Why brand matters more than ever in AI search

Brand has never mattered more to how you get found.

That sounds like the sort of thing an agency would say, so here's the number behind it. When AI search answers a question about your category, your own website supplies somewhere between 5 and 10% of the sources it reads. That figure comes from McKinsey's research on AI search.

The other 90% comes from places you don't control. Review platforms. Reddit threads. Comparison pieces. Journalists. Other entities talking about your brand.

AI engines form their view of your company much the way a stranger at an industry dinner would: mostly from what everyone else says.

You can publish flawless content on a beautifully structured site and still lose the recommendation to a competitor with a louder reputation. However good your content is, it can only ever be part of the picture.

The rest is your reputation, described and corroborated across the web, which is another way of saying your brand.

Over the past year, research from McKinsey, Edelman, Profound and Semrush has kept arriving at that same conclusion from different directions.

Brand, the slow, hard-to-measure asset that two decades of performance marketing taught us to put off, turns out to be the thing AI search runs on.

Here's what the numbers show, and what they mean if you're running marketing for a small or mid-sized business.

So what do we mean by brand?

It's worth pausing on the word, because brand means something different to everyone in the room.

Your logo, your palette and your tone of voice document all matter, mostly to the humans who come across you. AI engines never see any of it. What they work with is everything written about your company across the web, and the picture that forms when it's all read together.

There's an old line from Jeff Bezos about brand being what people say about you when you're not in the room. AI answers are quite literally the room you're not in. Someone asks a question about your category, an engine assembles a reply from what everyone else has published, and you find out afterwards whether you were mentioned and how you were described.

In practice that comes down to five things, all of which you can go and check today.

Recognition. Does the engine know your company exists as its own distinct thing, separate from similarly named businesses and from the generic category?

Description. When it does mention you, does it describe what you do accurately and consistently, or does it hedge and generalise?

Association. Which problems, categories and competitors do engines connect you to? This is what decides whether you show up for the questions your buyers actually ask, rather than only for your own name.

Corroboration. How many independent sources say roughly the same thing about you? A claim on your own website carries far less weight than the same claim echoed across reviews, press coverage and community discussion.

Sentiment. When you are discussed, is it warmly, neutrally, or with a caveat attached? Engines pick up qualifiers like "expensive", "hard to set up" or "great for small teams" and pass them straight into the answer.

None of that is abstract. All five turn up in AI answers as words on a screen, which means they can be read, tracked and worked on. That's the version of brand this piece is about, and it's a good deal more measurable than brand has ever been before.

The decisions have already moved

If this sounds like a problem for next year, McKinsey's numbers suggest otherwise. Half of consumers already use AI-powered search, and 44% call it their preferred way to research, ahead of traditional search at 31%.

McKinsey expects £560 billion of US consumer spending to flow through AI search by 2028, and warns that traditional search could lose between 20 and 50% of its traffic as decisions move upstream into the answer itself.

What's striking is the measurement gap. Only 16% of brands systematically track how they appear in AI search. Buyers have moved. Most marketing dashboards haven't caught up yet.

AI engines reach for brands they recognise

Traditional search ranked pages. AI search reasons about entities: the companies, products and people it has learned about from everything it has read. When an engine answers a commercial question, it reaches for the names it trusts.

Scale clearly helps. Semrush's 2026 AI Visibility Index, built on 126 million prompts, found that only 36 brands worldwide held top-100 visibility across ChatGPT, Gemini, AI Mode and AI Overviews at once.

McKinsey adds an honest complication, though. Even leading brands are missing from some answers in their own categories.

Existing power and weight in ad spend, market share or shelf presence, doesn't transfer to visibility automatically. What transfers is the version of your brand the machines can read: how often you're mentioned, how consistently you're described, and how strongly the sources engines rely on corroborate your story.

Being cited and being recommended are two different things

Semrush's ghost citations study offers the cleanest evidence that authority and brand are different currencies.

Across the domains studied, 62% of AI citations never produced a brand mention in the answer. Your page can do the work while somebody else's name gets the credit.

The pattern splits in two. The strongest brands in the study were named two to three times more often than they were cited.

Aggregator platforms like Medium were the mirror image: cited constantly, never named. Citations reflect domain authority. Mentions reflect brand trust.

Readers only ever act on the mention, so if you chase citations alone, you're optimising for the part of the answer nobody notices. There's more on this distinction in our guide to mentions vs citations in AI search.

Your brand is built on domains you don't own

Profound's analysis of 680 million citations shows where engines actually go to form their view, and each one reads a different web. Perplexity leans on Reddit for nearly half of its top sources.

Google's AI Overviews spread across Reddit, YouTube and Quora. ChatGPT favours established reference and news sites. A brand can be well represented in one engine's reading list and invisible in another's, and the places that matter most are the ones your buyers already use to talk shop.

We've mapped who reads what in our guide to the source ecosystem in AI search.

For B2B teams there's a sharper finding still. Profound's study of 1.4 million citations for professional queries found LinkedIn had become the most-cited domain in AI search, with posts and articles growing from 26.9% to 34.9% of LinkedIn citations in three months, overtaking company profiles as the biggest slice.

Every colleague who writes on LinkedIn is now part of your brand's citation footprint.

Founder and expert content has quietly become source material.

Brand trust is still what wins the deal

None of this works without the human layer underneath, and Edelman's 2026 Trust Barometer special report on brands measures it.

Across 17,688 people in 15 countries, 88% said trusting the brand is an important or critical purchase criterion, level with quality at 89% and value at 88%.

The voices that build that trust are earned ones: customers and peers carry more weight than anything a brand says about itself.

Look at what AI engines read and you'll recognise the same earned layer Edelman describes. Reviews, communities, independent coverage, peer discussion.

Machine trust is human trust, aggregated and compressed.

As AI assistants begin shortlisting suppliers and completing purchases on people's behalf, that compressed trust quietly decides which brands are ever considered.

Why brand advantage compounds

Three dynamics turn today's visibility gap into a lasting advantage for whoever holds it. Zero-click behaviour keeps growing, so the brand impression inside the answer is increasingly the entire touchpoint.

Agentic search narrows a whole market to one or two picks, made by a model that won't scroll to page two on your behalf.

And the loop feeds itself: AI recommendations create branded demand, branded demand creates more coverage and discussion, and that coverage becomes the next round of training data.

Every quarter spent absent from the answer makes the leader's head start a little harder to close.

Smaller brands can still win this

The same research shows the field is far from settled, which is worth saying plainly. Semrush found concentration varies enormously by category: the top three brands hold 82.9% of AI visibility in news and media, but just 41.4% in finance.

And McKinsey's finding that market leaders go missing from answers works in both directions. If fame doesn't transfer automatically, a focused challenger can overtake a complacent incumbent.

For most small and mid-sized companies, the realistic aim is to become the clearest, most consistently described answer to one specific problem, rather than to chase fame.

How to build a brand machines can read

The practical work sits closer to brand and PR than to traditional SEO, with some technical hygiene underneath. A few places to start.

Earn mentions where each engine reads. Digital PR and expert commentary in trusted publications, a presence in the Reddit and community conversations your buyers already have, and, for B2B, a steady LinkedIn writing habit across the team as well as the company page.

Keep your brand story consistent. The same name and the same one-line description of what you do, everywhere it appears: your site, LinkedIn, directories and review platforms. Inconsistency makes models hedge, and a hedging model will leave you out.

Keep owned content useful and structured. Your site may be a small share of the source mix, but it anchors how engines verify claims, and clear answers, FAQs and original data are what gets retrieved.

Measure mentions, sentiment and share of voice, prompt by prompt, alongside citations and rankings. Ghost citations mean citation counts on their own will flatter you; it's why we measure AI visibility this way at Tilio.

And give the work an owner. Only 16% of brands track any of this systematically, which is exactly why the ones that do are quietly building a lead.

Brand used to be the investment you made once the performance channels were exhausted. In AI search, brand is the algorithm. The teams who treat it that way now will be the names engines reach for by default in two years' time.

Frequently asked questions

Why does brand matter so much in AI search?

Because AI engines describe your company using the whole web rather than your website. McKinsey found owned sites make up only 5 to 10% of the sources AI search reads, so how widely and consistently your brand is discussed elsewhere largely determines whether you get recommended.

Is brand more important than backlinks for AI search?

Increasingly, yes. Backlinks still correlate with being cited, but the 2026 research shows earned brand mentions across trusted sources do more to get you named and recommended in AI answers. Citations reflect authority; mentions reflect the brand trust readers act on.

What is a ghost citation?

A ghost citation happens when an AI engine uses your website as a source but never mentions your brand in the answer itself. Semrush found this in 62% of AI citations, so citation tracking on its own significantly overstates real visibility.

How do I get my brand mentioned by ChatGPT?

Earn coverage in the trade press and industry publications ChatGPT treats as authoritative, describe your brand consistently everywhere it appears, and show up where your buyers already talk: review platforms, relevant Reddit communities and LinkedIn, now the most-cited domain for professional queries.

Does traditional SEO still matter for AEO?

Yes, and the two work best together. Organisations that fully integrated SEO and AI visibility strategies reported increased traffic or leads from AI 81% of the time, against 36% for teams running them separately. Crawlable, structured, authoritative content underpins both.

How long does it take to build brand visibility in AI search?

Expect months rather than weeks. Mentions driven by retrieval can move within weeks of earning coverage in trusted sources, but training-data effects build over quarters. The honest answer is to start measuring now, because the gap to category leaders widens the longer you wait.

How do I find out how visible my brand is in AI search right now?

Start with a snapshot rather than guesswork. Our AI visibility audit benchmarks how ChatGPT, Claude, Perplexity and Google's AI features mention, cite and describe your brand against your competitors, then gives you a prioritised list of what to fix first. For ongoing support, Tilio has monthly AI Search Visibility plans.