How to get cited by Google Gemini

In depth

The two engines have almost no shared plumbing. ChatGPT is a separate trained model with its own corpus — heavy on Reddit, news and long-form content. Gemini sits on top of Google's existing crawl and rank infrastructure, which means everything that affects your Google ranking also affects your Gemini citation rate.

The practical consequence: classical SEO is still the foundation for Gemini. Backlinks, page speed, mobile responsiveness, content quality — all still matter. The multiplier is structured data. A page with comprehensive schema markup (FAQ, HowTo, Organisation, Review) will out-cite a better-written page without schema.

That makes Gemini optimisation uniquely technical. The lever isn't 'publish more content' the way it is for ChatGPT — it's 'implement schema properly on the content you already have'. A mid-size site can get dramatic Gemini lift in 4–8 weeks by cleaning up structured data, while the equivalent ChatGPT lift would take 3–6 months of distributed content work.

In depth

AI Overviews don't invent their sources — they summarise pages that are already ranking. If you're not in the top 10 for the query, you're not in the AI Overview for it either. So traditional SEO remains the foundation: links, authority, intent-matched content.

Schema is the amplifier. Google's own documentation is explicit that structured data helps it 'understand content for rich results and AI features'. In our testing, pages with complete schema markup (multiple valid types, not just one) are selected for AI Overviews 2–3× more often than same-ranking pages without schema.

The final lever is structure. AI Overviews prefer content organised around explicit questions, with clean paragraph-level answers they can quote. A page with an FAQ section outperforms a wall-of-text page covering the same content. This is also why HowTo schema works so well — it's content in the exact shape AI Overviews want to extract.

In depth

Schema's impact varies sharply by engine. Google has been explicit for years that structured data feeds both traditional rich results and their AI summaries — and Gemini is directly downstream of that system. Sites with comprehensive, valid schema across FAQ, HowTo, Organisation, LocalBusiness and Product types see materially higher AI Overview citation rates in every category we've tested.

For the other engines, schema is quiet infrastructure. ChatGPT doesn't parse schema directly, but it does benefit from pages that are clearly structured (which schema tends to correlate with). Claude is similar. Perplexity sits in the middle — its ranker respects schema signals but weights them less than authority and freshness.

The practical programme is to invest in schema primarily for Gemini and let the other engines benefit as a side-effect. In most projects, getting schema right is a 2–4 week technical investment that keeps paying for years.

In depth

The 'ranks-but-isn't-cited' pattern is extremely common and almost always traces to content structure, not content quality. Google's AI Overview system needs to extract a clean, short answer it can surface with confidence. Pages that bury the answer in paragraph 6 of a long post get passed over in favour of pages that answer the question in the first 60 words.

The fix is mechanical. Break content into explicit question headers. Put a short, quotable answer immediately under each header. Follow with supporting detail. Add FAQ schema that mirrors those headers. This is the same structure that wins traditional featured snippets, and it works for AI Overviews for the same reason.

The second common issue is schema absence. If a competitor page at position 5 has complete FAQ and HowTo schema and you're at position 3 with no structured data, the AI Overview can easily pick them over you. Schema is the tiebreaker more often than most SEOs realise.

In depth

Google owns YouTube, and Gemini integrates video transcripts into its answer generation in a way no other engine does. For certain query categories — demos, tutorials, product walk-throughs, comparative reviews — the AI Overview frequently quotes a video, cites the creator and links to a specific timestamp.

The multiplier is transcript quality. YouTube's auto-captions are serviceable for English but miss jargon, numbers and brand names. Videos with manually-uploaded transcripts get cited at dramatically higher rates — our internal tests show 3–5× the citation frequency for the same content quality.

For categories where video is a natural medium — SaaS demos, professional services explainers, product unboxings, local service showcases — a basic video programme (one 4–8 minute video per month, with clean transcripts) pays off in Gemini citations within a quarter. For text-heavy categories (legal, finance writing, research) the lift is smaller but still non-trivial.

In depth

The AI-Overview-reduces-CTR complaint is real. Industry data shows 30–50% fewer click-throughs on top-10 results when an AI Overview sits above them — even when your page is one of the cited sources. That's a genuine cost.

But the alternative is worse. Not being in the AI Overview means you get the reduced CTR without the citation. The user reads the summary, sees competitor names, forms an opinion, and never even notices your blue link below. You lose both the click and the recall.

The brands winning this shift are treating AI Overview citation as a branding channel, not a traffic channel. A mention in an AI Overview for 'best X' is an impression — even if they don't click, they saw your name next to 'best'. That compounds into branded search ('your brand reviews', 'your brand vs competitor'), which still converts well. The total funnel math still favours being cited over being invisible.