Is GEO just another fad like NFTs? The differences, with data
If you have been running an agency for more than five years, you have survived several revolutions that were going to change everything: the metaverse where your clients had to buy a plot of land, the NFTs that were going to redefine brands, Clubhouse as the future of content marketing, voice search that was going to bury the keyboard. You bought into one of them, lost money or credibility over it, and learned the lesson: when every LinkedIn guru agrees that something is essential, it's time to be suspicious.
So when you hear "GEO" -Generative Engine Optimization, working on a brand's presence in the answers given by ChatGPT, Gemini, Perplexity and the rest- and you see the same swarm of courses, certifications and English-named tools, the pattern feels familiar. The question is legitimate and deserves a serious answer, not another helping of hype: is this different, or is it the metaverse in a new disguise?
Let's start by admitting you're right
The sceptic gets more right than the industry likes to admit:
Right about the pattern. The mechanics of the hype are identical: a new technology, a flood of self-proclaimed experts, tools costing hundreds of euros a month, and the threat of "falling behind". Anyone who lived through 2021 recognises the tune.
Right that there's smoke inside GEO. It exists: guarantees of "we'll get you into ChatGPT in 30 days" that nobody can deliver, because AI answers are volatile and no one controls them. "Secret" tricks that are basic SEO rebranded at a premium price. Certifications for something that's two years old. All of that is smoke, and it's worth calling it by its name.
And right that the name is marketing. "GEO" is a new label for work that is 60-70% the quality SEO we've always done, as we admit without dodging in GEO vs SEO: what's genuinely new.
With all that conceded, let's apply to GEO the same three tests that the metaverse and NFTs failed.
Test 1: do ordinary people use it, without anyone paying them to?
This is the test the metaverse never passed. Decentraland was once worth billions with a handful of real daily users; NFTs were driven by a minority speculating, not by consumers solving problems. There was never a moment when your mum used the metaverse.
With generative AI the numbers are of a different kind, and the course-sellers don't publish them:
- Frequent use of ChatGPT in Spain has gone from 4% to 28% in two years (Funcas, III Survey on AI, 2026).
- 37.9% of the Spanish population used generative AI in the last quarter of 2025, according to the INE. The INE never measured how many people entered the metaverse, because there was no need.
- 76% of Spanish SMEs use AI daily (II Hiscox Report, December 2025).
Your brother-in-law never asked you about a plot in Decentraland. He asks ChatGPT about restaurants, accountants and physiotherapists every week. That's the difference between a promise of behaviour and an actual behaviour.
Test 2: does it move your clients' money today, whether they believe in it or not?
The metaverse only affected those who chose to step in. You could ignore it completely and your bottom line wouldn't notice. The shift in search doesn't work like that: it affects your clients even if they don't know what GEO is.
- Where AI Overviews appear, the average organic CTR drops by 61% in Spain, while traffic from AI grows by 527% in a year (ismajimenez.com).
- Zero-click searches have gone from 56% to 69% in a year (data cited on stucom.com).
- Gartner forecasts 25% less traditional search before the end of 2026 (press release, February 2024).
Your clients are already losing clicks over this. The organic traffic drop you're explaining in reporting meetings -this one here- is the same phenomenon seen from the other side. With NFTs, doing nothing was a neutral option; here, doing nothing is also a decision with a cost.
Test 3: if the bubble bursts, what happens to the work done?
The decisive test. The money invested in virtual land was worth zero once the fad passed: it was work that only existed inside the bubble.
Look instead at what a serious AI visibility service involves: content that answers the real questions of the sector, structured data, corrected listings and directories, reviews worked on, brand authority in sources that third parties cite. Now imagine the most pessimistic scenario: generative AI stalls, chatbots lose steam, and by 2028 everyone goes back to Google. Which part of that work gets thrown away? None of it: it's exactly what Google has been rewarding for a decade. Well-executed GEO is a hedged bet: if AI wins, you appear in its answers; if AI loses, you're left with better SEO. The metaverse never offered anything close to that symmetry.
What this means for your agency
It's not "ride the wave now or die". It's more boring and more useful:
- Measure before you opine. Healthy scepticism is settled with your own data, not with articles (this one included). Take three clients, ask the four AIs what their audience would ask, and look at what comes out. In the study we ran with 9,865 Spanish SMEs across 30 sectors, 91% appeared in only 1 of the 4 AIs (full study): the most likely outcome is that your clients are invisible, and that's a finding that changes conversations.
- Don't buy smoke, and don't sell it. No guarantees of appearing, no "GEO tricks". Sell measurement, work on sources, and demonstrable progress.
- If you decide to offer it, do it for margin, not out of fear. The cold comparison with other services you could launch this year is in new services for your agency in 2026.
For step 1 you don't need to buy anything: we have a free AI visibility test where you enter a client's website and see in minutes which AIs they appear in. It's the cheapest way to check whether this is for you or not. And if the answer is yes, Surfeo for agencies runs that same measurement every week for your whole portfolio.
Frequently asked questions
But didn't they say the same about voice search?
Yes, and that's exactly why the right test is measured usage, not predictions. Voice search never had an INE saying that 37.9% of the population used it to make buying decisions, nor a Funcas showing a jump from 4% to 28% in two years. Predictions fail; data on usage that has already happened doesn't.
And what if ChatGPT disappears or changes radically in two years?
It might, just as Google changes its algorithm every year. That's why the criterion is test 3: only work on things that survive any scenario (content, structured data, authority, reviews) and treat whatever tool is in fashion as what it is: today's channel, not tomorrow's religion.
Isn't it too early to sell this to clients?
To promise guaranteed results, it will always be too early. To measure and work on the sources, it's the exact right moment: 91% of SMEs are invisible, which means almost no one has started and the first-mover advantage is real. By the time it matures and all your clients' competitors are working on it, the window will have closed.
How much should my agency invest to try it?
Almost nothing: the initial test is free and setting up the service on a tool costs from 55 €/month (the agency account plus the first client). Compare that to what "trying" the metaverse cost. If after three months you see no demand in your portfolio, you switch it off and you've lost less than on a team lunch.
The fastest way out of doubt isn't reading more articles: it's looking at the data. Take the free AI visibility test with your most sceptical client and decide with their result in front of you.
