Ranking #1 on Google used to be the finish line. Now there's a second race: getting your page picked up as a source inside ChatGPT answers, Google AI Overviews, and Perplexity results. Miss that race and you can rank well and still lose the click, because the AI answer already gave the user what they needed — with someone else's brand attached.

This is generative engine optimization (GEO), and it borrows from SEO without being identical to it.

Why AI answer engines cite what they cite

AI Overviews and chat-based answer engines work by retrieving a set of candidate pages, extracting passages that answer the query, and synthesizing a response with citations. Three things matter more here than in classic SEO:

  • Extractable answers. A model can lift a clean, self-contained paragraph far more easily than it can summarize a page that buries its point three scroll-lengths down.
  • Structural clarity. Headings that match real sub-questions, short paragraphs, and lists make it easy for a model to isolate a passage and attribute it correctly.
  • Corroboration. If three independent, credible sources say the same thing, a model is more likely to treat that claim as citable fact. Being the only source saying something unusual makes you less likely to get picked up, not more.

Practical changes that help

Answer the question in the first two sentences of a section. Don't build up to the answer — state it, then support it. Models weight the direct answer near the heading more heavily than context buried afterward.

Use real headings for real sub-questions. "What is X", "How does X work", "X vs Y" as literal H2s and H3s make your content easier to chunk and retrieve. Vague headings like "Let's dive in" give the model nothing to match against a query.

Keep facts self-contained. A stat, a definition, or a step should make sense pulled out of context, because that's exactly how it will be used. If a sentence only makes sense after three paragraphs of setup, it won't survive extraction.

Maintain crawlability for AI crawlers specifically. Google's AI Overviews use the standard Googlebot index, but ChatGPT search and Perplexity have their own crawlers (GPTBot, PerplexityBot, and others). Check your robots.txt isn't blocking them if you want to be eligible for citation there.

Add structured data. Article, FAQ, and HowTo schema don't guarantee citation, but they give the model an unambiguous, machine-readable version of your content to lean on alongside the prose.

Publish primary information, not just synthesis. Original data, first-hand testing, and direct quotes are more citable than another summary of what's already published elsewhere. If your page is the fifth rewrite of the same five facts, there's no reason for the model to pick you over the other four.

What doesn't work

Keyword-stuffing headings for AI doesn't move the needle — models aren't matching strings, they're matching meaning. Neither does adding an FAQ section that repeats the main content in question form; if it's not adding a distinct, extractable answer, it's noise.

The honest tradeoff

GEO and SEO pull in the same direction more often than not: clear structure, direct answers, and real expertise help with both. The difference is that AI Overviews can answer a query completely enough that the user never clicks through, even when you're the cited source. Some traffic loss on informational queries is a structural shift in how search works, not a problem you can fully optimize away. The response isn't to abandon informational content — it's to make sure the content that still drives clicks (comparisons, tools, transactional pages) is where you invest the deeper resources, and treat purely informational pages as brand-visibility plays rather than pure traffic plays.