From keywords to entities: Why SEO now depends on brand.

Google has always indexed pages. What’s changed is what it does with them.

The shift from keyword-matching to entity understanding didn’t happen overnight. It’s the cumulative result of years of algorithm development – Hummingbird, RankBrain, BERT – each one moving Google further from ‘does this page contain the right words?’ towards ‘does this brand deserve to rank for this topic?’. Three things did the heavy lifting:

• Semantic search. Google interprets meaning, intent, and context – not just exact-match terms.
• Knowledge graphs. Google maps entities – brands, people, topics – and the relationships between them. Being a recognised entity, not just a well-optimised page, shapes how and where you appear. In practice, this means showing up in AI Overviews, entity knowledge panels, and being cited in LLM responses – not just ranking on page one.
• LLMs and AI search. Large language models aggregate knowledge across sources and surface what’s already recognised. Isolated pages with no entity context barely register.

The practical effect is something SEOs have been noticing for a while. Across practitioner communities, the same pattern keeps coming up: brands with genuine discussion around them – reviews, forum mentions, community conversations – are outperforming technically superior sites. Better-optimised pages, weaker brands, losing ground. It’s consistent enough to be a signal in itself.

Google stopped reading your page. It started reading your brand.

Google’s quality systems no longer ask a technical question. They ask a brand question: who’s behind this content, are they credible, and do others reference them?

E-E-A-T (experience, expertise, authoritativeness, trustworthiness) is the framework Google uses to make that judgement. Its Search Quality Rater Guidelines reference E-E-A-T 121 times. And almost all of it maps to brand-building activity, not on-page optimisation. The data backs this up:

• Unlinked brand mentions now outweigh backlinks in AI search. Ahrefs found web brand mentions have a 0.664 correlation with AI Overview inclusion, against 0.218 for traditional backlinks. Who talks about you now matters more than who links to you.
• Branded search is a direct trust signal. A brand people search for is a brand Google trusts – and 45.7% of searches are already branded – while in the non-branded half, recognition still wins the click. Demand feeds straight into how search engines interpret relevance.
• Sites without brand authority are exposed. Analysis of Google’s 2024 Helpful Content Updates found the hardest-hit sites shared one trait: low brand authority. High domain authority without brand authority isn’t a strength. It’s a vulnerability.

Here’s the part I find most interesting as a practitioner: SEO now only works if you’re a genuinely good business. Search engines and LLMs are evaluating you as a brand – not testing whether you can play the game well. Which means a large part of an SEO’s results are now shaped by teams they don’t sit in. Customer service drives the reviews. PR earns the relevant mentions. Content demonstrates the E-E-A-T. Brand makes you the name people recognise and click. Social has become a search and discovery channel in its own right. Even your newsletters count – Google’s Personal Intelligence uses Gemini to analyse a user’s Gmail, so the brands landing in inboxes are the brands surfacing in AI Mode.

At a recent SEO panel at WordCamp Europe, Pam Aungst Cronin of Pam Ann Marketing put it plainly:

“Brand is the new backlink. That is what really you need to think about… branding is just such a bigger thing, and it’s about awareness building. That’s what you need to do to get the AIs to recommend your brand.”

Pam Aungst Cronin of Pam Ann Marketing

Traditional link-building got you authority. Now, authority comes from being a brand worth referencing.

Brand, PR, content – and teams well beyond marketing – aren’t supporting SEO. They’re generating the signals SEO depends on:

• Digital PR builds authority and mentions. A well-placed piece of coverage in a credible outlet does more for SEO than a manually built link ever could. It builds the entity graph around your brand – and contributes the unlinked mentions that now carry measurable weight in LLM outputs.
• Content builds topical depth and demonstrates E-E-A-T. Google rewards demonstrable expertise across a topic area, not keyword coverage – and content is where experience, expertise and named authors are shown. Generic AI content is already losing ground. Original, attributed insight is what holds.
• Brand builds recognition and demand. People click the names they recognise – in the SERP, in an AI answer, anywhere. And when people search for you by name, that’s the strongest trust signal there is.
• Social is a search channel in its own right. People discover and search on LinkedIn, Instagram and TikTok – and AI now powers the algorithms deciding those feeds. Social conversations and community presence feed the same entity signals search engines and LLMs draw on – and Google’s July 2026 rollout of Search Console platform properties, which reports on how Instagram, TikTok, X and YouTube content performs in Google Search and Discover, is a clear signal it now treats these platforms as part of Search itself.
• Customer service builds reviews and reputation. Strong reviews sit on exactly the platforms AI systems read, and they’re among the brand signals LLMs evaluate directly. A business that looks after its customers is, increasingly, a business that gets surfaced.
• Email is becoming personal context. In January 2026, Google brought Personal Intelligence to AI Mode – an opt-in feature where Gemini analyses a user’s Gmail and Photos to tailor responses. In early testing, brands present in a connected inbox appeared in 66.8% of relevant AI Mode responses, against 23.9% in a standard account. It’s an opt-in, US-only feature for now, but the direction is unmistakable: the brands in people’s inboxes are becoming the brands AI surfaces.

None of these are SEO outputs. They’re marketing outputs that SEO captures the benefit of. SEO no longer operates alone – that’s the integrated marketing argument in a single line.

AI search systems don’t rank pages – they synthesise answers from recognised sources. If your brand isn’t part of the entity graph, it barely exists in AI search.

The scale is already here. AI platforms now generate around 45 billion sessions a month worldwide, and 83% of that happens in mobile apps – so web-only measurement is dramatically undercounting adoption. Traditional search still dominates in volume – Google processes 15+ billion searches a day and holds roughly 90% of the market – but LLM-based search has doubled its share of US desktop search traffic in a year, reaching 5.6% by June 2025, with Semrush forecasts suggesting AI search visitors will surpass traditional search visitors in 2028.

And the economics are shifting. AI Overviews now appear on 58% of Google searches – and when they do, organic click-through drops by 61%. The flip side: AI-referred traffic converts at 14.2%, against 2.8% for traditional organic. Fewer clicks, far higher intent. The brands visible in AI answers capture more value per visit.

What earns that visibility? Mentions. Ahrefs’ research found brand mentions are the strongest off-site factor correlating with LLM citations. The brands surfaced in AI answers aren’t the best-optimised pages – they’re the most consistently referenced brands.

This isn’t speculation. In May 2026, Google published its own guide to optimising for generative AI features in Search – confirming that SEO best practice remains foundational, and that unique, non-commodity content is what gets surfaced. The message is unambiguous: the brands that earn visibility in AI search are the ones that built something genuinely worth referencing.

AI SEO isn’t a new channel. It’s the logical endpoint of entity-based SEO.

If you’re running SEO as a standalone channel, you’re already behind.

I’ve felt this shift directly – in how I build strategies, and in how we sell SEO as an agency. For SEO to deliver results now, it either needs to sit within a wider marketing strategy, or plug a specific gap for a business already doing the rest: PR, content, brand activity. SEO alone is a far harder route to results than it was even two years ago.

The foundations haven’t moved. I still start every strategy where I always have – technical health, metadata, schema, structured content. What’s changed is everything after. Optimisation no longer stops at your website: it extends to everywhere your audience searches and everywhere AI gathers its signals – LLMs, social platforms, review sites, not just Google. Content shifts from keyword-matching to conversations: the questions people actually ask, answered in language AI can extract and attribute. And expectations need resetting too – fewer clicks, visitors arriving from more channels, and search judged on the value it captures rather than the rankings it holds.

It’s a shift practitioners are naming out loud. As Jovana Smoljanovic Tucakov of Melograno Ventures told the same WordCamp Europe panel:

“In order to do good SEO, GEO or whatever and be included in AI generated answers, you need to look at SEO and marketing as a whole… you need to approach it as brand strategy, product marketing, SEO tactics we were already using, but just upgraded on a much higher level.”

Jovana Smoljanovic Tucakov, Melograno Ventures

Stop asking ‘what’s the SEO strategy?’. Start asking ‘what’s the marketing strategy – and how does search capture it?’.

We’ve unpacked all of this in more depth – including real examples – in our SEO in an AI World ebook, and you can watch our talk on the topic.

SEO performance is increasingly a byproduct of brand presence, authority, and recognition. That’s not a threat to SEO – it’s a clarification of what SEO actually depends on. The brands that understand this aren’t asking whether to invest in brand. They’re asking whether their brand is strong enough to compete in search.

And the window to build that advantage is narrowing. Entity authority compounds – brands consistently cited, mentioned, and recognised across the web today will be disproportionately hard to displace as AI search grows. Waiting until AI search is dominant means catching up against competitors who’ve been building for years.

There’s a structural question for agencies and in-house teams in this too. SEO delivered in isolation, disconnected from brand, PR, and wider marketing activity, is a harder and harder route to results. That’s not a criticism of SEO. It’s an argument for integrating it properly. The teams that figure that out now won’t just do better search. They’ll do better marketing.

That’s how we build search strategies at Invanity – as part of brand marketing and performance marketing, never in isolation. If you want a search strategy built for how Google – and AI – actually evaluate you, we’ll go after it with you.