AI Overviews now appear on most big-brand searches

Type "Nike" into Google and the first thing you see is no longer a Nike link. It is a paragraph Google wrote about Nike, stitched together from sources Google chose, in an order Google decided, with Nike's own site reduced to one citation among several. Search Engine Journal reports that tracking data now finds AI Overviews appearing on most big-brand name searches, a sharp escalation from the pattern of a year ago, when Overviews concentrated on informational, long-tail queries and left transactional and navigational searches mostly alone.

That distinction mattered. The working assumption inside most marketing departments, including plenty of sophisticated ones, was that AI Overviews were a long-tail phenomenon: a threat to how-to content and comparison articles, not to the searches that carry a company's own name. A CMO at a bank or an industrial conglomerate could reasonably have told the board that brand search was still a clean, controllable channel, the one part of the funnel where the company's own website would reliably sit at the top. Search Engine Journal's data says that assumption no longer holds. If Overviews now show up on most big-brand queries, then brand search, the highest-intent traffic a company gets, is being intermediated by a model that summarises before the user ever reaches a company-owned page.

Why brand search was the last redoubt

Brand search converts at multiples of generic search because the user has already decided what they want; they are typing your name to find you, confirm a price, check a store hours, or verify a claim. Owning the top result for your own name was never glamorous SEO, but it was close to guaranteed. AI Overviews puncture that guarantee in a specific way: they do not just sit above the organic results, they often answer the query outright, reducing the click-through that used to be close to automatic. Search Engine Journal's tracking suggests this is now the norm rather than the exception for recognisable brand names, which means the erosion is not confined to struggling or lesser-known companies. It hits the brands with the strongest name recognition precisely because strong name recognition is what generates the volume of branded search that Google's systems have apparently decided is worth summarising.

The mechanism is not mysterious. Google's Overviews draw on a mix of the brand's own site, third-party reviews, news coverage, forums, and structured data, then compress it into a synthesised answer. For a bank or insurer, that means a model-generated paragraph about fees, ratings, or complaints could appear before the institution's own disclosures, assembled from whatever Google judged most "helpful," a judgement that does not always track what the institution would choose to foreground. For an industrial group whose brand search volume is driven by investors, regulators, and procurement officers rather than retail consumers, the same dynamic applies with higher stakes: the summary a pension fund analyst sees when searching the company name may cite a trade publication's characterisation of an earnings call rather than the company's own investor relations page.

Citation, not ranking, is now the contest

The practical shift is this: optimising to rank first for your own brand name is no longer sufficient, because ranking first on a results page that now opens with a synthesised paragraph is a smaller prize than it used to be. The contest has moved one level up, to whether a company's own material is among the sources the Overview cites and how favourably that citation is framed. That is a different discipline from classic SEO. It rewards organisations whose public-facing content (newsroom pages, product documentation, regulatory filings, executive commentary) is structured clearly enough and distributed widely enough that Google's summarisation layer treats it as authoritative, rather than defaulting to secondary coverage.

For multilaterals and policy institutions this has a sharper edge still. A UN agency or standards body like ISO exists partly to be the authoritative source on its own mandate; if a branded search for "ISO 9001" or "UNDRR Sendai Framework" now surfaces an AI-generated summary built partly from third-party explainers, the institution's claim to being the primary reference point is being quietly outsourced to whichever secondary sources Google's model trusts most. Philanthropic funders face a similar problem when a search for their own name returns an Overview citing grantee commentary or critical press before the funder's own framing of its work.

None of this is reversible by better keyword targeting. It requires treating the brand-name search result itself as contested territory, auditing what the Overview currently says about the organisation, and working out which of the cited sources Google is trusting over the organisation's own. The companies that will handle this well are the ones that stop treating brand search as owned media and start treating it as earned media, because that, per Search Engine Journal's data, is effectively what it has become.