Short answer: In AI search, being mentioned means an AI answer names your brand, while being cited means your website is linked as a source. According to Semrush's AI Visibility Index 2026, fewer than half of the top 100 mentioned brands also appear among the top 100 cited domains across ChatGPT, Google AI Overviews, Gemini and AI Mode.
- ChatGPT has the widest gap: only 28.7 to 35.8 of 100 domains overlap between its mention and citation lists.
- Famous brands are often not cited: Ford, Toyota, Disney and Nintendo rank in the top 100 for mentions but outside the top 500 for citations in the US.
- Each platform cites different sources: ChatGPT favors news, AI Overviews favors YouTube and medical sites, and AI Mode concentrates on a few large platforms.
- Citations drive direct referral clicks; mentions drive indirect traffic through branded search.
- The fix: measure mentions and citations separately by platform, market and category, then audit the pages being cited instead of yours.
Your AI visibility is two positions, not one score
A brand can be named in ChatGPT answers constantly while its website almost never shows up as a cited source. That is the headline finding of Semrush's AI Visibility Index 2026, which tracked ChatGPT, Google AI Overviews, Gemini and AI Mode across the UK and US from January to July 2026 (Semrush).
The scale is significant: 2.5 million UK prompts and 25 million US prompts per month on ChatGPT, Gemini and AI Mode, plus a separate keyword pool of 4.35 to 4.80 million (UK) and 48.34 to 54.60 million (US) for AI Overviews.
For agencies and in-house teams running organic and paid search, the lesson is simple. If you report "AI visibility" as a single number, you hide the gap that decides what to fix next.
Cited vs mentioned: what each term means
| Mentioned | Cited | |
|---|---|---|
| What it records | The AI answer names your brand | Your domain appears as a source link |
| What it signals | Brand salience inside the answer | Your pages are used as evidence |
| What it does not mean | A recommendation or positive sentiment | Trust or editorial approval |
| Typical levers | Brand building, PR, third-party coverage, reviews | Crawlable, specific, current pages; technical access; source placements |
The mention-citation gap, in numbers
Fewer than half of the top 100 mentioned brands also sit in the matching top 100 citation list. That held across 22 categories, four platforms and every month measured.
Mention and citation lists rarely agree
ChatGPT shows the widest split. Being talked about there tells you almost nothing about whether your pages are being used as sources.
The Tesco and Asda example makes it concrete. Both rank in the top 100 for mentions on all four platforms every month. In July, ChatGPT ranked Tesco as its 25th most cited domain and Asda as its 606th. On Gemini the two sat close together at 29th and 60th. Same brands, same market, very different source positions depending on the engine.
The One-Way Door: famous brands, invisible websites
Semrush calls it the One-Way Door: a domain inside all four top-100 mention lists while sitting outside all four top-500 citation lists. The reverse (highly cited, rarely mentioned) almost never happens. Each month, 3 to 9 UK domains and 8 to 13 US domains held the One-Way Door position, while the reverse group was 0 in the UK and 0 to 1 in the US.
In the US, seven brand domains held the position every month: Chevrolet, Disney, Ford, Honda, Marvel, Nintendo and Toyota. Ford shows the pattern clearly.
ford.com: top 25 for mentions, past 700 for citations
The opposite case exists at category level. In US Finance on ChatGPT, Investing.com ranked 7th as a cited source but 50th as a mentioned brand. McMaster-Carr showed the same shape in Industrial.
Only Apple and YouTube achieved true balance in both markets: top 100 on both lists, on every platform, every month, with no rank gap above 25. Reddit and Walmart joined them in the US. Balance is rare, so do not assume a strong position on one list carries over to the other.
Each AI platform cites a different web
The same customer question produces different brands and different sources on each engine. Optimizing for "AI search" as one channel misses this.
ChatGPT links the most; Gemini names more brands than it links
| Platform | Leading source types | What it means for your plan |
|---|---|---|
| ChatGPT | News leads: 29.5% UK, 19.6% US | Digital PR and earned media in the publications it cites |
| AI Overviews | Streaming (YouTube heavy), news, medical reference, UK government | Video content and YMYL accuracy matter; closely tied to classic SEO keywords |
| Gemini | News in UK; streaming and online learning in US | Brand mentions matter more than links; mix changes by market |
| AI Mode | SaaS and community platforms (Facebook, Instagram, Pinterest, Quora); top 10 domains hold 23% to 39% of citation weight | A short audit of top sources covers a lot of ground; community presence counts |
AI Overviews also contributed the most single-platform domains every month (36 to 49 in the UK, 32 to 47 in the US), so classic Google rankings remain a lever that ChatGPT tracking will not reveal.
For paid search teams this matters too. AI Overviews and AI Mode sit on the same results pages where you buy clicks. Knowing which domains own the answer above your ad helps you decide where to bid harder, where to defend branded terms, and which query themes to brief into organic content instead.
Your market and category set the benchmark
A plan that works in one country does not automatically transfer. In every month, the four UK platforms agreed less on mentions and sources than the US platforms did. Only 16.7% to 26.0% of leading UK domains held a top-100 spot on all four platforms, against 23.5% to 36.5% in the US. Ford outranked Toyota for citations in the UK, while Toyota led in the US.
Category matters even more than platform. The spread between categories was wider than the spread between platforms in every US month and five of seven UK months.
- Sports: the UK's dual-role share (domains both cited and mentioned) trailed the US by at least 35 points every month. The US had four recurring core sources (CBS Sports, Fox Sports, MLB, Sports Illustrated); the UK had one (BikeRadar).
- Finance (UK): Lloyds Bank appeared on all four mention lists in July but none of the source lists, while MoneySavingExpert and Wise were cited on all four platforms every month.
- Real Estate: stable source cores in both markets (Zillow, Redfin, Realtor.com, Apartments.com in the US; Rightmove, Zoopla, OnTheMarket in the UK).
- Industrial, Food and Drinks: no recurring core at all, so teams need a tighter, custom competitor and question set.
Cited or mentioned: which drives more referral traffic?
The answer depends on your goal. Citations drive clicks; mentions drive awareness. Treat them as two separate tracks.
1. Traffic and lead track: citations
For when you treat AI engines like a search channel: a user clicks, visits and converts.
- The AI's logic
- LLMs do not like to guess. They use Retrieval-Augmented Generation (RAG) to support their assertions and look for hard, uncontradicted facts.
- Your strategy
- Structured, crawlable product data, transparent pricing and specific facts on your own site, plus a strong footprint on authoritative domains AI engines trust: review directories, Reddit, Wikipedia and mainstream tech blogs.
2. Brand equity track: mentions
For when you treat AI engines as top-of-funnel discovery or digital PR.
- The AI's logic
- The AI acts as a digital advisor. When a user asks for recommendations, you want your brand named as a trusted option. The user may not click, but a seed is planted.
- Your strategy
- Build semantic association so your brand is strongly connected to your category across the content AI models learn from and retrieve. Measure lifts in branded search volume and direct traffic.
| Strategy goal | Core target | Primary tactics | Success metrics |
|---|---|---|---|
| Referral traffic and direct conversions | Citations (the clickable links) | High-authority third-party PR; structured data and clear schema; inclusion in top listicles and reviews | Referral traffic from AI platforms; affiliate and direct conversions |
| Demand generation and consideration | Mentions (the textual callouts) | Category-affinity building; forum and community seeding (Reddit); broad digital footprint | Share of voice in AI outputs; branded search lift; direct domain traffic |
If your goal is measurable referral traffic and conversions, prioritize AI citations: crawlable, indexable pages with specific, accurate, up-to-date information, and placements on the authoritative sources AI platforms reference. If your goal is awareness and consideration, track brand mentions across relevant non-branded queries and measure branded search volume, direct traffic and assisted conversions over time.
What to do next: match the fix to the gap
| Your position | What to investigate | Likely owner |
|---|---|---|
| Mentioned, rarely cited | Which pages are cited instead (yours and third-party), and whether your pages are crawlable, specific and current | SEO and content |
| Cited, rarely mentioned | How answers describe your brand and whether cited pages make your role clear | Content and brand |
| Strong on both | Which pages and prompts drive it, and whether they recur | SEO and analytics |
| Absent from both | Broader discovery: relevant competitors, sources and prompts | Strategy |
A 90-day structure keeps this accountable:
Two cautions from the study itself: a pairing in AI answers does not prove two brands compete, and a visibility change does not prove a revenue change. Track business outcomes separately. Still, Adobe data cited in the report shows AI-referred visitors convert 60% better and generate 53% more revenue per visit, so the upside is worth measuring properly.
Need help with the audit? Our AI search optimization service builds separate mention and citation baselines and ties them to your SEO and Google Ads plan.
Sources
- Semrush (An Adobe Company), AI Visibility Index 2026: Where the top brands stand in the UK and US. Data January to July 2026.
Remar Barquilla is the founder of remar.digital and the developer of SiteAud!tLint. He writes about technical SEO, AI search and measurable growth. More articles by Remar ↗