There is a piece of advice that every person who launches a website hears in the first week, and it goes more or less like this: be patient, a new domain takes months before Google will trust it, so plant your content now and wait. It is reasonable advice. For twenty years it was also, without doubt, correct.
Then something changed, and we have now watched it happen on a domain that is not yet one month old.
The numbers
Kurli is an AI hairstyle advisor. You upload one selfie, the face is mapped in the browser to work out its shape and proportions, and the styles that genuinely suit that face are rendered onto your own photo before you commit to the chair.
The domain, kurli.me, was registered on 27 July 2026. That is not the client telling us so, it is what the public domain registry records, and anyone reading this can check it in a few seconds with a WHOIS lookup. The registrar is IONOS SE.
Here is where the visitors came from in the first 25 days of that domain's existence:
| Referrer | Visits | | --- | --- | | accounts.google.com | 69 | | chatgpt.com | 56 | | kurli.me | 38 | | www.google.com | 23 | | com.google.android.googlequicksearchbox | 5 |
Now, two of those five lines are not new visitors at all, and it is worth being honest about that before anyone gets excited. The first, accounts.google.com, is the Google sign-in flow, which means people coming back into the product after logging in. The third, kurli.me, is the site referring to itself, which is what analytics records when someone moves between pages. Neither one is a person discovering the product for the first time.
Strip those two away and what remains is the part that matters:
ChatGPT sent 56 visitors. Google Search sent 23. On a domain that was less than one month old, the AI assistant was the largest single front door, by more than two to one.
Why this is not a small thing
Let us be clear about the scale first, because we have no interest in dressing up a modest number as something it is not. Fifty-six referrals is not a flood. Nobody is retiring on this traffic.
But the size is not the point. The point is the ratio, and the age of the domain.
A brand new domain has nothing that classical SEO runs on. It has no backlinks, because nobody has had time to link to it. It has no history, no accumulated trust, no track record of satisfying searchers. Google is, quite properly, cautious about such a site, and that caution is exactly what produces the familiar advice about waiting several months. Twenty-three visits from Google in 25 days is not Google failing. That is Google behaving normally toward a stranger.
ChatGPT did not behave that way. It recommended the site to people while the domain was still measured in days.
This is the practical difference between the two systems, and it must be said plainly. Google ranks pages largely on accumulated authority, which is a thing that by definition takes time to accumulate. A language model with browsing recommends whatever best answers the question in front of it right now. It has no equivalent instinct to distrust a domain simply for being young. If your page is the clearest answer to "what haircut suits a heart shaped face", the age of your certificate does not enter into it.
For anyone launching something new, that difference is enormous. The channel that traditionally takes six months to open was outperformed, in the first 25 days, by the channel that everyone is still treating as experimental.
What this means if you have just launched something
We see a particular version of this problem constantly at the moment. Somebody builds a genuinely good application, often very quickly and with a great deal of AI assistance, puts it live, shares it in a few places, receives some kind comments, and then watches nothing happen. The building was the easy part. Getting one person who is not the founder to use the thing turns out to be the wall.
The instinct in that moment is to wait for Google. Alla fine, that is what everybody has always done.
What Kurli suggests is that waiting for Google may now be the wrong first move for a new product, or at least not the only move. The faster door, in the first month, was the AI assistant. And the things that get you through that door are not mysterious:
- Be readable. An assistant cannot recommend a page whose content only appears after JavaScript runs. If the answer is not in the HTML, for these systems it does not exist.
- Answer real questions directly. Kurli does not have one page that says "AI hairstyle tool". It has guides organised by face shape, by hair type, by length, by hair colour, for men and for women. Those are the shapes of the questions people actually ask, and an assistant looking for a specific answer finds a specific page.
- Say what you are, in plain words. Including the limits. Kurli's own about page states openly that it is a styling guidance tool and not a substitute for a stylist who can see your hair in person. Honesty of that kind reads as trustworthy to a human and gives a model something unambiguous to quote.
- Do not make people sign up to see the value. A wall in front of the product is also a wall in front of the crawler.
None of this requires authority you have not earned yet. It requires being clear, and being machine readable, from day one. That is the whole opportunity, and it closes as this becomes common knowledge.
The honest caveats
We would rather state these ourselves than have you find them.
This is one site, over 25 days, in a consumer category where people ask assistants very natural questions ("what haircut suits my face"). A business to business product in a dull category may see a completely different mix. The absolute numbers are small, so ordinary variation moves them around. And we are reading referral data, which captures the people who clicked through to the site, not the much larger number who may have seen the brand mentioned inside an answer and never clicked at all.
What we are not claiming is that Google no longer matters. Of course it matters, and over the coming year Google will very likely become the larger channel for this site as the domain matures.
What we are claiming is narrower and, we think, more useful: on a domain 26 days old, ChatGPT was the bigger front door. If you are launching something now, that is worth building for deliberately, rather than discovering by accident.
See where you stand
If you have built something and nobody is finding it, the first question is not what to fix, it is whether the AI systems can read your site at all. That takes a couple of minutes to check.
Run the free Get Found check and see what Google and the assistants actually see when they look at your site.