How We Measure AI Visibility

Our Approach

We measure AI visibility by testing real questions on real AI platforms and observing the results. This is fundamentally different from SEO metrics like keyword rankings or traffic.

What We Test

For each client, we:

Identify 15-20 buyer questions — the kind of questions potential customers ask AI engines about the client's industry, products, or services

Test each question on ChatGPT, Google Gemini, and Perplexity

Record each AI response (45-60 total responses per audit)

Analyze whether the client's business is mentioned, cited, or recommended

Check the accuracy of the information provided

Assess the sentiment (positive, neutral, negative)

Compare results with 3 direct competitors

Technical Audit

In parallel, we perform a technical audit covering:

robots.txt configuration for AI crawlers

llms.txt presence and format

Structured data (JSON-LD / Schema.org) implementation

Sitemap accessibility

Server response to AI bot user agents

CDN or firewall bot blocking

What We Don't Do

We do not use a proprietary "AI visibility score" algorithm

We do not predict future AI behavior

We do not guarantee specific outcomes

We do not create fake testimonials or fabricated results

Limitations

AI responses are non-deterministic. The same question asked twice on the same platform can produce different answers. Our observations are snapshots taken on specific dates. Results will change over time.

The relationship between technical optimization and AI citations is correlational, not guaranteed causal. We improve the signals, but AI engines make their own decisions.

Transparency

Every observation in our reports includes: the exact question asked, the platform used, the date of the test, and the response observed. Clients can verify any observation independently.