What Our Chiropractic AI Citation Study Revealed

Why we ran this study

Everyone in digital marketing has opinions about what gets businesses recommended by AI. We wanted data instead of opinions.

So we designed a pre-registered study — meaning we locked in our methodology before collecting any data, to prevent cherry-picking results. We tested 20 U.S. cities across ChatGPT, Perplexity, and Gemini using 120 query templates that covered specific conditions, general recommendations, and various ways patients ask for help.

The result was over 12,000 AI responses to analyze. Here's what we found.

Finding 1: Website structure is the #1 predictor

The way a chiropractic practice's website is organized — with clear page titles, proper heading hierarchy, and logical navigation — was the single strongest difference between practices AI recommends and those it ignores.

The effect size was d = 1.14, which in research terms is massive. For comparison, most meaningful findings in social science have effect sizes around 0.3 to 0.5. This was more than double that.

Finding 2: Domain authority is nearly irrelevant

The correlation between domain authority and AI citations was r = -0.05. That's essentially zero. Whether a practice had a domain authority of 15 or 55, it made virtually no difference in whether AI recommended them.

This is the finding that should make every practice owner rethink their marketing spend. If your agency is charging you for link building as your primary strategy, that money isn't helping you in AI search.

Finding 3: Condition pages are a moat

82-89% of practices that AI cited had dedicated pages for specific conditions — sciatica, lower back pain, sports injuries, and so on. Practices without these pages were nearly invisible to AI, regardless of their other qualities.

This is one of the most actionable findings: create a dedicated page for each condition you treat. It's a straightforward change that dramatically increases your chances of being recommended.

Finding 4: Each AI model picks different practices

Only 8.7% of practices were recommended across all three AI platforms. ChatGPT, Perplexity, and Gemini evaluate content differently and often reach different conclusions about who to recommend.

This means optimizing for just one platform isn't enough. You need to be visible across all of them — which fortunately comes down to the same fundamentals: clear structure, dedicated pages, and well-organized information.

Finding 5: Structured FAQ is almost unused

Only 1.6% of practices had their frequently asked questions set up in a format AI can directly read (FAQ schema markup). This is one of the simplest technical changes a practice can make, and almost nobody has done it.

What this means for chiropractors

The opportunity is wide open. Less than 2% of chiropractic practices have the structural elements AI needs to recommend them. The practices that move first in each market will have a head start that compounds over time.

And the changes needed aren't expensive or complicated. They're structural — reorganizing content, creating condition-specific pages, and adding the right technical signals. Most practices can implement them without rebuilding their entire website.