Updated May 2026: Added benchmarking section drawn from 75+ UK brands SearchIntel has scored in early 2026, vertical-average patterns, and the convergence-across-sources framing in response to the Brainlabs critique. Replaced the unverified Canel 30% schema stat with the qualitative SMX Munich confirmation.
Your brand's visibility is shifting from blue links to AI answers. And most marketing leaders have no idea how they're performing.
After 20 years in search, spanning Google, MediaCom (Adidas, Shell, Coca-Cola), IPG Mediabrands (LEGO), and TripAdvisor/Viator, I've watched every major algorithm shift. But this one's different. ChatGPT and Gemini aren't ranking websites. They're citing brands directly. Your content either makes the cut or it doesn't.
The problem? Most brands are flying blind. They're still optimising for Google's first page whilst AI models are already deciding which 3-5 brands deserve to be mentioned at all.
So we built a scorecard. Ten questions that separate brands winning the answer economy from those being systematically erased.
The Foundation Layer: Can You Even Compete?
These first three questions determine whether you're in the game at all.
1. Can you measure your visibility in ChatGPT and Gemini?
If you're not tracking AI citations, you're managing in the dark. Traditional rank tracking is increasingly insufficient when 60% of Google searches end without a click. You need to know: when prospects ask AI about your category, does your brand appear? How often? In what context?
Most marketing teams can't answer this. They're reporting on metrics from 2019 whilst their competitors are already optimising for language models.
2. Do you have "definitional" pages for your core terms?
AI models prioritise authoritative definitions. If someone asks "what is [your core product category]", do you have a page that directly answers that? Not a product page buried in sales copy. A clear, structured explanation that establishes expertise.
Banks without dedicated pages explaining "bridging loans" or "remortgaging" are losing citations to comparison sites. SaaS companies without clear "what is X" pages are being defined by industry blogs instead.
This is good SEO practice that directly transfers to AI visibility. Strong SEO foundations make your content retrievable by AI models in the first place.
3. Is your structured data implemented correctly?
Schema markup is no longer optional. It's how you explicitly tell AI models what your content means. Organisation schema, FAQ schema, product schema. These are the annotations that help language models understand and cite you accurately.
Microsoft's Fabrice Canel confirmed at SMX Munich 2025 that schema markup directly helps Bing's LLMs and Copilot understand page content. Practical lift varies by site, but pages without structured data are systematically harder for AI models to parse and cite accurately.
I've seen brands with genuinely superior content get overlooked because they didn't bother with structured data. The AI couldn't parse what they were actually saying.
Competitive Position: Where Do You Stand?
Questions 4-5 assess your position relative to rivals.
4. Do you appear more or less than your top competitor?
In the old Google world, ranking position mattered. In AI, it's simpler and more brutal: are you mentioned or not? And if multiple brands are mentioned, are you first, third, or absent entirely?
Run 20 category-relevant prompts across ChatGPT, Gemini, Claude and Perplexity. Count mentions. If your main competitor appears in 15 responses and you appear in 6, you're losing citation share. That gap compounds over time as AI models reinforce what they've previously learned.
5. Are you framed positively or neutrally?
This is where brand positioning shows up in AI outputs. Models don't just list brands randomly. They contextualise. "Best for enterprise teams" carries different weight than "another option to consider."
If you're consistently positioned as a fallback rather than a leader, your brand messaging isn't translating to AI training data. That's fixable, but only if you're measuring it. We call this your share of memory, how AI models remember and recommend you.
Citation Gravity: Does the Internet Endorse You?
The middle layer (questions 6-8) measures whether third-party sources validate your authority. This is where AI visibility goes beyond traditional SEO.
6. Do third-party sites cite you in comparison content?
AI models train on the entire web. If industry blogs, review sites, and comparison platforms consistently mention your brand, models learn to include you. If they don't, you're functionally invisible.
The credible research is consistent. Omniscient Digital's analysis of 23,000+ AI citations found that on branded queries, the vast majority of citations come from earned media: reviews, news articles, forum discussions and industry coverage. Your own website is one input among thousands. Being part of the conversation matters more than ever. Guest articles, expert commentary and case studies on partner sites all build citation gravity.
7. Are you mentioned in Wikipedia or industry reports?
Wikipedia and Reddit consistently rank among the most-cited sources in AI training data and live retrieval. Across major engines, Reddit hovers around 40% of citations per the Semrush 150,000-citation analysis and the 5WPR AI Platform Citation Source Index 2026, with Wikipedia consistently in the top three. Share is volatile: ChatGPT's Reddit citation rate fell sharply in late 2025 before recovering. Analyst reports from Gartner, Forrester and category-specific research firms also carry significant weight. If your brand appears in these sources, language models are far more likely to cite you.
Most mid-market brands ignore this entirely. They assume Wikipedia is for enterprises or that analyst relations is too expensive. Meanwhile, their well-cited competitors are locking in AI visibility that compounds over time.
8. Have you earned PR in the last 90 days?
Fresh citations signal relevance. Brands that regularly appear in news articles, industry press, and trade publications maintain momentum. Brands that went quiet three years ago fade from AI responses, even if their product is objectively superior.
This doesn't mean press releases. It means actual earned coverage that journalists and bloggers cite naturally. Consistently appearing in high-quality editorial coverage is how AI systems learn to trust and recommend brands.
The Strategic Layer: Priority or Afterthought?
The final two questions reveal organisational readiness.
9. Do you track AI visibility as a KPI in exec reports?
If your board deck still shows Google rankings and organic traffic but no AI citation metrics, you're optimising for yesterday's game. Leadership needs to see: citation volume, share of voice in AI responses, competitive positioning in model outputs.
Until AI visibility sits alongside revenue and customer acquisition metrics, it won't get the investment it requires. AI Overviews reduce clicks to the top-ranking page by 58%. That's a board-level concern, not a marketing footnote.
10. Can you prove ROI of content using citation metrics?
Content marketing has always struggled with attribution. But AI citations offer a direct line: this article was cited 47 times by ChatGPT this quarter, resulting in tracked conversions from users who mentioned the AI recommendation.
If you can't connect content investment to AI visibility to pipeline, you can't defend the budget when the CFO asks why you're still publishing 40 articles a month.
What we've seen scoring 75 UK brands
This scorecard isn't theoretical. SearchIntel has run the underlying methodology against more than 75 UK enterprise brands across pensions, wealth management, travel, comparison and publishing in the first four months of 2026. A few patterns worth sharing before you score yourself.
Vertical averages diverge sharply
Across 19 UK comparison and marketplace brands tested in April 2026, the mean visibility score was 71/100. Auto Trader, Trustpilot and Just Eat all scored 96+. Within the same batch, Bauer Media UK scored 16/100. Across 20 UK pension brands tested the same month, the mean was 62/100. NEST scored 100. Capita Pensions scored 0.
The spread inside a single vertical is wider than the spread between verticals. That tells you the question to ask isn't "is my industry visible." It's "what are my best-cited competitors doing that I'm not."
Third-party citation gravity correlates with score
Looking across the dataset, brands scoring 60+ consistently shared a profile: Wikipedia article live, regulatory or government source pages citing the brand, trade-press quotes from named executives in the last 90 days, Reddit or community discussion in category subreddits. Brands under 40 had thinner off-site presence regardless of how strong their own website was.
This isn't a backlinks correlation. It's a citation-context correlation. AI models weight what others say about you, in the language they say it. Volume of links is downstream.
How precise is your score, really?
Worth being honest about: a single visibility score is a snapshot, not an audit metric. SparkToro's research found a less than 1-in-100 chance that two identical AI prompts return identical brand recommendations, and closer to 1-in-1,000 for the same ordering. Anyone showing you a precise number from a single run is selling false precision.
Our methodology runs each prompt three times across five platforms, takes the median frequency and reports a band rather than a point estimate. The number you should track over time is the trend, not the score. A brand moving from 38 of 50 prompts to 44 of 50 in six weeks is the signal. The headline number is just the cover slide.
Convergence beats any single measurement
Brainlabs Digital published a useful piece in May 2026 arguing that no single measurement method is fully accurate, direct prompt testing has selection bias, clickstream panels have panel skew, server logs only catch click-throughs which are now under 1% of AIO interactions. They're right. The way to get to truth is to triangulate across methods that have different biases.
If your scorecard reading converges with what your GA4 shows for chatgpt.com and perplexity.ai referrals, and converges with what panel-based tools show for your category, you have a load-bearing signal. If only one source says it, treat it as a hypothesis. This is the level of methodology hygiene executive reporting needs in 2026.
Score Your Brand
Count your "yes" answers honestly.
What Happens Next
The answer economy isn't coming. It's here. ChatGPT handles over 2.5 billion queries per day. Google's AI Overviews dominate the SERP. Perplexity processes over 60 million queries per day. These platforms are answering questions that used to drive traffic to your site.
Brands that score well on this scorecard aren't lucky. They're deliberate. They measure what matters, they build citation gravity, and they've made AI visibility a strategic priority, built on strong SEO foundations with the additional layers that AI requires.
The rest are slowly disappearing from the conversation.