Step 1
Define the market
Choose the commercial questions buyers ask and the competitors they are likely to compare. Keep this prompt set stable between reporting windows so movement means something.
AI share of voice software
Measure AI share of voice with the evidence behind every score
Foglift reports competitive share of voice by engine and keeps the prompt, answer, competitor mentions, position, sentiment, and returned source URLs behind every score. Start with weekly Perplexity monitoring for $0, or monitor all five engines with developer access from $49 per month.
ChatGPTPerplexityClaudeGeminiGoogle AI Overviews
The short answer
AI share of voice is a competitive mention percentage
Divide your brand mentions by all tracked brand mentions for the same prompt set, competitor set, engine set, and measurement window. Multiply by 100. Keep answer coverage beside the result because a competitive percentage alone does not show how many answers named no tracked brand.
Report every engine separately before looking at a combined view. ChatGPT, Perplexity, Claude, Gemini, and Google AI Overviews can retrieve and cite different sources for the same buyer question.
AI share of voice formula
Your brand mentions
All tracked brand mentions
× 100 for the declared comparison window
Zero-mention answers remain in the evidence set, while the share calculation uses the brands actually mentioned. Pair the percentage with answer coverage so silence is not hidden.
Do not blend the outcomes
Mentions, citations, referrals, and crawler fetches answer different questions
A defensible report labels each signal and its denominator. It never treats a crawler request as proof that an engine cited the page or sent a person to it.
| Metric | Definition | What it tells you |
|---|---|---|
| Answer coverage | Answers that mention your brand divided by successful answers | Shows how often your brand appears at all, including answers where no tracked brand is named. |
| Competitive share of voice | Your brand mentions divided by all tracked brand mentions | Shows your portion of the competitive conversation without hiding the denominator. |
| Citation presence | Whether the answer returns a brand-owned URL as a source | Shows cited evidence by engine. It is not interchangeable with a brand mention. |
| AI referral | A recognized human visit from an AI engine referrer | Shows when an answer sends someone to your site. Report landing pages separately. |
| Crawler fetch | A request from an identified AI crawler or user-fetch agent | Shows retrieval behavior. A fetch is not proof of a mention, citation, or human visit. |
Measure the evidence behind each win
A useful share-of-voice tool connects the headline percentage to the answer, source, and competitor evidence that explains it. That makes the metric actionable for content, product marketing, and digital PR.
| Signal | What Foglift keeps | Decision it supports |
|---|---|---|
| Brand mention | Whether your brand appears in each answer | Shows where buyers can discover you and where you are absent. |
| Competitor mention | Every named competitor in the same answer set | Keeps the share-of-voice denominator visible and comparable. |
| Cited source | The exact URL and domain cited by the engine | Reveals which owned or third-party pages support the winning answer. |
| Answer sentiment | Positive, neutral, or negative brand framing | Separates a useful recommendation from a mention that harms trust. |
| Prompt history | Prompt, engine, answer, and run date | Makes changes auditable instead of treating one volatile answer as a trend. |
From score to action
A repeatable AI share-of-voice workflow
Consistency turns volatile generated answers into a useful trend. Keep inputs fixed, retain the raw evidence, and review weak commercial prompts before changing the measurement design.
Step 2
Run the same prompts
Monitor those questions across the engines your buyers use. Foglift keeps the prompt and engine attached to every answer instead of blending away the underlying evidence.
Step 3
Count mentions consistently
For each answer, record your brand, every tracked competitor, cited sources, position, and sentiment. The score should be reproducible from those rows.
Step 4
Fix the losing prompts
Use cited pages and competitor patterns to decide whether the next step is a clearer product page, stronger comparison evidence, technical cleanup, or third-party coverage.
Find prompt gaps
Separate category discovery, comparison, use-case, and branded prompts. A healthy overall score can still hide absence on the few questions closest to a purchase decision.
Trace winning sources
See whether a competitor wins through its product page, documentation, original research, customer proof, or a third-party comparison. The cited source points to the evidence gap worth fixing.
Watch durable movement
Compare stable windows rather than reacting to one answer. Prompt history distinguishes a persistent gain from normal engine variation and gives teams an audit trail for releases and campaigns.
Published plan boundaries
Start free, then add engines and cadence when the program grows
Every plan keeps the answer evidence behind the score. Paid monitoring adds all five engines and the developer surfaces needed to move the same evidence into an existing reporting workflow.
Free
$0
One brand, weekly Perplexity monitoring while active
A permanent starting point for one prompt program
Launch
$49/month
Three brands, all five engines, daily to monthly schedules
Includes developer access and up to 4,000 monitoring tokens each month
Growth
$129/month
Ten brands, all five engines, up to twice-daily schedules
Includes up to 11,500 monitoring tokens each month
Frequently asked questions
What is share of voice in AI search?
Share of voice in AI search is the percentage of tracked brand mentions that belong to your brand across a declared prompt set, competitor set, engine set, and measurement window. It measures presence inside AI-generated answers, not traditional search traffic.
How do you calculate AI share of voice?
Divide your brand mentions by all tracked brand mentions for the same prompt set, competitors, engines, and measurement window, then multiply by 100. Keep the denominator and answer history available so a team can reproduce the score.
What is a good AI share of voice score?
There is no universal evidence-backed target. Compare your score against the same competitors and commercially important prompts over time. A rising score is useful only when the denominator, engine mix, and sampling cadence remain stable.
Why should each AI engine be reported separately?
AI engines retrieve and cite different source sets, so a blended total can hide a weak platform. Report every engine separately first, then use a combined view only when its weighting is declared and supported by your audience evidence.
How is AI share of voice different from SEO share of voice?
SEO share of voice typically estimates visibility from ranked search results and keyword volume. AI share of voice counts brand presence inside generated answers. An engine can recommend a brand or cite a third-party page even when the brand does not hold the equivalent classic ranking position.
See where your brand earns the answer
Start with weekly Perplexity monitoring on Free, or compare all five supported engines from $49 per month.
Start monitoring free