GEO vs SEO: what's genuinely new and what's just rebranded
You're about to read something almost no one selling GEO will tell you: between 60 and 70% of GEO work is the quality SEO we've always done. Content that answers real questions, a technically healthy site, authority built over time. If your agency does good SEO, it already does two thirds of GEO without knowing it.
So is it all a marketing label? Not that either. The other 30-40% is genuinely new, and it's precisely the part where a classic SEO agency gets it wrong if it doesn't understand it. This article separates the two parts without selling smoke, so you can decide what you really need to learn and what you already know under another name.
First, a one-line definition: GEO (Generative Engine Optimization) means working on a brand's presence in the answers given by AI search engines -ChatGPT, Gemini, Perplexity, Claude- just as SEO works on its presence in Google. If you need to explain it to a client without jargon, don't send them this article: send them the simple glossary version, written for SMEs.
The part that's the same (and why it pays to admit it)
Content that answers specific questions. AIs build their answers from content that properly addresses a specific question. That's exactly what good SEO has been doing for years, ever since semantic search arrived. If your articles already answer "how much does X cost" or "how to choose Y", you're set.
Technical health. A site that Google's crawlers read well is also read well by the AI crawlers. Clear architecture, speed, content accessible without JavaScript flourishes that hide the text.
Authority and trust. AIs prefer to cite sources with reliability signals: brands mentioned in the media, with a consistent presence, with identifiable authors. The functional equivalent of the same old E-E-A-T.
Why labour the point? For two practical reasons. First: it saves you buying training for things you already master. Second: it dismantles the seller of "secret GEO"; anyone charging you a premium to teach you that "AI rewards useful content" is reselling you your own trade. We cover that side of the hype at length in is GEO just another fad like NFTs?.
What genuinely changes: four operational differences
1. From keywords to conversations
In SEO you optimise for "employment lawyer madrid". Nobody types that to an AI: they type "I've been let go while on sick leave, which lawyer would you recommend in Madrid?". The unit of work stops being the keyword and becomes the prompt: the full question, with context, the way people actually speak. This changes keyword research at its root: you no longer mine search volumes in a tool, because there are no public prompt volumes; you build the question map from your knowledge of the client's business and validate it by testing.
2. From positions to mentions (and measurement turns statistical)
In Google there are ten positions and yours is a number. In an AI answer, either it mentions you or you don't exist, and the same question can give different answers on Monday and Thursday. Measurement stops being deterministic: you measure coverage (in how many of the key questions you appear), frequency (what percentage of the time) and sentiment (what it says about you), over samples repeated across time. This has a direct commercial consequence: there are KPIs you can commit to in a proposal and KPIs you can only report, and confusing the two is expensive — here's the list of which is which.
3. Offsite matters far more: reviews, forums, directories
This is the difference that throws SEO agencies most. In Google, your site is the main asset and links are the support. AIs, by contrast, draw heavily on third-party sources: reviews, Reddit and forum threads, sector directories, other people's comparisons. You can have the best site in the sector and still be invisible because the conversation about your category happens in places where you're not present. The new work: auditing which sources the AIs cite in your sector and getting a presence there, a discipline closer to PR than to on-page SEO.
4. There isn't one engine: there are four, and they don't agree
Ranking in Spain was synonymous with ranking in Google. Now ChatGPT, Gemini, Perplexity and Claude use different sources and criteria, and the results reflect it: in the study we ran across 9,865 Spanish SMEs in 30 sectors and 10 cities, 91% appeared in only 1 of the 4 AIs (full study). Being strong in one and gone from the other three is the norm. In practice this means measuring and working each engine separately, not assuming "AI" is a single homogeneous thing.
Add to this structured data with a different purpose: in SEO, schema served rich snippets; in GEO it serves to let the machine extract your data without error -prices, hours, services- and cite you with correct information instead of inventing it.
What this means for how your agency operates
- You don't need a new department. You need the SEO team to add three habits: thinking in prompts as well as keywords, treating offsite (reviews, forums) as your own territory, and measuring across 4 engines.
- You do need new measurement. The rank tracker doesn't know what ChatGPT says about your client. Checking dozens of prompts across 4 AIs every week by hand doesn't scale past the first client; that's what we use Surfeo for, which does it automatically for every client in the portfolio and leaves it in a PDF report for your reporting.
- It does change the sales pitch. Selling "positions" no longer captures the value; selling "presence where your client decides" does. The step-by-step of the first audit to launch the service is in the first AI visibility audit, step by step.
Frequently asked questions
If 60-70% is SEO, can I sell GEO without knowing SEO?
Bad plan. GEO without an SEO foundation is a façade with no building: you can fix listings and reviews, but if the content and the technical side are shaky, the AIs will have nothing solid to cite. The natural path is an SEO agency that adds GEO, not the other way round.
Should I bill it as a separate service or inside the SEO?
As a separate module within the same retainer: its own deliverables (prompt coverage, mentions by AI, corrected sources) and its own price. If you give it away inside the SEO "because it's almost the same", you've decided that 30-40% of new work is worth zero. Real price ranges in how much to charge for AI visibility services.
Does GEO cannibalise my SEO service?
Quite the opposite: it props it up. The real threat to your SEO retainer is the drop in clicks (organic CTR falls by an average of 61% where there are AI Overviews in Spain, according to ismajimenez.com) without a story to explain it. GEO gives you the story and the service that goes with it.
What do I tell a client who asks the difference?
The short version: "SEO works to get you found when people search on Google; GEO works to get you recommended when they ask an AI. The foundation is the same; the shop window is different." And so they can read it at their leisure, pass them the glossary explanation, which is written for SMEs, jargon-free.
Want to see the difference with a real case rather than in theory? Run the free AI visibility test on a client you have in Google's top 3: checking whether their SEO translates into presence in the AIs is the fastest way to understand which part of the work you're missing.
