Key takeaways
- Citation share matters more than rank position for queries that trigger an AI answer, because most of those searches now end without a click.
- Self-contained paragraphs get quoted; narrative build-up does not. If a paragraph needs the one before it to make sense, an engine will skip it.
- Schema markup roughly doubled citation frequency across the pages we tested, by removing the need for an engine to infer what a page is about.
- Content locked behind client-side JavaScript is largely invisible — most AI crawlers do not execute JS the way Googlebot does.
For most of the last decade, the job was simple to describe: rank in the top three organic results and collect the click. That job has quietly changed. On a growing share of commercial queries, the answer now assembles above the results — and increasingly, outside Google altogether.
What actually changed
The shift is not that AI arrived. It is that the unit of visibility moved from the link to the citation. When an answer engine composes a response, it selects a handful of sources to draw from and names them. Being ranked fourth and being cited is now worth more than being ranked first and being ignored.
Across our accounts we started tracking citation share — the percentage of relevant AI answers that name a client as a source — alongside conventional rankings. The two correlate far less than you would expect. Pages that ranked well but read as marketing narrative rarely got cited. Pages that ranked mid-page but answered a question cleanly got cited often.
The content structures that get quoted
The pattern that emerged is unglamorous: engines quote paragraphs that make sense on their own. A paragraph that opens with "As we mentioned above" or "This is where it gets interesting" is unusable out of context, so it gets skipped regardless of how good the writing is.
Answer-first paragraphs
The highest-cited blocks in our data shared a shape: a direct claim in the first sentence, then two to four sentences of support. No throat-clearing, no "it depends" opener, and the subject named explicitly rather than referred to as "it" or "this".
If you cannot copy a paragraph out of your page, paste it into a document with no other context, and have it still read as a complete answer — an engine cannot quote it either.
Explicit structure over implied structure
Headings that state the question a section answers outperformed clever headings by a wide margin. So did comparison tables, which get pulled into AI Overviews far more often than equivalent information written as prose.
Why schema mattered more than expected
Structured data was the single highest-leverage technical change we made. Adding comprehensive schema — Article, FAQ, Organization, Breadcrumb — to pages that already ranked roughly doubled how often they appeared as a named source.
The mechanism is simple enough: schema removes guesswork. Without it, an engine has to infer what a page is, who wrote it, and whether it is authoritative. With it, those things are stated. Given a choice between a source it can verify and one it must infer, an engine takes the verifiable one.
The JavaScript problem nobody mentions
Googlebot renders JavaScript. Most AI crawlers do not, or do so inconsistently. We found several client sites where the majority of body content was injected client-side — visible to users, visible in Google, and effectively absent to every other engine we tested.
If your site is built on a framework, the fix is server-side rendering or static generation for anything that matters. This is a build decision, not an SEO one, which is exactly why it tends to get missed until traffic starts falling.
What we would do first
- Audit whether your key pages render their content without JavaScript enabled.
- Rewrite the opening paragraph of every important section to answer, not to set up.
- Add comprehensive schema markup, starting with your highest-traffic templates.
- Convert at least one prose comparison per page into an actual table.
- Start tracking citation share as a metric, not just rank position.
None of this replaces conventional SEO. Technical health, topical depth and authority still decide whether you are in the consideration set at all. What has changed is that being in the set is no longer sufficient — you also have to be quotable.
