TL;DR:

The phrase AI SEO agency covers two different businesses. Some agencies use AI internally to produce ordinary SEO work faster and cheaper. Others work on getting your brand named and cited inside ChatGPT, Perplexity, and Google’s AI Overviews. The deliverables barely overlap, and most vendors advertise the second while selling the first. Ahrefs studied 75,000 brands and found the strongest signals for AI visibility are off-site brand mentions, while the number of pages on a site correlates at roughly 0.194. So vet on measurement, not on volume: ask what they track, at what level, and what the number is today before you sign anything.

Somebody sends me an AI SEO agency proposal to look over about twice a month. It usually arrives after two sales calls, from someone who cannot work out why quotes for the same service range from $1,500 to $18,000 a month.

It reads like the same service because the label covers two different businesses.

Two businesses, one label

The first kind uses AI internally. Drafting, clustering, audit triage, reporting. What you get is ordinary SEO, produced faster and cheaper for them. That is a genuine efficiency and I have nothing against it. But it is a margin story, and the deliverable on your desk looks like 2019 SEO with a shorter turnaround.

The second kind works on whether AI systems mention you at all. This is where AI SEO services expand beyond traditional SEO: entity clarity, technical SEO, content architecture, and structured data all shape whether crawlers can access and understand your content before AI systems can cite it. That scope is part of ai seo and geo, sometimes framed as generative engine optimization geo, and it is different from traditional seo because traditional seo focuses on rankings in traditional search engines while this work is about ai search visibility across AI generated search experiences. A real ai search approach also accounts for third-party coverage, entity optimization, content structured so a model can lift a self-contained answer out of it, and tracking at the prompt level instead of the keyword level. In practice, that means an ai seo agency helps brands build search visibility across search engines, traditional search, and ai powered search, rather than selling generic seo services under a new label. A strong ai seo agency should be able to explain that ai discovery still depends on traditional SEO fundamentals being in place. Different work, different people, different invoice.

The confusion is profitable, which is why it persists. I looked at who ranks for the term before writing this. The first page is agency service pages, listicles named some version of best AI SEO agencies, and a Reddit thread. Almost none of it explains the distinction.

What the ones doing real work actually do differently

The honest case for treating AI visibility as its own service rests on one finding: AI SEO improves visibility in AI-generated answers because rankings and citations have come apart.

Moz ran nearly 40,000 queries through Google’s AI Mode and found that 88% of AI Mode citations do not appear in the organic top 10 for the same query. Only about one citation in ten matched an exact URL from the top ten results. AI Mode fires off several related searches behind a single question and pools sources from all of them, so being first for your head term is one input among many rather than the whole game.

Then there is the question of which signals move it. Ahrefs studied 75,000 brands across ChatGPT, AI Mode, and AI Overviews. YouTube mentions correlated most strongly at roughly 0.737. Branded web mentions landed between 0.66 and 0.71. Backlinks, in the earlier AI Overviews version of the same study, sat at 0.218. Domain Rating in ChatGPT came in at 0.266. And the number of pages on a site correlated at about 0.194, which is close to nothing at all.

Two caveats I would want if I were on the other side of this. Ahrefs sells Brand Radar, a product these findings happen to make a strong case for, so trust the direction more than the decimal places. And Ahrefs says plainly in the study itself that correlation is not causation. Brands that get mentioned everywhere tend to be brands that earned it through something other than an SEO retainer.

Even with both caveats attached, the practical read is hard to dodge. If an agency’s answer to AI visibility is thirty blog posts a month, they are pushing on the input with the weakest measured relationship to the outcome. The real work is authority building and content clarity so AI systems can reference you more reliably.

The traffic conversation to have before you sign

Ahrefs compared Search Console data across 300,000 keywords between December 2023 and December 2025 and found AI Overviews cut the average clickthrough rate for top-ranking pages by 58%, up from 34.5% eight months earlier.

I raise it because of what it does to your reporting. Good reporting for an ai seo agency should connect visibility to qualified leads and sales, not treat traffic as the only KPI, because shifts in search rankings and Google rankings now influence both referral traffic and ai referral traffic across the funnel. You can improve AI visibility and watch sessions fall in the same quarter. Companies using AI SEO report $3.7M in attributed pipeline, and AI-driven visibility can triple qualified leads within 12 months. An agency that has not warned you about that has either never run a campaign through it or is planning to show you a different chart in month four instead of explaining referral revenue, including cases where AI SEO can increase LLM referral revenue by 49x.

Five questions worth asking on the call

What exactly are you measuring, and can I see it now?

Ask for the baseline before you sign, not after. If they cannot show your current mention rate across a defined set of prompts today, they will construct a flattering starting point later.

Where are the mentions going to come from?

The answer should name publications, communities, review platforms, or video channels. If the answer is their own blog, they are describing content marketing with a newer label on it.

How much of the retainer is content volume?

Not a gotcha. There are good reasons to publish, and topical depth still matters. Automation helps most when it gives an in house team more room to focus on strategy and content quality, not when it just inflates output. But if publishing is most of the budget, weigh it against a correlation of 0.194. Productive AI work in b2b saas also includes keyword research, content optimization, technical audits, and search-intent analysis, with AI tools analyzing search data to surface high-value keywords and customer intent. Good teams still apply human review to AI-generated drafts to protect quality and originality.

What happens to my organic traffic and search visibility while this works?

You want to hear something about clickthrough compression before you hear anything about growth. It is the fastest way to tell whether someone has done this with real clients.

What are you doing that I could not do with a $99 tool?

Moz, Semrush, and Ahrefs all sell AI visibility tracking now, and the cheapest entry point is under a hundred dollars. Monitoring is not the service. The service is what gets done about what the monitoring shows. If you want the tooling side of that question first, we went through it in the 2026 SEO stack.

What it costs, and why nobody publishes a number

Nobody publishes a number because the work genuinely varies. From what crosses my desk, US agencies doing real AI visibility work land somewhere between roughly $3,000 and $15,000 a month, and most of that spread is about how much earned media sits inside the scope. Pricing also moves because search behavior keeps changing, so ai search work has to stay adaptable and data-driven in the ai era. That is an observation from proposals I see rather than survey data, so treat it as directional.

What I am more confident about is the bottom end. Anything advertised at a few hundred dollars a month is the first business model wearing the second one’s marketing.

I should say plainly that Shortlist sells this work. We run SEO and generative engine optimization services for clients, so I have a commercial interest in you concluding it is worth paying for. Read the five questions above as things I would want asked of us, and if you want the conceptual groundwork first, start with answer engine optimization before you start taking sales calls.

FAQ

An agency that either uses AI to produce SEO work more efficiently, or works to get a brand cited inside AI-generated answers, or both. The label does not distinguish between them, so the first question on any call should be which one you are buying.

They handle parts of it well. Crawl triage, clustering, first drafts, spotting anomalies in reporting. Judgment about what is worth doing remains the bottleneck, and I have not seen that shift.

It ranges widely. Efficiency-focused providers can be cheap because AI lowers their production cost. Visibility work costs more because it depends on earned coverage, which is slow and human.

Mostly the same practice under different names. Google’s published position, from May 2026, is that these are all just SEO. That is roughly right for Google Search and less right off Google, where ChatGPT and Perplexity run on their own indexes and their own retrieval rules.