A featured snippet quotes one page above the search results. An AI Overview writes its own answer from several pages, and it has taken the position the snippet used to hold. If your traffic changed over the last two years, this swap is a big part of why, and the work that used to win the old box is not quite the work that wins the new one.
What a featured snippet is
A featured snippet is an extract: Google lifts a paragraph, list or table verbatim from one ranking page and displays it above the organic results with a link. One source, quoted, attributed, clickable. Winning it meant structuring a page so your answer was the cleanest lift available, and for a decade it was the most valuable real estate in search. Snippets still appear, but for fewer queries every year, because their successor eats the same questions first.
What an AI Overview is
An AI Overview is a composition: Google's model reads a shortlist of pages and writes a new answer, citing some of its sources in a panel beside the text. Several sources, synthesised, selectively credited. The difference sounds technical and is commercial: a snippet guaranteed that being the source meant being seen, because the source was the content. An Overview can use your page's facts, credit you in a collapsed citation few people expand, and describe your category using three competitors' names.
What changed for your business
Three shifts follow. Position one organic now starts below an answer that may not mention you, so rankings overstate your visibility. Clicks fall on the queries Overviews answer fully, so traffic understates whether your content is working: your facts can be doing the persuading inside an answer that sends you nothing. And the contest moved from winning one box with one page to being a repeatedly cited source across a set of questions, which is a different discipline with a different scoreboard.
How to show up in AI Overviews
The good news: everything that won snippets still counts, because Overviews retrieve from the same index and prefer the same clean structure. Answer-first sections under plain headings, one idea per passage, specific claims that survive being lifted alone, real numbers where competitors have adjectives. What's new is everything around the page: a clear entity so the citation lands on your name, corroboration across the web so the model trusts the claim, and structured data labelling what's what. Snippet optimisation was a page-level game. Overview visibility is a business-level one.
One practical note on measurement: snippet wins were binary and easy to track. Overview presence is probabilistic, varies between runs, and needs a fixed question panel checked on a cadence to read honestly. We track both inside the Search and AI Visibility retainer, one report, $2,000 to $10,000+ a month, detailed in our cost guide.
FAQ
Did AI Overviews replace featured snippets?
Largely, for the question-style queries snippets used to answer. Both can still appear, but where an AI Overview shows, it takes the top position and the snippet's old job.
Is snippet optimisation still worth doing?
Yes, because the same structure that won snippets is what Overviews lift from: direct answers, clean headings, specific claims. The work transfers. The expectation that one box guarantees visibility doesn't.
Why did my traffic drop if my rankings held?
Very likely because answers above your ranking are satisfying the query before the click. The fix isn't recovering the old clicks. It's being named and cited inside the answers doing the satisfying.
