Two chiropractic practices in Cleveland. Same modalities. Similar years in business. Both have websites that look fine.
Someone in the city wakes up at 6am with pain shooting down the back of their leg. They open ChatGPT and describe it. They ask what it might be, what they should do about it, and eventually who they should see in Cleveland.
By the time they ask that last question, the model has already pulled from a handful of practice websites to build its answer.
One practice gets named. The other does not.
The difference is not budget. It is not tenure. It is not which one has the nicer logo.
One site says something specific. The other says nothing a model can use.
What Generic Means To A Machine
When we say a website is generic, we usually mean it is boring. That is a human judgment about tone.
The machine does not care about tone. It cares about signal.
To an AI model deciding who to cite, generic means indistinct. It means there is nothing on the page that ties this specific practice to this specific question in a way worth quoting.
"Serving the Cleveland community since 1998" reads warm to a person. To a model it is noise. Every practice serves its community. Every practice has a year it started. None of that helps the AI decide that you, specifically, are the answer to a question about a specific condition.
The model is scanning for what makes you citable. Sameness reads as invisibility.
If your homepage could belong to any practice in any city, the AI treats it as nothing.
The Patterns That Just Died
Most chiropractic websites are built from the same handful of templates. You have seen them. You might be running one.
Here is what stopped working:
The mission-statement homepage. "We are passionate about helping you live pain-free. Our caring team treats every patient like family." This says nothing. It describes a feeling, not a practice. The AI has nothing to extract.
The stock-photo condition page. A photo of someone holding their lower back, two sentences about how back pain is common, and a button that says Book Now. No depth. Nothing the model can quote as evidence that you actually understand the condition.
The conditions-we-treat link wall. Twenty conditions listed, each linking to a thin page with one paragraph. Wide and shallow. The AI reads shallow as unqualified.
The interchangeable service page. Your adjustment page could be copied onto any practice site in North America and nobody would notice. If it is interchangeable, it is invisible.
The AI-written filler. This one is newer and it is everywhere. Pages generated to "have content," full of sentences that are grammatically perfect and informationally empty. Models are getting very good at recognizing their own exhaust. It does not get cited.
Why This Got Worse, Not Just Different
Here is the part most chiropractors miss.
Generic content used to still work a little. Under old search, a thin page could rank on volume. Stack enough keywords, build enough backlinks, and Google would list you even if the page said nothing memorable. It was weak, but it was present.
AI retrieval does not work that way.
When a model answers a question, it pulls specific passages from specific sources and cites them. It needs a reason to extract your page over the next one. A distinctive, specific claim is a reason. A generic mission statement is not.
Remember the retrieval mechanism from Part 2. The model does not browse a ranked list of ten blue links and let the patient scroll. It retrieves passages, weighs them, and writes one answer naming a few practices. Generic content gives it nothing to retrieve.
There is no participation credit. The model is not ranking a list where you sit at position nine. It is choosing two or three practices to name. You are either specific enough to be chosen, or you are not in the answer at all.
Picture the two Cleveland practices again. The first has a sciatica page that names the nerve, describes the assessment, explains what the first three visits look like, and is honest about when a patient should be referred out. The second has a page that says sciatica is painful and they can help. When the model builds its answer, only one of those pages holds a sentence worth quoting. The other was never in the running.
Generic content was always weak. Now it is actively invisible.
A named technique, described in enough detail that the page demonstrates you actually do it.
A named condition covered with real depth, the kind that answers the follow-up questions a patient would ask.
A named outcome, with the honesty to describe who it works for and who it does not.
A named population. The runners. The desk-bound. The post-surgical. The pregnant.
A named place. Not "the greater metro area." The actual neighborhoods you serve.
A Miami practice that documents exactly how it treats a specific condition, for a specific kind of patient, in plain language, beats a site that says "we treat back pain" every time. Not because it is louder. Because it is quotable.
The AI does not need you to be good. It needs you to be specific.
The shift is uncomfortable because specificity feels risky. Generic feels safe. Generic offends nobody. But safe is now the same as invisible.
Quotable beats optimized.
The Part That Stings
Most chiropractic websites are generic by default, including the expensive ones. They were built to look professional. Polished, balanced, inoffensive. Nobody built them to be citable, because until recently citable was not a thing a website had to be.
That is no longer optional.
In Part 4 we take the next step. We break down what a citable practice page actually looks like, element by element, so you can see exactly where generic ends and quotable begins.
