By the end of this year, traditional search engine volume is projected to fall by roughly 25% as generative AI platforms intercept queries that once went straight to Google, according to Gartner data cited by Elementor.
The shift is already visible in the numbers.
AI search engines like AI Overviews now appear in close to 48% of tracked Google searches, and Microsoft has reported that AI referrals to top websites jumped 357% year-over-year in June 2025, reaching 1.13 billion visits.
For brands, creators, and marketers, understanding how these systems actually choose what to cite has become as important as classic SEO once was.
The good news is that the fundamentals haven’t been thrown out. They have been layered with new technical and structural requirements. In this guide by Brand Mender, you will understand what really matters in digital content, how content is getting ranked, and a bit of technical update your company must implement.
Feel free to book a no-obligation consultation with us if you want the best for your brand from SEO and GEO.
What is E-E-A-T
E-E-A-T is Google’s framework for writing content that will get ranked on AI models and the traditional search results.
Google has been explicit that how content is produced matters far less than whether it’s genuinely useful. In its official guidance, Google states that its ranking systems reward original, high-quality content demonstrating E-E-A-T:
- Experience
- Expertise
- Authoritativeness
- Trustworthiness
This is regardless of whether AI played a role in creating it.
Automation has powered and influenced SEO ranking and helpful content for years, from sports scores to weather updates, and Google draws a clear line – using AI to manipulate rankings violates its spam policies, but using AI to help produce genuinely helpful content does not.
That distinction carries directly into generative search. Because large language models are prone to inventing facts, they lean on trust signals to decide which sources are safe to cite.
E-E-A-T as AI ranking factors effectively functions as an AI’s misinformation filter, and it operates at the author level as much as the brand level — meaning a detailed, credentialed author bio can matter as much as the article itself.
How AI Selects and Cites Information
Modern AI search engines, from Google’s AI Overviews to Copilot and ChatGPT, are largely powered by retrieval-augmented generation (RAG).
Meaning:
Rather than answering purely from static training data, these systems search an index or vector database for relevant passages, feed those passages back into the model, and generate a response grounded in that retrieved content.
This is also why they can cite sources. This means content has to win two separate contests.
First,
It must first be technically easy to retrieve.
Second,
Be trustworthy enough to be selected for the final answer.
Microsoft describes this parsing process plainly:
- Assistants like Copilot break pages into smaller, structured pieces, evaluate each for authority and relevance, and assemble the strongest fragments into a single answer.
That’s why Microsoft recommends clear, descriptive title tags, H1, and headings that function like chapter markers, so that AI systems can isolate a complete idea from the surrounding page.
Structure Is No Longer Optional
Across all sources, one theme repeats – content built for skimming humans doesn’t automatically work for machines.
HubSpot’s research found that leading with a direct 1–3 sentence answer, in plain language, is the single most reliable tactic for earning citations, and recommends structuring sections around the natural-language questions people actually type into AI tools.
Microsoft suggests Q&A formatting, noting that assistants can often lift a well-phrased question-and-answer pair word for word into a generated response.
Every H2 and H3 section should function as a self-contained, citable answer rather than part of a sprawling narrative.
| Did you know? Both Microsoft and Elementor also caution against burying key facts in tables, PDFs, or image-only content, since AI parsers like ChatGPT and Claude read linearly and often can’t reconstruct information locked in those formats. |
Schema & Off-Site Authority Still Count
Structured data remains one of the clearest ways to remove ambiguity for machines. Microsoft notes that schema markup, delivered in JSON-LD, labels content as a product, review, FAQ, or event so systems can interpret it with confidence, while HubSpot points out that FAQ Page and Article schema help Google’s AI Overviews in particular, even though schema itself doesn’t directly move traditional rankings.
Authority built off the page matters just as much. Research suggests brands are roughly 6.5 times more likely to be cited by AI through third-party sources than through their own websites, which means guest contributions, expert quotes, consistent brand data across platforms, and genuine community presence all feed into whether a brand gets treated as a credible source.
This is often called a bifurcated reality – Google’s AI Overviews correlate strongly with traditional top-10 rankings, while standalone tools like ChatGPT and Perplexity lean more heavily on community and encyclopedic sources such as Reddit and Wikipedia, largely independent of Google rank.
Conclusion
Despite different architectures, Google, Bing, and the major AI search engines converge on the same underlying logic:
- Help real people first, make that help easy for a machine to find and extract, and back it up with verifiable expertise.
Brands that treat AI visibility as a new coat of paint on old keyword-stuffing tactics tend to be filtered out. Those that combine clear structure with genuine authority are the ones showing up inside the answer box.
Frequently Asked Questions
- Does using AI to write content hurt search rankings?No. Google has stated that AI-generated content is not against its guidelines unless it's created primarily to manipulate rankings. Quality and helpfulness matter more than the tool used to produce it.
- What is RAG, and why does it matter for SEO?Retrieval-augmented generation is the framework behind most AI search tools. It retrieves relevant content passages from an index, then generates an answer grounded in them, which is why passage-level structure matters more than whole-page optimisation.
- Is schema markup necessary for AI search visibility?It isn't a ranking factor on its own, but FAQPage, Article, and Organisation schema make content far easier for AI systems to parse accurately and are considered a strong supporting signal.
- Do Google AI Overviews and ChatGPT rank content the same way?Not exactly. AI Overviews correlate closely with traditional top-10 Google rankings, while tools like ChatGPT and Perplexity often cite sources with little connection to Google rank, favouring community and encyclopedic content instead.
- What's the fastest way to improve AI citation potential?Adding a concise, direct answer near the top of a page — typically one to three sentences — combined with clear headings and credible author information, tends to produce the quickest gains in citability.
