When someone asks ChatGPT, Claude, or Perplexity to recommend a product, the AI doesn't Google it in real time and pick the top result. It draws on everything it has already learned, and some platforms on what it can currently find on the web, to decide which brands are worth mentioning. The signals that shape that decision are called AI visibility signals, and most brands have no idea they exist, let alone that they're already being judged by them.
This guide explains what AI visibility signals are, what "top picks" and "brand mentions" actually mean in this context, and what brands should (and definitely shouldn't) be doing if they want to show up when it matters most.
What Are AI Visibility Signals?
AI visibility signals are the pieces of information that artificial intelligence platforms use to decide whether your brand gets recommended, mentioned, or ignored when a consumer asks for advice. Think of them as the behind-the-scenes evidence that tells an AI: "yes, this brand is credible, well-regarded, and worth suggesting."
They're not the same as traditional SEO signals. You can't just tweak a title tag or build a few backlinks and expect your AI visibility to improve overnight. AI platforms are looking for a completely different kind of proof, one that comes less from what you say about yourself, and far more from what other people say about you.
Research across more than 125 UK consumer products found a clear and consistent ranking of which signals carry the most weight with large language models:
- Retailer reviews (volume, recency, consistency) - Impact score of 100
- Authority references (expert bodies, institutions) - Impact score 72
- Brand longevity (how long you've been around) - Impact score 45
- Community discussion (forums, Reddit, social) - Impact score 38
- Editorial reviews (press, magazines) - Impact score 25
- Brand Messaging (your own marketing content) - Impact score 8
That last row is the one that surprises most brands. Your own website copy, your Instagram captions, your beautifully crafted brand story, they score 8 out of 100 when it comes to influencing AI recommendations. We'll come back to that.
What Is a "Top Pick" in AI Search?
A top pick is when an AI platform doesn't just mention your brand, it names you as the single best option for a given query. It's the difference between being in a list of five recommendations and being the one the AI leads with or singles out as its number one choice.
In a study of 100 real UK consumer queries run across Claude, GPT-4o, and Perplexity, Gymshark was named as the outright top pick 131 times. Lululemon (despite having a slightly higher overall visibility score) was named as top pick just 86 times. Being visible and being the top pick are two very different things, and the commercial value of that distinction is enormous.
When an AI gives someone a top pick, that's often the brand they go and buy. There's no scrolling through search results, no comparison shopping in the traditional sense. The AI has done the work. If you're not in the running for that top pick slot, you're effectively invisible at the most important moment in the purchase journey.
What Are Brand Mentions in AI?
A brand mention in AI is simply any time your brand name appears in an AI-generated response to a consumer query. The AI might include you in a list, reference you in a comparison, or bring you up unprompted when discussing a category.
Brand mentions matter for two reasons. First, they're a direct measure of how visible you are in AI-mediated discovery, the new version of being on page one of Google. Second, they compound. The more platforms mention you, the more your brand becomes part of the established conversation around your category, which in turn makes it more likely you'll be mentioned again.
The gap between brands that get mentioned and those that don't is staggering. Research tracking 12 activewear brands across 300 AI responses found that the top five brands captured 91% of all brand mentions. The remaining seven brands shared just 9%. One brand received zero mentions across all 300 responses, not because its products were worse, but because its digital presence gave AI nothing to work with.
"AI is not a level playing field. It is a winner-takes-most environment, and the winners were decided before most brands knew the game had started."
The Signals That Actually Move the Needle
Retailer reviews: the single most important signal
Nothing comes close to retailer reviews when it comes to influencing AI recommendations. Volume, recency, and consistency across platforms are what AI uses as its primary evidence when deciding whether to recommend a product.
Here's why: retailer reviews sit close to the actual purchase. They're written by real people who have bought and used the product, they aggregate at scale, and (crucially) they have no obvious reason to be biassed in the brand's favour. That combination makes them incredibly reliable evidence in the eyes of an AI that is trying to reduce uncertainty and give a trustworthy answer.
If your reviews are thin, out of date, or inconsistent across platforms, AI doesn't respond by being cautious about recommending you. It responds by substituting weaker evidence and still making a confident recommendation, which often means recommending someone else, or worse, misrepresenting you entirely.
Authority references
Being cited by respected institutions, industry bodies, or expert sources carries significant weight. If a trusted third party has staked its reputation on saying you're good, that's a meaningful signal to an AI working out who deserves a recommendation. Think trade press, professional associations, NHS guidance for health brands, accreditation bodies in your sector.
Brand longevity
The longer you've been around and the more consistent your presence, the more likely AI is to consider you a safe bet. This isn't just about age, it's about the depth and density of your digital history. A brand that's been written about, reviewed, and discussed for ten years has a richer evidence base than a newer entrant, even if the newer brand's products are objectively better.
Community discussion
Conversations on Reddit, forums, and social platforms where real people talk about your brand in an unsponsored, unprompted way carry genuine weight. This kind of discussion is low in persuasive intent, which is exactly what AI platforms find trustworthy. Someone asking "is Gymshark worth it?" on Reddit and getting 47 positive responses is a powerful signal.
Editorial coverage
Press coverage, magazine features, and independent editorial reviews do contribute, but less than most brands expect, and significantly less than retailer reviews. Where editorial content does help is in establishing category relevance and topical authority over time.
What Brands Definitely Should Know
Absence is riskier than bad reviews
This is the finding that most surprises brands. If AI can't find much about you, it doesn't quietly exclude you from its answers. It fills the gap with whatever weaker evidence it can find, and makes a recommendation anyway, often a confident one that may not reflect your product fairly. A handful of mediocre reviews is actually a better position than no reviews at all.
Different AI platforms behave very differently
Claude, GPT-4o, and Perplexity don't all give the same answers, and the reasons for their differences matter. Perplexity is search-grounded, meaning it draws on live web results rather than just its training data. This creates stark gaps for some brands. In the activewear research, one major brand dropped 72% in mentions between Claude and Perplexity, strong historical brand authority hadn't translated into a strong current web presence, and Perplexity found them out.
Gymshark, by contrast, posted near-identical scores across all three platforms. That kind of consistency is the clearest signal that a brand's digital presence is genuinely robust and not reliant on reputation built years ago.
The window is open, but it won't be for long
AI models are updated continuously. A brand that builds the right kind of digital authority now can meaningfully shift its visibility over a six to twelve month period. But every week of inaction is a week that dominant brands extend their lead and compound their advantage. The brands that are visible today didn't necessarily have the best products. They had the best-represented ones.
Your visibility score varies by what people are asking
Different query types produce very different visibility results. Purchase-intent queries ("where should I buy activewear?"), activity-specific queries ("best leggings for running"), price-point queries ("affordable gym wear UK"), each surfaces brands differently. A brand might be well-represented in purchase-intent queries but almost invisible when someone asks about sustainable sportswear. Understanding where you show up and where you don't is the starting point for any meaningful AI visibility strategy.
What Brands Should Stop Doing
Assuming your marketing content is working harder than it is
Brand-owned content scores just 8 out of 100 in the signal hierarchy. Your website copy, your social media posts, your press releases, they explain who you are, but they don't recommend you. They support, they don't persuade. AI platforms are specifically designed to treat brand-owned content with scepticism, because the intent to persuade is obvious. Stop expecting your content to do the heavy lifting in AI discovery.
Treating reviews as a hygiene metric
Review volume, recency, and consistency aren't nice-to-haves. They are the primary infrastructure of your AI visibility. A brand that doesn't actively manage its review environment across retailer platforms is essentially leaving its AI visibility to chance, and based on the research, that chance is not good.
Focusing exclusively on Google
Traditional SEO still matters, but AI discovery is a separate game with separate rules. The brands winning on Google are not automatically the brands winning in AI recommendations. Your strategy needs to account for both, and increasingly, AI visibility needs its own dedicated attention.
Assuming visibility on one platform means visibility everywhere
A strong showing on ChatGPT doesn't mean you're doing equally well on Claude or Perplexity. Given that Perplexity is growing in consumer adoption and its search-grounded model makes it sensitive to current web presence rather than historical authority, brands that rely on legacy reputation without maintaining fresh, consistent output will increasingly find themselves falling behind on the platform most likely to win new users.
Frequently Asked Questions About AI Visibility
What is AI visibility and why does it matter?
AI visibility is a measure of how often and how prominently your brand appears in AI-generated responses to consumer queries. It matters because more people than ever are using AI tools like ChatGPT, Perplexity, and Claude to research products before they buy. If you're not visible in those responses, you're missing a growing share of the consideration set.
How do I know if my brand is being mentioned by AI?
You'd need to run a systematic test, querying multiple AI platforms across a range of prompts relevant to your category, without seeding your brand name, and recording the responses. This is what an AI visibility audit does. The results are often very different from what brands expect.
Can I pay to appear in AI recommendations?
No. AI recommendations are not influenced by paid advertising in the way that search results are. The signals that matter are organic, reviews, editorial coverage, community discussion, authority references. There is no shortcut that replicates the credibility of genuine post-purchase feedback at scale.
Does having a good website help with AI visibility?
A well-structured website contributes marginally, but it is one of the lowest-impact signals in the hierarchy. What matters far more is your presence and reputation across third-party platforms (particularly retailer review environments) rather than the quality of your own site content.
How quickly can AI visibility be improved?
Research suggests meaningful improvement is achievable over six to twelve months with a focused approach. The key inputs are review volume and recency, editorial coverage, and community presence. These compound over time, which is why starting early gives a significant advantage.
Does AI visibility differ from SEO?
Yes, significantly. SEO is primarily about how well your content ranks in Google's search results. AI visibility is about whether and how your brand appears in AI-generated answers. Some of the inputs overlap (editorial coverage, for example) but the weighting is very different. Retailer reviews, which are relatively minor in traditional SEO, are the dominant signal in AI visibility.
The bottom lineThe brands that AI recommends aren't necessarily better. They're better evidenced. Their customers talk about them, review them, discuss them, and reference them in ways that give AI something credible to work with. If your brand isn't building that kind of presence actively, it is being evaluated passively, using whatever evidence happens to exist, for better or worse.
