The published advice contradicts itself. One agency argues that asking an SEO firm for AEO is like asking a plumber to do electrical work. Another argues that if your SEO is solid, your GEO is largely handled already. A third says splitting the work between two vendors duplicates production and breaks measurement.
Each of those is written by someone selling the conclusion they reached. The buyer gets three incompatible answers and no way to adjudicate between them.
There is a way, and it comes down to a single question about your current agency.
Key Takeaways
- The deciding variable is earned media: whether your agency already places content on third-party sites.
- Most SEO agencies do not: their model is built around on-site production.
- Asking anyway produces schema and FAQs billed as GEO: real work, wrong half of the problem.
- Two vendors creates a specific failure: conflicting positioning and a broken measurement loop.
- Test it in one question: ask what third-party placements they secured in the last six months.
The Question That Decides It
Ask your SEO agency what third-party placements they have secured in the last six months. Not backlinks acquired. Editorial coverage, review platform positioning, and content published somewhere other than your domain.
If they can name several, they have the capability that matters and extending them into GEO is reasonable. If the honest answer is none, they are not equipped for the part of the work that determines the outcome, and no amount of willingness changes that.
This is not a criticism of SEO agencies. Their model is built around on-site production and technical work, which is where SEO outcomes are largely determined. Earned media is a different discipline with different relationships and a different delivery rhythm.
Why That Variable and Not Another
Analysis of 25 million cited links found earned media accounts for the large majority of AI citations, while paid and advertorial placements account for a fraction of a percent. An agency working only on your domain is working on the smaller input.
Everything else your SEO agency does transfers usefully. Technical hygiene, schema, content structure, and crawlability all matter for AI retrieval, and an agency doing them well is doing something that counts.
What does not transfer is the corroboration layer. Models weight independent sources describing you more heavily than your own account of yourself, particularly in categories where a wrong answer carries consequence. That work involves editors, review platforms, and communities, and it cannot be produced from a content calendar.
So the question is not whether your SEO agency is good. It is whether their scope reaches the input that decides citations, which is the same test that separates third-party mentions and authority from on-site optimisation generally.
Three Options, Honestly Compared
| Extend current agency | Add a second vendor | Consolidate elsewhere | |
|---|---|---|---|
| Works when | They already do earned media | Mature SEO, want a specialist layer | SEO is on-site only |
| Main risk | Schema relabelled as GEO | Conflicting positioning | Transition cost |
| Cost | Modest increase | Two retainers | One retainer |
| Measurement | Unified | Split, often incoherent | Unified |
| Speed to start | Immediate | Fast | 4-6 weeks |
The two-vendor option gets recommended less often than it deserves and fails more often than it should. It works when your SEO programme is genuinely mature and you want specialist citation work layered on top. It fails when the two vendors describe your company differently, because models synthesise whatever they retrieve and inconsistency is itself a cause of poor performance.
If You Extend Your Current Agency
Two things prevent the common outcome, which is paying more for the same work under a new name.
Specify the earned component explicitly. Not "GEO services" but a stated number of third-party placements or corrections per quarter, with named target publications or platforms. Vague scope here reliably resolves into more content.
Require source-level reporting. Ask which specific sources produce your current answers, and expect that list to change over the engagement. An agency that cannot produce it is looking at your site rather than at outputs, which is the practical difference between GEO and SEO with a new label.
If both requests are met with enthusiasm and a clear method, extending is the efficient choice. If they produce a deck about schema and content structure, you have your answer.
The Data Disagreement Worth Knowing About
One reason the advice conflicts is that the underlying data does too, and the disagreement is large enough to matter.
Some agencies argue AI citations come overwhelmingly from organic top ten results, which supports the conclusion that strong SEO is sufficient. Independent studies point the other way: Moz's analysis of nearly 40,000 queries found 88% of Google AI Mode citations do not appear in the organic top ten, and Ahrefs' analysis of 1.9 million citations converged on roughly 12% overlap.
Two large independent studies agreeing against a vendor claim is worth weighting accordingly. The practical reading is that ranking helps you be discovered and does not determine whether you are used, which is why the two disciplines overlap without being interchangeable.
What Happens If You Do Nothing
The default outcome of an unresolved decision is that your existing agency quietly adds schema markup and FAQ sections, reports on them as AI optimisation, and nothing changes in what models say about you.
This is the most common state in the market right now, and it is comfortable because it produces reporting. Nobody is being dishonest. The work is real, it is just addressing the accessible half of a problem whose outcome is determined elsewhere.
You find out by measuring rather than by asking. Run your category questions, record which sources produce the answers, and note whether any third-party source has changed over the past two quarters. That is what visibility tracking should surface, and if the source mix has not moved, neither has the thing that matters.
A Short Diagnostic
| Ask your SEO agency | Good answer |
|---|---|
| Third-party placements last six months? | Several, named |
| Which sources produce our AI answers? | Specific pages and threads |
| What is our baseline and method? | Prompts, platforms, dates, frequency |
| What is in scope off our domain? | A concrete list |
| When would you advise against this? | Named conditions |
An agency answering four or five of these can extend into GEO credibly. One answering fewer than two is being asked for something outside their model, and the resulting engagement will reflect that regardless of intent.
Getting to a Decision
Most of the difficulty here is that the question gets asked in the abstract, when it is answerable with evidence about your specific situation.
What models currently say about you, which sources produce those answers, and whether any of them have changed recently tells you whether your existing programme is reaching the input that matters. It also tells you what a second vendor would actually be brought in to do, which makes that conversation concrete rather than a comparison of pitches.
That read is what Lureon starts with, and it is equally useful whether the conclusion is to extend your current agency, add a specialist, or consolidate.
FAQs
1. Can my SEO agency just do GEO as well?
It depends on whether they already do earned media. Agencies securing editorial coverage and third-party placements have the capability that determines citations. Those working only on your own domain do not, and extending them typically produces schema and content work billed as GEO.
2. How do I test whether my agency is capable?
Ask what third-party placements they secured in the last six months, and which specific sources produce your current AI answers. Both questions are answerable in minutes by an agency doing the work and unanswerable by one that is not.
3. Is it a mistake to hire a separate GEO agency?
Not inherently, but it has a specific failure mode. Two vendors describing your company differently creates inconsistent signals, and models synthesise whatever they retrieve. It works when your SEO programme is mature and positioning is centrally owned; it fails when neither is true.
4. Does strong SEO mean my GEO is already handled?
Partly. SEO makes your content discoverable, which is a prerequisite. But Moz found 88% of AI Mode citations do not appear in the organic top ten, and Ahrefs' larger analysis converged on similar figures, so ranking does not determine whether you are cited.
5. What should I ask for if I extend my current agency?
A stated number of third-party placements or source corrections per quarter with named targets, plus source-level reporting showing which specific pages and threads produce your answers. Vague GEO scope reliably resolves into additional content production.
Figures reference Muck Rack's analysis of AI-cited links, Moz's analysis of Google AI Mode citations, and Ahrefs' citation research. Published vendor claims on this topic vary considerably and frequently align with the service being sold.