Last updated: 11 October 2026
AI Share of Voice Explained: A UK Business Guide for 2026
AI share of voice is the percentage of AI-generated answers — across ChatGPT, Perplexity and Google AI Overviews — that mention your brand rather than a competitor's for a category-relevant question. Most B2B brands appear in under 30% of relevant queries regardless of SEO performance, according to LLM Pulse (2026), making it a scarce, competitive metric worth tracking directly.
Key Takeaways
- AI share of voice measures the percentage of AI-generated responses in a category that mention a specific brand, not website traffic or Google rankings.
- AI Overviews appeared on roughly 48% to 50% of all US Google Search queries by February 2026, up from 6.49% in January 2026, according to BrightEdge via Omnibound (2026).
- Only 17% of AI Overview citations came from pages ranking in the organic top 10 by 2026, down from 76% in mid-2026, according to BrightEdge and ALM Corp data via Omnibound (2026) — so top Google rankings no longer guarantee AI visibility.
- Most B2B brands appear in under 30% of relevant AI queries regardless of their SEO performance, according to LLM Pulse (2026).
- ChatGPT generates 87.4% of all AI referral traffic to websites, with Perplexity a distant second, according to Conductor via AirOps (2026).
What is AI share of voice?
AI share of voice is a competitive visibility metric that quantifies how often generative AI tools mention, cite or recommend a brand compared with its named rivals across the same set of prompts. It measures presence inside a synthesised answer — the output of a large language model ("LLM", a system trained to generate human-like text, such as those powering ChatGPT or Gemini) — rather than a position in a ranked list of search results.
HubSpot defines AI share of voice as "a competitive metric that expresses what percentage of brand mentions in AI-generated answers belong to your company versus others in your category." A related definition from OptimizeGEO states it more simply: "AI Share of Voice is the percentage of AI-generated responses in your category that mention your brand."
GEO Metrics / TryGeometrics notes that AI share of voice "is not a single number — it is a multidimensional metric segmented by engine, by prompt, by position and by time period." A brand might lead in ChatGPT but be invisible in Perplexity, or dominate one product category while trailing in another. A Profound analysis found only 6.82% overlap between ChatGPT citations and Google's top 10 results, meaning roughly 80% of AI-cited sources never appear on Google's first page for the same query (AuthorityTech, citing Profound, 2026).
How AI share of voice differs from traditional share of voice
Traditional share of voice measures paid media spend, organic search rankings, or social mentions relative to competitors. AI share of voice measures something structurally different: whether a generative model chooses to name a brand at all when synthesising an answer, with no ranked list to climb.
The model either cites the brand, paraphrases its content without attribution, or ignores its category entirely in favour of a rival.
This distinction matters practically. A business can rank #1 on Google and still score zero in AI share of voice, because only 17% of AI Overview citations come from pages ranking in the organic top 10, according to BrightEdge and ALM Corp data reported via Omnibound (2026).
This pattern is also visible between individual AI engines: only 11% of domains cited by ChatGPT also appear among domains cited by Perplexity, according to Digital Applied (2026) — so tracking a single engine routinely misses most of a brand's actual AI visibility.
How do you calculate AI share of voice?
Calculating AI share of voice means running a defined set of category-relevant prompts through multiple AI engines, recording every brand mention, and dividing your brand's mention count by the total mentions across all competitors in the same result set. The formula is: (your brand mentions ÷ total category mentions) × 100.
Two established methodologies exist. Alex Birkett's guide distinguishes entity-based measurement — counting how often a brand name appears anywhere in an AI response — from citation-based measurement, which counts only instances where the AI engine explicitly links to or cites a website as a source. Entity-based scores are typically higher and reflect brand awareness; citation-based scores are stricter and more closely tied to referral traffic potential.
A common shortcut many teams take when building this kind of measurement is reporting share-of-voice from a handful of hand-picked prompts. Tiny prompt sets produce whatever story a team wants them to: running a stable, sizeable prompt set and keeping the misses in the data produces a trend that moves slowly month to month — and that steadiness is precisely what makes it trustworthy.
The core measurement steps
- Build a representative prompt set. Include informational queries ("what is the best CRM for small UK businesses"), comparison queries ("Brand A vs Brand B"), and transactional queries ("where can I buy X in London"). Aim for at least 30-50 prompts rather than a handful — a small sample produces results that swing wildly month to month with no underlying change in content.
- Run prompts consistently across engines. Test the same prompts on ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Microsoft Copilot, since responses vary meaningfully between them.
- Record every brand mention, noting whether it is a named mention, a hyperlinked citation, or a recommendation with ranking language ("the best option is...").
- Segment by prompt category, engine and time period, because — as GEO Metrics notes — a single blended score hides where a brand is actually winning or losing.
- Benchmark against named competitors, not an abstract "industry average".
- Repeat monthly or quarterly, since AI model outputs shift as underlying training data, retrieval indexes and search integrations update.
Semrush's guide to measuring AI share of voice recommends using a dedicated AI visibility toolkit to automate this process rather than manually prompting each engine, since manual testing does not scale once dozens of prompts are being tracked across six engines. A free AI visibility checker can be a useful starting point before committing to a paid tracking workflow.
Entity-based vs citation-based measurement
| Method | What it counts | Best for | Typical score range |
|---|---|---|---|
| Entity-based | Any mention of your brand name in the AI response text | Measuring brand awareness and category presence | Generally higher, often 20–60% for known brands |
| Citation-based | Only instances where the AI links to or names your website as a source | Measuring referral traffic potential and GEO ROI | Generally lower and stricter, often under 30% |
Why is AI share of voice important for UK marketing teams?
AI share of voice matters because generative AI is displacing traditional search as the first touchpoint for research-stage buyers, and brands absent from AI answers lose visibility at the exact moment prospects form their consideration set. Gartner predicted in a 2026 forecast that traditional search engine volume would drop 25% by 2026 due to AI chatbots and virtual agents, according to Omnibound.
This shift is visible in behaviour data. A Pew Research Center study from March 2026 covering 68,879 queries found AI Overviews appeared on roughly 18% of all Google searches.
More tellingly, AI summaries appeared in 60% of search queries that began with question words such as "who," "what," "when," or "why." Question-based queries are precisely where B2B buyers research vendors, so a UK software company or professional services firm invisible in AI answers is missing exactly the moments that shape a shortlist.
Why traffic alone understates the stakes
Direct AI referral traffic still looks small in aggregate. Conductor's data via AirOps shows AI referrals accounted for just 1.08% of total traffic across the industries studied in 2026.
But this figure understates influence, because most AI-driven brand consideration happens without a click — a user reads a ChatGPT summary recommending three vendors and contacts them directly, with no referral ever logged in Google Analytics or GA4.
ChatGPT drives 87.4% of all AI referral traffic that does register, with Perplexity a distant second — so a UK business prioritising GEO (generative engine optimisation, the practice of structuring content so AI systems cite it directly) should weight ChatGPT visibility heaviest in its measurement plan.
Which tools measure AI share of voice?
AI share of voice can be tracked using tools ranging from manual prompt-testing spreadsheets to automated citation-tracking platforms that monitor multiple engines continuously. The right choice depends on how many prompts, competitors and engines a business needs to monitor, and how often it needs fresh data.
Manual approaches — running prompts by hand in ChatGPT or Perplexity and logging results in a spreadsheet — work for a small business testing five or ten prompts once a quarter. This approach becomes impractical fast: testing 50 prompts across six engines monthly means 300 manual queries, each requiring careful logging of mention type, position and sentiment.
Automated platforms solve this by running scheduled prompt batches, tracking citations across engines simultaneously, and surfacing competitor comparisons without manual re-entry. Aether AI, for example, is a self-service GEO platform that tracks citations across six AI engines — ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Microsoft Copilot — alongside keyword and competitor tracking and Google Search Console integration, giving UK marketing teams a single dashboard rather than six separate browser tabs. Aether AI publishes its pricing openly and offers a free AI-visibility audit at aether-ai.co.uk/audit as a starting point before any commitment.
Choosing between manual tracking and a dedicated platform
| Factor | Manual prompt testing | Dedicated AI SOV platform |
|---|---|---|
| Time cost | High — grows linearly with prompts × engines | Low — automated scheduled runs |
| Engine coverage | Usually 1–2 engines tested consistently | Can cover 6 engines (ChatGPT, Perplexity, AI Overviews, Claude, Gemini, Copilot) |
| Competitor benchmarking | Manual comparison, error-prone at scale | Built-in competitor and keyword tracking |
| Historical trend data | Difficult to maintain consistently | Automatically logged over time |
| Best suited to | Very small prompt sets, occasional checks | Ongoing GEO strategy and reporting |
Budgets typically scale with the number of prompts, engines and competitors monitored. A solo founder checking a handful of core topics across two or three engines usually fits an entry-level or free-tier product; a mid-size marketing team tracking competitor benchmarking and all six engines with Search Console integration typically needs a standard paid tier; and an agency managing several brands usually needs a higher tier with multi-brand dashboards and exportable reporting.
Who should own AI share of voice tracking within a UK marketing team?
Ownership of AI share of voice tracking should sit with whoever already owns organic search performance — typically an SEO lead, head of content, or in smaller businesses, the founder or marketing manager — because the discipline extends existing search strategy rather than replacing it. In practice, this person needs three things to act on the data: access to the content management system to update or create pages, visibility into Google Search Console data, and authority to brief writers on structural changes such as adding definitions, statistics or comparison tables.
For larger UK teams, a monthly review cadence involving both SEO and content leads tends to work better than assigning the task to a single junior analyst, since interpreting citation data well requires editorial judgement as much as reporting.
How is AI share of voice different from organic search share of voice?
Organic search share of voice measures ranking position and click-through share across a defined keyword set on Google, while AI share of voice measures whether a brand is named or cited inside a synthesised AI answer regardless of ranking position. A page can hold position one on Google and never appear in the AI Overview above it — and conversely, a page ranking on page two can be the exact source an AI model chooses to cite.
This means a UK business measuring only Google rankings — via tools like Google Search Console or traditional rank trackers — is measuring an increasingly incomplete picture of its actual search visibility. See this site's guide to brand visibility in AI tools for more on why GEO and traditional SEO now require separate, parallel measurement.
What is a good AI share of voice benchmark for B2B brands?
A good AI share of voice benchmark depends heavily on category competitiveness, but the available data suggests most B2B brands should not expect to dominate their category. LLM Pulse reports that most B2B brands appear in under 30% of relevant queries regardless of SEO investment, meaning a score in the 20–30% range against named competitors already represents strong category presence for many sectors.
Because AI share of voice is multidimensional — varying by engine, prompt type and time period, according to GEO Metrics — a single blended benchmark number can mislead a marketing team. A UK fintech scoring 40% in ChatGPT but 5% in Google AI Overviews doesn't have a "35% average" problem; it has a specific, fixable AI Overviews visibility gap.
Setting realistic internal targets
- Establish a baseline first — a brand cannot improve what it hasn't measured against its own historical scores and named competitors.
- Segment targets by engine, since ChatGPT (87.4% of AI referral traffic) deserves more weight than lower-traffic engines in most GEO strategies.
- Segment targets by query intent, treating "best X" comparison prompts separately from purely informational "what is X" prompts.
- Review quarterly at minimum, given how quickly AI Overview prevalence has moved — from 6.49% in January 2026 to roughly 48–50% by February 2026 in the US market.
How do you improve your AI share of voice score?
Improving AI share of voice requires publishing content that generative AI models can easily extract, attribute and cite, rather than content optimised purely for human skimming or keyword-matching search crawlers. This means structuring pages with clear definitions, named entities, and self-contained factual statements that survive being lifted out of context and quoted in isolation.
Academic research from Princeton University supports this approach directly. Its GEO-bench study tested roughly 10,000 queries across nine datasets and found that specific content interventions — adding citations, quotations and statistics to source pages — could boost a source's visibility in generative engine responses by up to 40% (Aggarwal et al., 2026).
Practical levers include structuring content around direct question-and-answer formats — since 60% of question-word queries trigger AI summaries, according to Pew Research Center — naming real competitors and industry bodies rather than vague generalities, and maintaining consistent, accurate brand information across a company's own site, review platforms like Trustpilot, and third-party directories that LLMs draw on during training and retrieval.
Structured data, clean heading hierarchies, and up-to-date statistics with clear sourcing all increase the odds that an AI system treats a page as a reliable, citable source rather than paraphrasing a competitor instead. See this site's AI SEO software guide for more on the content and technical levers involved.
Your AI share of voice checklist
- Define a representative prompt set covering informational, comparison and transactional queries relevant to your category.
- Build the set to at least 30-50 prompts, since small samples produce unstable, misleading month-to-month swings.
- Test prompts consistently across ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Microsoft Copilot.
- Record whether each mention is a named brand reference or an explicit citation with a link.
- Benchmark your score against named competitors, not a vague industry average.
- Segment results by engine and query type rather than relying on one blended percentage.
- Track Google Search Console data alongside AI citation data to compare organic and AI visibility trends.
- Re-run the full prompt set at least quarterly, given how fast AI Overview prevalence is changing.
- Publish content with clear definitions, named entities and self-contained factual statements that AI engines can extract cleanly.
FAQ
What is AI share of voice?
AI share of voice is the percentage of AI-generated answers within a category that mention a specific brand, measured across engines like ChatGPT, Perplexity and Google AI Overviews. It is a competitive visibility metric distinct from Google rankings, since only 17% of AI Overview citations now come from top-10 organic pages, according to BrightEdge and ALM Corp data via Omnibound.
How do you calculate AI share of voice?
You calculate AI share of voice by dividing your brand's mention count by total category mentions across a defined set of prompts, then multiplying by 100. Measurement can be entity-based (counting any brand mention) or citation-based (counting only explicit source links), as outlined in Alex Birkett's methodology guide.
What is a good AI share of voice benchmark for B2B brands?
Most B2B brands appear in under 30% of relevant AI queries regardless of SEO effort, according to LLM Pulse, so a score in the 20–30% range against named competitors already represents solid category presence. Benchmarks should be segmented by engine and query type rather than treated as one blended figure.
Does AI share of voice affect website traffic and conversions?
AI referral traffic remains small in aggregate — just 1.08% of total traffic across industries studied, according to Conductor via AirOps — but this understates influence, since most AI-driven consideration happens without a tracked click. ChatGPT alone drives 87.4% of the AI referral traffic that does register.
Measuring your AI share of voice with Aether AI
Understanding AI share of voice conceptually only helps if a UK marketing team can measure its own brand against named competitors, across every engine that matters. Aether AI tracks citations across six AI engines — ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini and Microsoft Copilot — combined with keyword tracking, competitor benchmarking and Google Search Console integration in a single dashboard. Aether Agency Ltd runs this same engine for its own clients, including Priority First, Aether Agency and Pulse Operations.
To see where a brand currently sits before investing further in GEO content, run a free AI-visibility audit at Aether AI's audit tool.