business with fewer than a dozen pages that actually matter, doing the work yourself is entirely reasonable — you already know your own facts better than anyone you'd hire, and the volume of pages is manageable in a focused weekend or two. Start with the free version: run your own prompts across ChatGPT, Perplexity, and Google, see where you're missing, and fix the highest-impact gaps first using our DIY AI visibility audit as the checklist.
The moment DIY stops making sense is the moment the work stops getting done — when month two arrives, the business gets busy, and the monitoring cycle that's supposed to repeat quietly stops happening. That's not a failure of will; it's what happens to any recurring task that isn't someone's actual job.
When paying for it is the right call
Pay when the number of pages that matter is large enough that consistency, not cleverness, becomes the bottleneck — multi-location businesses, e-commerce catalogs, or any brand where the same five fixes need to happen across dozens or hundreds of URLs and stay maintained. Pay also when the category is competitive enough that losing a citation to a competitor has a real, calculable cost, and a ranked, expert-reviewed plan is worth more than the hours it would take you to build one from scratch.
Either way, judge any quote against the five categories above. If a proposal can't tell you, in plain language, which of those five it covers and on how many pages, you're not looking at a scoped engagement — you're looking at a subscription to a dashboard.
The teardown: pricing a real page
Take a single service page. Passage rewrite for the two or three facts worth quoting: roughly 60–90 minutes done carefully. Schema markup for that page type: 20–40 minutes once you know the vocabulary, longer the first time. Fact-consistency check against your Google Business Profile and top three directories: 15–20 minutes. Crawler accessibility check with curl: under 5 minutes using the method in why your best page is invisible to AI crawlers. Call it two to three hours of focused, non-glamorous work per page, the first time through. That's the real unit economics behind every retainer quote you'll ever receive — the price is that per-page labor, multiplied by however many pages, plus the discipline to repeat it.
Questions people ask
There is no fee you pay an AI engine to be cited — citation is earned by having content structured and factual enough for the engine to quote and trust. The real cost is the labor: rewriting pages into extractable passages, fixing schema, keeping facts consistent across your site and third-party listings, and repeating that work as your business changes. Agencies charge $1,500 to $5,000 a month for a small business engagement, and $2,000 to $25,000+ a month for a full program, largely to cover that ongoing labor.
It is cheaper in cash and more expensive in time. A business owner who already understands their own facts, offers, and customers can do the extraction and structuring work themselves for the cost of their own hours plus small tool costs. What an agency actually sells on top of that labor is consistency across dozens of pages, ongoing monitoring as engines change, and the discipline to keep doing it after month one — which is exactly where most solo attempts stall.
A one-time audit is a snapshot. AI visibility is a moving target because engines retrain, re-crawl, and change what they cite on an ongoing basis, so the retainer price is really paying for repeated monitoring and rewriting, not a single deliverable. That is also why the honest agencies price it as a monthly retainer instead of a flat project fee — the work does not end when the first page is fixed.
Want this done for you?
Want to know if AI is recommending you or your competitor? Get an AI visibility audit from Acromatico.
Get a free AI Visibility Audit →