If your newsroom, agency or SEO team chases referral spikes from breaking stories, Google has quietly shifted one of those spikes into its AI-generated search surface. For some developing-topic queries, AI Mode now places a horizontal carousel of article cards — with image, headline, source and date — directly inside the generated response.

What changed

Robby Stein, Google Search’s VP of Product, announced on X that link carousels for developing topics are now live in AI Mode (they previously appeared in AI Overviews). For eligible queries about unfolding stories, Google can insert a row of article cards between sections of the AI-generated answer. Stein illustrated the feature with an example showing three article cards for a query about newly discovered gold bars in Belgium.

Why this matters to publishers and SEOs

The carousel is a visually prominent link placement — larger and more image-forward than the compact citations Google typically appends to generated text. That increased prominence can make impressions more noticeable and may drive additional clicks to original reporting, particularly for timely coverage where users want multiple perspectives.

Operationally, links surfaced in AI Mode can generate Search Console impressions and clicks. Google does not yet provide a Search Console filter that isolates impressions from this specific developing-topic carousel, so publishers will need to infer its impact from broader query and impressions trends rather than a labeled metric.

How this fits into Google’s broader rollout

The developing-topic carousel first appeared in May as part of Google’s effort to connect AI Search users with original reporting and different perspectives. Stein’s announcement confirms the format is now present across both AI Overviews and AI Mode, although Google says the carousel will appear only for some searches about developing topics, not universally.

Google has also expanded its Preferred Sources program: the company reports more than 600,000 unique sources are now selected as Preferred Sources, up from over 345,000 in May. Preferred Sources can be highlighted within the carousel, which may affect which outlets receive the most visible placements for breaking stories.

What to monitor and what to do next

Expect this to behave like a new discovery channel for timely reporting. Practical actions publishers, newsrooms and SEO teams should take now:

  • Monitor Search Console for sudden changes in impressions and referral clicks tied to breaking queries. Look for query-level spikes around developing topics and compare timing to published timestamps.
  • Ensure article pages include clear publication dates, accurate metadata and a high-quality lead image — the carousel displays images and dates, and those elements affect rendering quality.
  • Streamline rapid-publishing workflows. Visibility in the carousel favors timely coverage; speed plus accuracy improves the chance of being surfaced.
  • Track Preferred Sources signals. If your site is included among Preferred Sources, you may see differential placement within the carousel.

Because Google is still testing different link designs inside AI experiences, continue to treat the carousel as an experimental but potentially meaningful referral source. Without a dedicated Search Console filter, attribution will rely on correlation across timestamps, query activity and referral patterns.

What to watch

Two developments will determine how important this channel becomes: whether Google adds Search Console filters or documentation that explicitly surfaces AI Mode placements, and whether the carousel’s eligibility expands beyond narrowly defined developing topics. Both would make the feature easier to measure and potentially wider in impact.

For now, publishers should sharpen metadata, images and publishing speed for breaking coverage and prepare analytics queries to detect correlated shifts in impressions and clicks. If the carousel starts to appear more often, the winners will be the outlets that combine fast, accurate reporting with clean metadata and strong visual assets.