LLM Share of Voice — Definition & Meaning
What is LLM Share of Voice?
LLM Share of Voice is the percentage of relevant AI prompts across ChatGPT, Perplexity, Gemini, and Claude that cite or mention your brand versus your competitors. It's the AI-search equivalent of organic ranking share and is fast becoming the most predictive top-of-funnel metric for B2B SaaS.
Key points
- Measured by running a fixed prompt set (usually 50-500 prompts) against each LLM on a set cadence and counting brand mentions + citations.
- A citation (linked source) is worth ~3x a mention (unlinked text) because it drives referral traffic and reinforces the LLM's next-crawl weighting.
- Prompt sets should mirror your buyer's actual research journey — comparison prompts, "best X for Y" prompts, and objection-handling prompts.
- Track share of voice weekly, not daily — LLM outputs are non-deterministic and daily noise obscures the real trend.
- Increasing SoV requires: strong SEO foundation, comparison content, third-party mentions on high-authority sites the LLMs already trust.
Example
For 100 prompts in the "rank tracking" category, if ChatGPT cites you 22 times, Ahrefs 41 times, and Semrush 55 times, your ChatGPT SoV is 22% — a benchmark you can move quarter-over-quarter.
Frequently asked questions
How is LLM Share of Voice different from Google Share of Voice?
Google SoV weights by keyword volume and position. LLM SoV weights by prompt frequency and citation prominence. The two correlate but diverge on comparison and long-tail prompts.
How often should I run the prompt set?
Weekly for trend tracking, daily for launch-week monitoring. Running less than weekly means you miss the impact of LLM model updates (which happen every few weeks).
Which prompts should I include in my set?
Start with 3 buckets: "what is X" (category education), "best X for Y" (comparison), and "X vs Y" (head-to-head). LemRank's AI-search wizard suggests prompts based on your Semrush keyword profile.