Okay, we need to talk about a number that should be causing significantly more panic than it currently is.

A new study analysed 12,000 AI responses across Google AI Mode, AI Overviews, Gemini, and GPT — and buried in the data is a finding that basically invalidates about 80% of the "AI SEO" advice currently circulating on LinkedIn. (Yes, including some of mine. Growth is painful.)

Here it is: when someone asks an AI a category-level question — "what's the best project management tool?" or "which CRM should I use for a small law firm?" — only 7.2% of brand mentions in those responses actually come with a citation. When someone asks about a specific brand by name? That jumps to 39%.

That's a fivefold gap. And it tells a story the industry has been circling for eighteen months without quite managing to land on: brand awareness isn't the fluffy upstream metric that sits politely above the real work. In AI search, brand awareness is the real work.

The study — conducted by Authority Hacker using Xofu's tracking platform — looked at roughly 200 brands across ten industries and three prompt types. Its central finding is one of those things that sounds simple until you sit with it for a minute and feel your strategy deck quietly disintegrate.

AI systems mention brands far more often than they cite those brands' content. Only 23.1% of brand mentions carry a citation. But flip it around, and 69.9% of citations also include the brand name in the response text.

What this means in practice is properly uncomfortable. Your brand can be mentioned constantly in AI responses while receiving approximately zero referral traffic, because the AI is pulling your name from its training data — the accumulated weight of every mention, review, and Reddit thread about you across the internet — rather than actually visiting your website and linking to it.

It's the AI equivalent of someone name-dropping you at every party in town but never actually introducing you to anyone. Flattering, technically. Useless, practically.

Conversely, your content can be cited as a source without your brand being mentioned at all. Which gives you a link but no recognition. You're essentially ghost-writing your own marketing. Congratulations, I think?

This isn't a measurement quirk. It's a structural feature of how AI search actually works. And it has serious implications for where marketing budgets should be going.

Here's where the data gets properly interesting — and by "interesting" I mean the kind of interesting that makes you want to quietly close your laptop, take a long walk, and reconsider every budget allocation you've made this year.

The study broke responses down by three prompt types, and the results diverge so dramatically they're essentially describing different systems.

Branded queries ("Is Asana good?" / "Tell me about HubSpot"): 39% of mentions carry a citation. The AI knows what you want, finds the relevant content, and serves up a coherent response connecting the brand to its source material. Everyone's happy. Well done. Gold star.

Head-to-head comparisons ("Asana vs Monday" / "Mailchimp vs ConvertKit"): Citation rate drops to 35.8%. The AI starts pulling in comparison articles, review sites, and third-party analysis, which dilutes the direct brand-to-citation connection. Still workable, but messier. Like a group project where everyone does fine but nobody gets individual credit.

Open-ended category queries ("best project management tools" / "which accounting software for freelancers?"): 7.2%. Seven point two percent. The AI is essentially free-styling, constructing responses from its aggregated knowledge rather than actively retrieving and citing sources. These are the queries where AI systems are most autonomous — and they're the exact queries that matter most for reaching people who don't already know you exist.

The uncomfortable implication: if you've been treating AI search optimisation as purely a content and technical discipline — structured data, schema markup, citation-optimised articles — you've been solving the wrong problem. Or more precisely, you've been solving the right problem for people who already know your name, while everyone else gets served a response assembled from the AI's general vibes about your category. And your carefully optimised FAQ page has approximately no influence on those vibes.

The Xofu study isn't operating in isolation here. Multiple research streams are converging on the same conclusion, which is either reassuring or alarming depending on your current brand awareness strategy. (If you don't have a current brand awareness strategy, it's alarming. Just saving you time.)

Amsive found that branded queries with AI Overviews see an 18% increase in click-through rate, while generic queries experience a 34–46% decline. Seer Interactive's research across 53 brands showed that brands cited within AI Overviews earn 35% more organic clicks and 91% more paid clicks than those that aren't. The AirOps report found that brands earning both mentions and citations are 40% more likely to resurface across consecutive queries.

Every single one of these findings points in the same direction: the brands winning in AI search are the ones people already know. When someone types your brand name into a prompt, the AI activates a different retrieval pathway — one that connects entity recognition with source citation, produces coherent responses, and actually drives clicks. When someone asks a generic category question, the AI is making autonomous choices based on the cumulative weight of your existence across the internet.

Which means that boring, expensive, notoriously-hard-to-measure work of making people aware you exist? That's not the fuzzy top-of-funnel stuff you do when you've got spare budget and an optimistic CFO. That's the foundation everything else depends on.

(I know. I'm sorry. I wanted it to be about schema markup too.)

Here's the finding that should quietly terrify every content marketing team currently running a "publish and pray" strategy: 85% of brand mentions in AI responses originate from third-party pages, not brand-owned domains.

The AI doesn't primarily form its opinion of you from your own website. It forms its opinion from review articles, comparison pieces, journalist coverage, industry analysis, expert roundups — the ecosystem of content that other people create about you. Your blog post is one signal. The fact that an industry publication mentioned you in a comparison guide, a consultant referenced you in a LinkedIn post, and three Reddit threads argued about whether you're better than your competitor — that's the signal that actually builds entity recognition in the model.

This doesn't mean owned content is irrelevant. When the AI does decide to cite a source, your well-structured, clearly written, genuinely useful content is what it links to. But it has to decide to mention you first. And that decision is driven by the broader web's consensus about whether you're a legitimate player in your category.

The sequence matters: awareness first, then citation. Not the other way around. It's like being invited to a job interview — your CV might get you the offer, but the referral gets you in the room. No referral, no room, no matter how immaculate the CV.

Three things fall out of this data, and they apply whether you're running a SaaS company, a professional services firm, a local business, or anything else where AI-mediated discovery is becoming relevant (which is, increasingly, everything).

First: brand and performance aren't separate budget lines anymore. The conventional split between "brand awareness" and "performance marketing" has always been somewhat artificial, but in AI search it becomes actively counterproductive. Brand awareness is the upstream condition that determines whether your AI search optimisation works at all. All the structured data and citation-optimised content in the world won't compensate for a weak entity signal when someone asks a category question. You need to be known before you can be found.

Second: you need to track mentions and citations separately. Most teams are monitoring one or the other — checking whether their domain appears in AI citations, or whether their brand name shows up in responses. The data makes clear these are two different signals driven by two different mechanisms. Mentioned but not cited? You've got a content retrievability problem. Cited but not mentioned? That's a brand awareness problem. Each diagnosis requires a different intervention, and treating them as interchangeable produces a distorted picture of your actual AI visibility.

Third: this gap is going to widen, not narrow. As AI systems become more autonomous — as agents start mediating decisions, as conversational search replaces keyword search, as AI Mode handles an increasing share of discovery — the advantage compounds for brands that exist as strong entities in the model's understanding of the world. A brand consistently cited in comparison articles, frequently mentioned in expert coverage, and well-represented across review ecosystems will surface even when nobody asks for it by name. A brand that exists primarily as a well-optimised website will not.

The data from 12,000 AI responses is saying something the industry doesn't particularly want to hear: brand building — that slow, expensive, notoriously difficult-to-justify work of making people know you exist and associate you with a specific category — isn't the nice-to-have that lives in the "we'll get to that next quarter" column of every marketing plan ever written.

It's the single most important input into whether AI systems mention you, cite you, and recommend you to the growing number of people who now ask a chatbot before they ask Google.

The brands that figured this out eighteen months ago are the ones showing up in AI responses today. The brands that figure it out now still have time. But the compounding advantage of early investment means the window won't stay open forever.

And no, you can't just schema-markup your way out of this one. I checked.

Jo Lambadjieva is an entrepreneur and AI expert in the e-commerce industry. She is the founder and CEO of Amazing Wave, an agency specializing in AI-driven solutions for e-commerce businesses. With over 13 years of experience in digital marketing, agency work, and e-commerce, Joanna has established herself as a thought leader in integrating AI technologies for business growth.

😍It was great, all I needed

🙂Good, not great

🙄It sucked