AI Overviews are Google's AI-generated answers that appear above the traditional results on a search engine results page. Google assembles them by retrieving passages from multiple pages, synthesizing them into a short answer, and citing a handful of the sources alongside. They rolled out to US search in 2024 and have expanded steadily in coverage and query types since.
The mechanism matters because it determines what a brand can influence. Google runs its own retrieval against the index, selects passages it judges responsive to the query, and generates prose from them. The cited sources are not simply the top organic results — pages ranking outside the top ten are regularly cited, and pages ranking first are regularly omitted.
The practical consequence: an AI Overview citation is a separate outcome from an organic ranking, won or lost on whether a specific passage answers a specific question cleanly, not on whole-page authority alone.
Not all lost clicks cost the same. Definitional traffic often converts poorly, so losing it may cost close to nothing. The expensive error is assuming the same about visits that fed email capture, retargeting pools, and eventual orders. Price the loss before reacting to it. If a page sent 1,200 visits a month and those visits now have to be bought back through paid search at an assumed $1.50 a click, that is $1,800 a month added to customer acquisition cost, for traffic that may never have been worth that much. Both figures are illustrative inputs, not measured ones. The order of operations is fixed: pull revenue and assisted conversions for the affected query set first, then decide. Where the traffic never earned, accept the loss and stop paying anyone to rewrite the page. Where it did earn, the replacement is a real budget line.
The direction is not in dispute: when an AI Overview fully answers a query, click-through to the results below falls, often sharply. The effect concentrates on informational and definitional queries — the ones a short synthesized paragraph can satisfy completely.
It is far weaker where the answer cannot be finished in a paragraph: transactional queries, queries needing a specific product or price, comparison shopping, local intent, and anything where the user needs to do something rather than know something. For an ecommerce brand this is the useful split. Definitional content loses clicks to AI Overviews; product, category and commercial-intent pages largely do not.
Search Console does not label AI Overview impressions separately, so the effect has to be inferred. The pattern to look for is impressions holding steady or rising while clicks fall on the same query set — particularly on definitional and "what is" queries. A page sitting at position three with a click-through rate far below what that position should produce is the classic signature.
The mistake worth avoiding is reading that signature as a content-quality failure and rewriting a page that was never going to earn the click. The right response is usually to accept the loss on the definitional query and shift the page's job toward the questions an overview cannot finish.
Structuring content so it survives — and ideally gets cited — is the substance of answer engine optimization, and part of the wider organic growth picture now that a meaningful share of informational demand never reaches a blue link.
Because a citation is won at the passage level rather than the page level, the practical question most brands face is which existing content is worth restructuring for citability and which definitional pages are better left to lose the click — a distinction worked through in answer engine optimization: how brands get cited by ChatGPT, Perplexity, and Google AI Overviews.
.png)
.png)
