The honest guide to AI SEO services and AI visibility

What are AI SEO services, really?

AI SEO services help your brand get surfaced and cited inside AI answers — Google AI Overviews, ChatGPT, Perplexity, Claude, Gemini — instead of just blue links. That is the honest definition. The dishonest version, sold by most agencies, frames it as a mysterious new discipline that requires a new vendor and a new line item.

Here is the quiet part we say out loud: about 80% of what gets billed as "AI SEO," "GEO," or "AEO" is classic SEO fundamentals done right — crawlability, clean structure, real authority, fresh content. We run an engine across more than a dozen brands, and the pattern holds consistently. The genuinely new 20% is four specific things, and that is where your money should go. If a proposal does not name those four, you are paying premium rates for fundamentals you could have fixed for free.

We get into the terminology fully in how GEO, AEO and SEO differ, but the short version: SEO is being findable, AEO (answer engine optimization) is being the answer, and GEO (generative engine optimization) is being cited inside a generated answer. They overlap heavily. Treating them as three separate purchases is how vendors triple a single invoice.

The 20% that is actually new

The four things that genuinely changed with AI search are extractability, cross-web consistency, third-party citations, and mention measurement. Everything else is SEO you already knew about.

• Extractability. AI engines lift clean, answer-first, parseable sentences. Lead every section with the direct answer, then support it. Burying the answer in paragraph four means the model never finds it. This is the cheapest lever and the most ignored.
• Cross-web consistency. The same entity facts — your name, what you do, who you serve, pricing — need to match everywhere AI reads: your site, your directories, your profiles, third-party write-ups. Models triangulate facts across sources. Contradictions get you dropped.
• Citations on trusted third-party sources. This is the real driver, and the one sub-$1k programs skip. AI engines show a systematic bias toward earned and third-party media over brand-owned content. Schema alone does not generate recommendations; third-party mentions and review volume do. We cover the mechanics in how to earn ChatGPT citations.
• Measuring mentions inside answers. Rankings are the wrong scoreboard now. What matters is how often you get named, and whether the model cites you or a competitor. We will show you how to track this without a $500/mo tool below.

Why does AI recommend my competitor and not me?

AI recommends your competitor because the engine found more consistent, trusted, third-party evidence about them than about you. AI answers typically name only three to five brands. Perplexity, for example, uses retrieval-augmented generation: it reads a set of pages per query and cites only a handful of them. If your competitor occupies those slots, it is because trusted sources mention them by name and the facts line up.

Three common reasons you are left out: your content is not extractable (the model cannot cleanly pull an answer), your facts are inconsistent across the web (the model cannot trust the entity), or you have no third-party citations (nobody credible vouches for you). Notice none of those are fixed by schema markup alone — schema helps machines read you, covered in why schema is the language AI reads, but it does not make them recommend you.

Can a brand-new startup with no authority rank in AI answers?

Yes, but you have to solve the cold-start problem honestly, and most guides pretend it does not exist. A startup has no domain authority and no earned media — the two things AI engines lean on hardest. So you start where you can win: long-tail, specific queries. Longer, more specific queries are substantially more likely to trigger an AI Overview, and the brands named in those niche answers face far less competition than head terms.

The cold-start play for a startup: publish genuinely extractable, answer-first content on the narrow questions your buyers actually ask; get your entity facts consistent across every profile and directory; and earn a handful of citations on trusted third-party sources in your niche — a podcast, a roundup, a partner's blog, a real review. That is slow, unglamorous work. It compounds. Budget 9 to 12 months before the returns show up. Anyone promising faster is either lucky or lying.

How do you actually measure AI visibility without buying a tool?

You measure AI visibility by running a fixed prompt set across the engines on a schedule and recording how often you get mentioned and cited. You do not need a $500/mo platform to start. Here is the manual method we use before any client buys tooling:

1. Write 20 to 30 prompts a real buyer would type — "best X for Y," "how do I solve Z," "alternatives to [competitor]."
2. Run them across ChatGPT, Perplexity, Claude, and Gemini on a fixed cadence (we do weekly). Use fresh sessions so memory does not skew results.
3. Record three numbers per prompt: mention rate (are you named?), share of voice (what fraction of named brands are you?), and citation frequency (does the answer link you vs. a competitor?).
4. Configure GA4 to capture AI referrers so you can see actual click-throughs from AI surfaces alongside the mention data.

That gives you a real baseline in an afternoon. Tools automate this at scale and are worth it later — but a vendor who cannot explain this manual method, or who hides the baseline behind their dashboard, is hiding how little is moving.

How much do AI SEO and GEO services cost — and is it worth it?

In our experience working with founders, programs typically run somewhere around $1,500 to $5,000/mo for small business engagements and $2,000 to $25,000+/mo for full programs — and GEO work tends to carry a premium over classic SEO because earning citations is hands-on. Treat these as the ranges we observe in the market, not fixed rates. Whether it is worth it depends entirely on whether the program funds the new 20% or just rebadges fundamentals.

The honest cost test: ask the vendor to itemize. If the line items are "schema implementation," "content optimization," and "technical audit" with no budget for earning third-party citations, you are buying SEO at a GEO markup. The citations — the actual driver — are the expensive, hard part, which is exactly why cheap programs skip them. Pay for the part that moves the needle, or do the fundamentals yourself and spend the budget on earned media.

One reality check on the stakes: AI Overviews now appear on a large and growing share of Google searches — trackers put it anywhere from roughly 1 in 4 to nearly half, depending on methodology — and an AI Overview correlates with sharply lower click-through for the top organic result. Ahrefs' study of AI Overviews found the top-ranking page lost roughly a third of its clicks on affected queries. Meanwhile ChatGPT appears to drive a large share of AI referral traffic, though across most publishers AI referrals remain a small fraction of total visits today. Translation: AI visibility today is about being the recommendation, not yet about referral volume. We unpack that shift in how to win in a zero-click world.

Can anyone guarantee #1 in ChatGPT or Google AI Overviews?

No. Nobody can guarantee #1 in ChatGPT or Google AI Overviews, and any vendor who promises it is selling the one thing they do not control. AI answers are generated, non-deterministic, and reranked constantly. You can dramatically raise your odds of being named and cited — that is the whole game — but a guaranteed slot does not exist.

What does reliably move visibility is well-documented. The Princeton-led GEO study (Aggarwal et al., accepted to KDD 2024) found that adding statistics, cited sources, and quotations can lift AI visibility by up to roughly 40%, with the strongest compound effect coming from a statistic paired with a source and a year. Recency matters too: recently-updated content tends to appear more often in AI answers, and a large share of AI Overview citations come from content published within the last couple of years. So the playbook is real — it is just probabilistic, not a guarantee. That refusal to overpromise is itself a trust signal, and trust is what gets you cited.

How to choose an AI SEO partner without getting sold

Choose a partner who can name the 80/20 unprompted, shows you a measurement baseline before pitching tooling, and budgets real money for third-party citations. Run the proposal through three questions: Does it separate the cheap fundamentals from the new 20%? Does it show how it will measure mentions, not just rankings? And does it refuse to guarantee a #1 slot? A "yes, yes, yes" is a partner. A "guaranteed top spot in ChatGPT" is a red flag.

That is exactly how we built our own engine — first-party data across the brands we run, plain reporting on what moves, and no promises we cannot keep. The brands that win in AI search are the ones whose facts are clean, consistent, and vouched for by sources machines already trust.

Italo Campilii

Founder, Acromatico · runs AI visibility & brand systems across 10+ live brands