For fifteen years, the instruction to a client was simple: rank on Google, and the traffic follows. That instruction is no longer complete. A growing share of the people who would once have searched, clicked, and compared five results are now asking ChatGPT, Perplexity, or Gemini a question directly and acting on whatever answer comes back, without visiting a single website at all.

Ranking first on Google is still necessary. It is measurably no longer sufficient. A page can hold position one and still be absent from the answer an AI gives to the exact question that page was built to win.

Search used to end at a list of links. Increasingly, it ends at a single paragraph, and if your business is not the source of that paragraph, the ranking underneath it has stopped mattering to the person who asked.

Why the old playbook stops short

Traditional SEO optimises for a crawler that indexes a page and a ranking algorithm that orders it against competitors. Answer engines do something different. They read a page, extract a claim, and decide whether that claim is trustworthy enough to state as fact to a user who will never click through to verify it. The unit of competition has moved from the page to the sentence.

Most businesses have not updated their content for that shift. Their pages are written to persuade a visitor who arrives and reads the whole thing. They are not written to survive being extracted, summarised, and quoted out of context by a language model that has three seconds to decide what to say.

What actually earns a citation

Getting cited by an AI Overview or a Perplexity answer is not a matter of writing more content, or writing it more persuasively. It rewards a narrower, more mechanical kind of clarity, one that most content strategies were never built to produce.

The mechanics that matter

  • Structured data: schema markup that states plainly what an entity is, what it does, and how it relates to other entities, removing the guesswork a model would otherwise have to do
  • Entity consistency: the same name, description, and facts about the business repeated identically across the site, directories, and social profiles, so no model has to reconcile conflicting versions of who you are
  • Answer-shaped content: a direct, quotable answer to the specific question in the first two sentences of a section, with the supporting detail and nuance placed after it, not before
  • An llms.txt file: a plain-language map of the site written specifically for models rather than crawlers, pointing them toward the pages that answer the questions your business actually gets asked

None of these tactics work in isolation. A page with perfect schema and a buried, hedge-filled answer will still lose the citation to a competitor whose answer is stated plainly in the first sentence, however weaker their markup happens to be.

A model does not reward the most thorough page. It rewards the page it can quote without embarrassment. Clarity is no longer a style preference. It is the mechanism by which you get cited at all.

Where this leaves Google

None of this replaces conventional search work. Google still sends the majority of commercial traffic for most of the businesses we work with, and will continue to for some time. The two disciplines are not in competition with each other so much as increasingly the same discipline, viewed from two directions: one measured in rank, the other measured in whether a model is willing to put your business’s name in its answer.

A site with strong technical SEO and no entity clarity will keep ranking while slowly losing the answer layer above it. A site built only for answer visibility, with thin technical fundamentals, will struggle to earn the crawl depth and authority signals that answer engines still lean on to decide who to trust in the first place.

We treat this as one workstream now, not two competing budgets. Every new engagement gets structured data, entity consistency, and answer-shaped writing as a baseline, not an add-on billed separately once a client asks why a competitor started showing up in ChatGPT and they did not.

The businesses that treat this shift as a passing trend will keep optimising for a search result page an increasing share of their customers no longer visit. The ones paying attention are already being quoted by name, in answers their competitors never see, by an audience that never clicks through to notice who lost.