# The Difference Between Being Built and Being Found

Category: Findability
Published: 2026-07-08
Author: Jean Dorff
Publisher: Jean Dorff Consultancy

> An AI website builder can produce a professional site in an afternoon. What it cannot produce is a site that AI search engines will recommend.

There has never been a cheaper moment to own a good-looking website. An AI builder will take a paragraph about your business and return a responsive, competently typeset site before lunch. The layouts are clean. The copy is serviceable. The photography is tasteful in the way that stock photography is tasteful. For the first time in the history of the web, looking professional online is close to free.

Which is precisely why looking professional online no longer distinguishes anyone. The scarce thing has moved. It is no longer the container — it is whether the systems that now mediate discovery can read what is inside it, understand who wrote it, and decide it is worth quoting when someone asks a question.

## Two different problems

Building a website is a design and production problem. It has a visible output, an obvious finish line, and a client who can judge the result by looking at it. Being found is an architecture problem. It has no visible output at all. Nobody browsing your site will ever notice your schema markup, your heading hierarchy, your canonical URLs, or the plain-language summary sitting at /llms.txt. The only audiences for that work are crawlers and models — and they are the audiences currently deciding whether your name appears in an answer.

Solving the first problem does not solve the second. This is not a criticism of AI builders; it is a description of what they were designed to do. They optimize for the thing the buyer can see. Findability is the thing the buyer cannot see and therefore cannot ask for.

## Layers, not alternatives

The discipline is often presented as a sequence of fashions — SEO gave way to AEO, AEO is giving way to GEO — as though each new acronym retires the last. That framing is wrong and it is expensive. They are layers. Each one sits on the one beneath it, and skipping a layer weakens everything above it.

SEO remains the foundation: clean metadata, semantic structure, fast loading, mobile-first layout, a heading hierarchy that reflects meaning rather than font size, internal links that describe where they go, alt text that says something. This is not legacy work. AI crawlers read the same HTML that search crawlers read, and they are no more forgiving of a page built out of unlabeled divs.

AEO is the layer that makes a business citable rather than merely present. It depends on structured data that states plainly what the organization is and what it offers, on entity information that stays consistent across every surface that mentions you, and on content organized so that a model can extract a defensible answer rather than a vague impression. When someone asks an assistant who does findability consulting in Dallas, AEO is what determines whether a name surfaces at all.

GEO goes one step further, into how generative systems select, weight, and synthesize what they have found. It is the difference between being findable and being quotable — structuring an argument so that when a model does reach for you, it reaches for a sentence that survives being lifted out of its paragraph and still says what you meant.

EEAT — Experience, Expertise, Authoritativeness, Trustworthiness — runs through all three. It is not a layer so much as a constraint on the content itself, and it is the layer most often faked and most rarely earned.

> None of this is visible to the person browsing your site. All of it is visible to the systems deciding whether to recommend you.

## Where businesses actually stall

The common failure is not laziness. It is a reasonable belief that the site is finished because it looks finished. A business launches, the owner is pleased, traffic stays flat, and the conclusion drawn is that the site needs more marketing — more posts, more ads, more presence on whichever platform is currently loud. What it needed was structure, which no amount of downstream promotion can retrofit.

The diagnostic question is simple enough to ask yourself this afternoon. Open an AI assistant and ask it about your field of expertise in your city. Not your name — your category. If you do not appear, the problem is almost never that your site is ugly. It is that nothing on it told the model what you are, and nothing off it confirmed the claim.

## The order of operations

The sequence that works is unglamorous. Establish what you want engines to say about you before writing a line of copy. Write the site before designing it. Build the structural layer — schema, semantic HTML, sitemap, robots configuration, llms.txt, markdown mirrors — as part of the build rather than as a later optimization pass. Then verify by fetching each configuration directly, because a summary that says the crawler file is fine is not evidence that the crawler file is fine.

A website built this way costs more attention than one generated in an afternoon. It also does something the generated one cannot: it participates in the conversations where buying decisions are increasingly made, rather than waiting politely at a URL nobody has been given a reason to visit.