How AI Visibility Works
The theory behind AI recommendations, scoring, and competitive analysis.
How AI Models Decide What to Recommend
Large language models don’t have a simple ranking algorithm like Google. Instead, they synthesize recommendations from multiple sources: their training data, real-time web searches (when available), and the specific context of the user’s query. The result is a nuanced, conversational response that feels like advice from a knowledgeable friend.
Several factors influence whether an AI recommends your business:
- Training data authority — Press coverage, industry publications, and high-authority websites that mention your business become part of the model’s knowledge. The more reputable the source, the stronger the signal (learn how each platform uses different data in The Five AI Platforms).
- Content freshness — Models with web access prioritize recent content. Outdated websites, old blog posts, and stale business information weaken your presence.
- Entity recognition — Schema markup, consistent business information (name, address, phone), and structured data help AI models identify and correctly categorize your business.
- Sentiment patterns — Reviews, testimonials, case studies, and how people talk about your business online shape the AI’s perception. Consistently positive sentiment leads to stronger recommendations.
- Category association — Niche expertise beats generalist positioning. A business clearly associated with a specific category gets recommended more often for that category than a business that tries to be everything.
These factors are measured by 11 distinct visibility metrics.
Why Each Platform Is Different
Not all AI platforms work the same way, which is why Vizaeo tests across all five. Each has different data sources, different user bases, and different tendencies.
ChatGPT
The largest AI platform by user base. ChatGPT combines its training data with real-time web browsing. It tends to recommend well-known, broadly visible businesses and is heavily influenced by website content and structured data.
Claude
Relies primarily on training data without real-time web search. Claude values depth and nuance, often providing more detailed reasoning about why it recommends a business. Strong authority signals in training data matter most here.
Gemini
Deeply integrated with Google Search, making it the most SEO-responsive AI platform. Businesses with strong Google presence — Google Business Profile, reviews, local SEO — tend to perform well on Gemini.
Grok
Built on X (formerly Twitter) data and social signals. Businesses with active social media presence and real-time engagement tend to surface more frequently on Grok.
Perplexity
Always searches the web before answering, making it the most responsive to current SEO and fresh content. If you’ve recently improved your web presence, Perplexity will reflect those changes fastest.
Consensus Scoring
AI responses are inherently variable. Ask the same question twice and you might get different businesses mentioned. This randomness makes single-query testing unreliable — you could get lucky or unlucky on any given run.
Vizaeo solves this with consensus scoring. Each query is run multiple times per platform, and only consistent mentions are counted. If a business appears in 4 out of 5 runs for the same query, that’s a strong signal. If it appears in only 1 out of 5, it’s likely noise. This approach produces reliable, reproducible data rather than lucky one-offs.
How Share of Voice Is Calculated
Share of Voice is calculated with a straightforward formula: your mentions ÷ total mentions × 100.
Here’s a concrete example. Suppose Vizaeo runs 40 queries across all 5 platforms for the category “family dentist in Austin.” Across all responses, there are 200 total business mentions. Your practice appears 30 times. Your Share of Voice is 30 ÷ 200 × 100 = 15%.
That 15% means that roughly one in seven AI mentions of family dentists in Austin includes your practice. The leading competitor might have 25%, meaning they appear nearly twice as often as you do. This gap is precisely what Vizaeo helps you understand and close. See how SoV translates into your Score tab.