Google put AI answers above search results in the US

The revenue mechanism

On 14 May 2024, at its annual developer conference, Google announced that AI Overviews, AI-generated summary answers shown above the traditional list of search results, would roll out to 'all users in the US' that week, with a stated ambition to reach over a billion people by the end of the year. The mechanism this changes for publishers is the click: a search results page has always sent some share of searchers onward to a third-party site, and that referral click is what funds display advertising, subscription conversion and affiliate revenue for the sites Google links to. An AI-generated summary that answers a query directly, before a person scrolls to a traditional listing, changes how many of those clicks happen and, publishers have argued, on which fraction of queries.

What the documents show

Google's own announcement states that 'people are visiting a greater diversity of websites' under AI Overviews and that 'the links included in AI Overviews get more clicks than if the page had appeared as a traditional web listing', framing the change as neutral or positive for referral traffic in aggregate; it does not name any specific publisher or quantify a site-level traffic change. A different kind of document, Chegg's own annual report filed with the SEC, describes the same feature from one named publisher's position: it states that in August 2024 'Google began to significantly expand its AIO search experience', and that this expansion 'has created and is expected to continue to create headwinds for our industry and our business, most notably reductions in traffic to our website and customers subscribing to our services'.

The assumptions exposed

Google's aggregate claim about click quality and site diversity is Google's own self-reported measurement, made without a published methodology a third party can check, and it does not rule out that individual publishers, including ones whose content AI Overviews can summarize directly, experienced a different outcome than the aggregate. Chegg's filing is the reverse: an SEC disclosure naming a specific business, but it is Chegg's own attribution of a business headwind to a Google feature, not an independently measured estimate of how many visits AI Overviews removed. Neither document, read alone or together, supports a general claim about what any other specific publisher experienced.

What to check before you rely on it

A publisher assessing AI Overviews' effect on its own traffic should measure its own referral data rather than borrowing either Google's aggregate claim or another named company's disclosure. This is an editorial checklist, not guidance from Google or Chegg.

• Does my own search-referral data, before and after AI Overviews appeared for my query types, show a measurable change?
• Is the traffic figure I am reading Google's aggregate claim, a single company's self-reported headwind, or my own measurement?
• Which of my content types are most likely to be summarized directly in an AI Overview, versus types that typically still require a click?

Both documents are genuine, but they describe different scopes; a publisher's own numbers remain the only direct evidence of what changed for that publisher specifically.