What changed with AI Overviews and answer engines?
Search moved from a ranked list of links to a synthesized answer with a small set of cited sources. The answer now sits above organic results — and the choice of which sources to cite is made by a model, not by a SERP algorithm tuned only for clicks.
For most informational queries the user no longer needs to click. They get an answer, optionally with two to six citations. If you are one of those citations, you keep mindshare, brand recall and a fraction of the traffic. If you are not, you lose all three — regardless of where you would have ranked organically.
This is the first structural change to search since the move from desktop to mobile. The companies that adapt early will set the defaults for their category.
What models actually reward
- • Entity clarity: the model can resolve who you are without ambiguity.
- • Answer-shaped content: a one-paragraph direct answer it can quote verbatim.
- • Structured data: JSON-LD that mirrors visible content.
- • Retrievable chunks: short, self-contained passages models can extract.
- • External corroboration: mentions on sources the model already trusts.
- • Machine-readable surface: llms.txt, predictable URLs, clean sitemaps.
What this does not mean
It does not mean SEO is dead. It means the surface area expanded. Classic ranking still feeds retrieval. Technical hygiene still matters. The change is that being technically excellent is now table stakes — the new advantage is being citable.