Let me tell you about a website that, by every conventional measure, had nothing going for it.
A twelve year old built it. They were interested in hair styling, they had an idea for a small app, and they built it the way a great many people build things now: by describing what they wanted to an AI and letting it write the code. Vibe coding, as everybody calls it these days. And the app worked. It was decent. It did what it promised to do.
And nobody came.
Zero traffic. Literally zero. Over the entire life of the app, subscriptions brought in a grand total of thirteen dollars and ninety five cents. One might reasonably conclude, at that point, that the whole thing was a waste of a summer.
We took the site on and pointed our GEO and SEO agents at it for thirty days. What happened next is the reason I am writing this.
What the dashboards show
Today the site is receiving around 400 clicks a day arriving from ChatGPT. This comes from the analytics installed on the site, so it is measured traffic rather than a projection.
Google, meanwhile, has still not put the site on page one for anything meaningful. Search Console shows the site has begun to register, but the visitors are not coming from there.
So here we have a website getting real, daily, high intent traffic from an audience that Google has not yet agreed to send.
The test that matters
Numbers on a dashboard are one thing. What I wanted to know was simpler: if a stranger asks ChatGPT about this topic, does the site actually come up?
So I asked, in an incognito window, with nobody logged in. This detail is important. If you ask ChatGPT a question while logged in to your own account, your history can shape the answer, and you end up admiring your own reflection. In a clean session, with no account and no history, the assistant recommended the site in second place.
That is a stranger's view. That is the real one.
Why this should surprise you
Consider what this site does not have.
It went live on July 27, 2026, so at the time of the test it had roughly one month of existence behind it. It has no backlinks. It has no Google Business Profile. It has no brand that anybody has heard of, no press coverage, no years of accumulated authority.
By the standard SEO playbook, this site has not yet earned the right to be recommended by anyone. Domain age, link profile, established trust: these are the things we have all been taught you must have first, and it has none of them.
ChatGPT recommends it anyway.
The belief I would like to question
There is an idea repeated constantly in our industry, and it goes something like this: GEO is simply SEO with a new name, so win at SEO first and AI visibility will follow on its own.
This case does not support that. The two surfaces reached completely different conclusions about the same website, in the same month, from the same content. Google is still deciding. ChatGPT already decided.
The honest explanation is that these systems weigh a site's past differently. Google has always been slow and careful with a young domain that nobody links to, and there are good reasons for that caution. An assistant answering a specific question has a different job: it needs a source that answers that particular question well, right now. A site with one month of history can satisfy that requirement. It cannot satisfy Google's, at least not yet.
What this means if you have vibe coded something
The truth is that most people who build an app this way quietly give up on traffic. They assume that without a marketing budget, without an agency, without years of history, they will never be found, and so they never try. That assumption deserves to be challenged.
If you are in that position, the practical first step is to find out what AI assistants currently say about you. Ask them, in a clean session, the questions your customers would ask. If your site does not appear, you now know where the work is. That answer is available to you today, and it does not require you to wait eighteen months for Google to make up its mind.
The limits of one case
It must be said clearly: this is one website. One case does not become a law of nature because it produced a pleasing result, and I would be doing you a disservice by pretending otherwise. A different topic, a more crowded field, a subject where established brands dominate the conversation, any of these could produce a very different outcome.
What one case can do is disprove a rule stated as absolute. The rule was that GEO requires SEO first. Here is a site with no SEO position to speak of, being recommended by ChatGPT. Whatever the general pattern turns out to be, that particular rule cannot be true as stated.
I am genuinely open to being argued with on this, and I would like to hear from anyone seeing the opposite in their own work.
This piece first appeared on our Medium, where you can see the dashboard screenshots.