One Agency or Two? The Real SEO-vs-AEO Tradeoff

Search "best AI SEO agency" today and the first page reads like a hall of mirrors. "11 Best AI SEO Agencies (GEO & AEO) in 2026." "10 Best Agencies Combining Paid, SEO, and AEO." "8 Top AEO Agencies for Lead Generation." Click into a few and the pattern gets funnier: several of the roundups are published by agencies that also happen to appear, conveniently, near the top of their own list. Nobody fact-checks a "best of" post written by an interested party, and in 2026 that's most of what a business owner researching this decision actually finds.

Underneath the noise is a real, unglamorous decision every business with a website now has to make. Do you hire one agency that claims to handle "both SEO and AEO," split the work between a traditional SEO shop and an AI-visibility specialist, or try to build the capability in-house? None of the three answers is obviously wrong, and none of them is obviously right either — which is exactly why the listicle industry has rushed in to fill the gap with confident-sounding nonsense. The honest version of this question is more interesting, and considerably less tidy, than any of those posts let on.

Why This Decision Actually Matters Right Now

Start with a prediction that didn't pan out the way it was framed. In February 2024, Gartner VP analyst Alan Antin predicted that traditional search engine volume would drop 25% by 2026 as AI chatbots and virtual agents took over. By April 2026, that hadn't happened — Google still commands over 90% of the search market, and traditional search volume never fell anywhere near that mark. The honest read isn't that Gartner was wrong to flag a shift; it's that the shift is real but slower, messier, and more channel-specific than a single headline number suggested. That gap between the dramatic prediction and the duller reality is precisely the environment agencies are selling into right now, and it cuts both ways — some genuinely understand the nuance, and some are happy to let the scarier number do their selling for them.

The actual data, where it exists, tells a more useful story than either the doom or the dismissal. Semrush's 2026 AI Visibility Index, built on 126 million U.S. AI search prompts collected between January and April 2026, found that 45% of marketing leaders say they cannot accurately measure their brand's visibility in AI-generated answers, and only 9% have tools that track all the relevant metrics across platforms. That's not a small measurement gap — it means most of the businesses currently paying someone for "AI visibility" have no independent way to check whether the work is doing anything at all. The same report found something sharper still: organizations running integrated SEO and AI-visibility strategies together reported increased traffic or leads 81% of the time, versus just 36% for organizations managing the two disciplines separately. That single number is close to the actual crux of the "one agency or two" question, and it's worth sitting with rather than treating as a marketing agency's talking point, because it comes from Semrush's own data rather than from anyone selling combined retainers.

Meanwhile Ahrefs' Q1 2026 AI Search Benchmark report, based on 75,000 brands, is a useful reality check against the more breathless framing: Google still sends roughly 190 times more traffic to websites than ChatGPT does, and ChatGPT's search volume sits at around 12% of Google's. AI Overviews, meanwhile, now appear on 21% of all keywords across 146 million SERPs Ahrefs tracked, and reduce clicks to the top-ranking organic result by 58% when they show up. Put together, that's the actual shape of the problem: traditional search still carries most of the volume, AI answers are eating an increasing and disproportionately important slice of it, and almost half of marketing leaders can't reliably tell which of their efforts is moving either number. That's the gap an agency engagement is supposed to close — and it's also exactly the gap a mediocre one can hide inside.

What a Real AI SEO Agency Engagement Should Actually Include

Set the marketing copy aside and ask what the deliverables should concretely look like. First, a real baseline — not a vague "visibility score" but two separate, honestly measured starting points: where your organic rankings actually sit today, and what ChatGPT, Google's AI Overviews, Perplexity, and Gemini actually say when someone asks about your business or category right now. If an agency can't show you that second baseline before you sign anything, they don't have the measurement infrastructure to prove the engagement worked later either.

Second, entity consistency work — making sure your business's name, description, credentials, and relationships are stated the same way across your own site, your Google Business Profile, review platforms, industry directories, and anywhere else an AI system might cross-reference to decide it trusts a claim about you. This is unglamorous, structural work, and it's exactly the kind of thing that gets skipped when an agency is really just running the same SEO playbook under a new name.

Third, real structured data — Schema.org markup that gives a model something unambiguous to parse instead of a paragraph it has to interpret, plus the technical groundwork Search Engine Land's technical guide to AI-agent search lays out clearly: managing which AI crawlers can access your site, separating core facts from boilerplate so answer engines can extract them cleanly, and keeping "last updated" signals current since retrieval-augmented systems penalize staleness more visibly than traditional search ever did.

Fourth — and this is where the relabeling problem shows up most often — content that's actually built to be cited, not existing SEO content with an AI-sounding paragraph bolted onto the top. Citation-worthy content answers a specific question precisely, states facts and numbers a model can lift cleanly, and doesn't bury the useful part under three paragraphs of scene-setting written to satisfy an old keyword-density habit. This is the substance of what our own AI SEO Agency engagement is built around — coordinating both disciplines from two honest, separately tracked baselines rather than blending them into one number that hides which one is actually working.

The Real Debate: Converging Disciplines, or Two Different Jobs?

This is where the genuine disagreement lives, and it's worth reading the actual argument rather than picking a side by instinct. Digiday's reporting on the GEO-versus-SEO debate, published in March 2026, quotes Jeremy Moser, co-founder and CEO of uSERP, putting it about as bluntly as an industry practitioner is willing to: "If a GEO service does not openly tell you that success in AI visibility is 80 percent good fundamental SEO, they are selling you snake oil." Lily Ray, VP of SEO Strategy and Research at Amsive, made a related but gentler point in the same piece — that this pattern echoes past cycles like AMP and featured snippets, each initially marketed as a distinct discipline requiring separate expertise before settling back into "just good SEO, applied to a new surface."

But the same article is careful not to flatten the argument into "it's all the same thing," because Michael King, founder and CEO of iPullRank, made the case for genuine novelty: retrieval-augmented generation and the architecture behind how large language models actually select and rank sources is real, additional technical work that most traditional SEO teams simply haven't built. That's not a small caveat — it means the honest answer to "is AEO really different from SEO" is neither "completely" nor "not at all." It's roughly 80% shared foundation and 20% genuinely new mechanics, and an agency that can't tell you which parts of your engagement fall into each bucket probably hasn't thought about it carefully enough to charge you for it.

You can watch a rougher version of the same disagreement play out among engineers rather than marketers. On a Hacker News thread discussing how generative engine optimization "rewrites the rules of search", commenter bravesoul2 dismissed the framing entirely — "If SEO was literally Search Engine Optimisation it'd be fine. They really mean SEG (search engine gaming)" — while kkaatii predicted GEO would follow the same decline: "Not too different from how SEO ends. Just spammy content written by bots." A more measured comment from user edwin, citing independent testing, noted that AI answer sets are genuinely unstable day to day, with Google's AI Mode and ChatGPT recommending the same top result only around 47% of the time on identical queries — a real, current data point that makes any agency promising a fixed "AI ranking" sound less credible on its face.

What People Are Actually Arguing About: Is This Worth Paying For At All

The sharpest, most honest version of this debate isn't happening in a vendor's sales deck — it's underneath a Y Combinator launch post. In early 2026, a startup called Sitefire launched on Hacker News with the pitch "Automating actions to improve AI visibility," essentially productizing a chunk of what an AI-visibility agency retainer does by hand, and the comment thread turned into exactly the argument a business owner evaluating agencies should want to read.

"Ugh. The worst of SEO, but a bunch more of it? Noooooo."

— ceejayoz, on Hacker News

That's the skeptic's whole case in one line: automate the wrong incentives and you don't get better visibility, you get more of what already made SEO annoying, just faster and cheaper to produce. It's a fair worry, and it applies just as well to a human agency team running the same playbook at scale as it does to a piece of software. The founder, replying as vincko, framed the product's actual differentiation more narrowly than the pitch implied — positioning it against existing players like Profound and Airops by arguing "our goal is to be more hands-off, so you don't need a team," which is really a build-versus-buy argument dressed up as a product feature. And commenter XCSme pushed on the part most of these debates skip entirely: at roughly $249 a month, "is [that] something most small brands/shops can afford? Many have a few $ks in total revenue." That question — can the business this is being sold to actually justify the spend — belongs in every conversation about hiring an agency too, and it rarely gets asked in the sales call.

How to Actually Evaluate One Before You Sign

Given all of that, the evaluation questions write themselves. Ask to see real, current AI answers about your business — not a generic case study from another client, but what ChatGPT, Perplexity, and Google's AI Overviews actually say right now when someone asks the questions your customers ask. If the agency can't produce that in the pitch meeting, they don't have the tooling to prove anything later, and given that Semrush found only 9% of marketing leaders have tools covering all the relevant platforms, that gap is common enough that you shouldn't assume it away.

Ask specifically what changes between month one and month three, in concrete terms — which schema gets implemented, which entity inconsistencies get fixed, which pages get rewritten for citation rather than keyword density, and what the measurement looks like at each checkpoint. A vague answer here is the clearest tell of a retainer that's really the old SEO deliverable list with "AEO" pasted over the top of it.

Ask them to name, unprompted, which parts of the work are shared SEO fundamentals and which parts are genuinely new — the Moser-versus-King framing from the Digiday piece is a good filter here. An agency that answers "basically everything is new" is probably overselling; one that answers "basically nothing is different" is probably underselling and skipping the entity and structured-data work that actually matters for citation. The honest answer sits in between, and a team that's actually done the work can usually tell you where.

Who Each Path Actually Fits

A single combined agency engagement tends to fit a business without internal marketing capacity to manage two vendors and translate between them when priorities conflict — which, per Semrush's numbers, is the option correlated with the strongest reported results, at 81% versus 36%. Splitting the work between a specialist SEO shop and a dedicated AI-visibility team can make sense for a larger business that already has strong organic performance and wants deep, focused expertise on the newer discipline specifically — but only if someone internally is actually coordinating the two, since the Digiday reporting and the entity-consistency logic above both point to the same risk: an SEO tactic optimized in isolation can measurably hurt AI citation, and nobody catches that if the two teams never compare notes. Building it in-house is realistic only for a business with genuine technical and content capacity already on staff, since the baseline measurement work alone — tracking citation across four or more AI platforms honestly — is closer to a standing function than a one-time project.

None of these paths is free of the core risk this whole piece has been circling: paying for a retainer that's mostly relabeled SEO, sold with the newer, scarier vocabulary. The way to protect against that isn't picking the "right" structure — it's asking for the baseline, the month-by-month specifics, and the honest breakdown of what's actually new, before any contract gets signed. If you want a starting point for that first conversation, our own AEO Score tool gives you a real, evidenced read on where your structured data and entity signals stand today — the same kind of baseline any agency worth hiring should be able to produce for you in the first meeting, not the third invoice.