AI share of voice
Definition
AI share of voice is your brand's proportion of mentions in AI-generated responses relative to competitors, measured across a defined set of prompts and platforms. It adapts the traditional share of voice metric for AI search — where visibility lives inside chat answers and AI Overviews rather than ranked links or media impressions.
How it works
AI share of voice is calculated by running a prompt set against AI platforms — ChatGPT, Claude, Perplexity, Gemini, Google AI Overviews — and counting brand mentions across responses. Your share is the number of mentions of your brand divided by the total mentions of all tracked competitors, expressed as a percentage.
The formula:
AI Share of Voice = (Your brand mentions ÷ Total mentions of all tracked competitors) × 100
Tracking is most useful when prompt sets reflect real buyer behavior. Category-level prompts like "best project management tools for remote teams" or "alternatives to [competitor name]" surface vendor lists that AI systems treat as authoritative recommendations.
11%
Of domains are cited by both ChatGPT and Perplexity for the same queries
Industry analysis, 2026
40%
Higher reappearance rate for brands earning both a mention and citation in AI responses
AirOps research
50–500
Recommended prompt set size for statistically meaningful AI SOV measurement
Indexly
Why it matters
Traditional share of voice tracks brand visibility across press, social, and advertising. AI share of voice does the same job for the channel where buyers increasingly research, evaluate, and decide — AI answers. As AI search captures larger portions of informational queries, AI share of voice becomes the leading indicator for category leadership.
AirOps research has found that brands earning both a citation and a mention in AI responses are 40% more likely to reappear across consecutive answers. AI share of voice captures this compounding visibility effect — and shows when competitors are accumulating it faster.
How to measure it
Four steps for credible measurement:
-
Define a prompt set of 50–500 queries. Mix category-level prompts ("best X tools"), problem-statement prompts ("how do I solve Y"), and competitor-comparison prompts ("[brand A] vs [brand B]"). The prompt set should reflect how real buyers ask AI systems for help.
-
Run prompts across multiple platforms. AI responses vary significantly between ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews. Run each prompt 5–10 times per platform to smooth out sampling variance.
-
Parse responses for brand mentions. Count both citations (with links) and unlinked mentions. Both contribute to AI SOV; mentions without links still build brand recognition for buyers who later search the brand name directly.
-
Aggregate over a rolling window. Weekly snapshots are too noisy. 30-day rolling averages reveal genuine trend changes without overreacting to short-term volatility. Indexly automates this end-to-end across platforms.
Frequently asked questions
How is AI share of voice different from share of model?
AI share of voice is competitor-relative — your mentions divided by total competitor mentions. Share of model is prompt-set-relative — the percentage of relevant prompts in which your brand appears. Both metrics matter; they answer different questions. AI SOV tells you competitive position; share of model tells you absolute visibility.
Should unlinked mentions count toward AI share of voice?
Yes. AI systems mention brands by name even when they do not include a citation link. Unlinked mentions still drive brand recognition — buyers who hear your brand in an AI answer often search for it directly afterward. Excluding unlinked mentions undercounts real visibility.
How many prompts do I need to track?
A statistically meaningful baseline starts at 50 prompts. Most brands operating in a single category settle at 200–300. Brands with multiple product lines or geographies often track 500+ prompts segmented by intent and audience.
Does AI share of voice vary across platforms?
Significantly. Only about 11% of domains cited by ChatGPT are also cited by Perplexity for the same queries. A brand can have 50% AI SOV on Perplexity and 5% on ChatGPT. Per-platform measurement is essential — a single aggregate number hides the platforms where you are losing.
How often should I track AI share of voice?
Weekly measurement with 30-day rolling averages is the practical standard. AI responses are stochastic and platforms update their indices at different rates, so single-day snapshots are unreliable. Quarterly executive reporting on rolling 30-day data gives stable trends.
Share of model
Share of model is the percentage of relevant AI-generated answers in which your brand appears, measured across a defined set of prompts and platforms. It is the AI-search equivalent of share of voice and the headline metric for tracking GEO performance.
Citation probability
Citation probability is the likelihood that an AI system will cite a specific URL when generating a response to a target prompt. Unlike share of model, which measures brand visibility across a prompt set, citation probability is a per-URL metric — it tells you how strong an individual page is at earning citations.
AI brand mentions
AI brand mentions are the instances of your brand name appearing inside responses generated by AI assistants — ChatGPT, Claude, Gemini, Perplexity, Grok, and Google AI Overviews. Unlike traditional brand monitoring across social and press, AI mentions surface inside the answer a buyer is reading, making them a high-leverage demand signal for Generative Engine Optimization (GEO).
AI visibility score
The AI visibility score is a single composite number — typically on a 0–100 scale — that summarizes a brand's standing across AI assistants (ChatGPT, Claude, Gemini, Perplexity, Grok, AI Overviews) by blending mention frequency, citation rate, ranking position, sentiment, and AI-referred traffic. It is the executive-friendly headline metric for Generative Engine Optimization (GEO) programs.