How to Navigate the New SEO Landscape with Expanded AI Overviews

What changed with Expanded AI Overviews?

Expanded AI Overviews can now load a fuller AI answer and an AI Mode follow-up prompt without requiring a user to select a Show more control. Google told Search Engine Roundtable that AI Overviews may dynamically expand for queries where its systems determine the format is useful. The reported experience moves standard organic listings farther down the results page for those affected searches.

Before this change, a user could encounter a shorter AI Overview and choose whether to reveal more of the response. In the expanded state, the fuller response occupies more of the first screen before a conventional organic result appears. The product change is a search-interface change, not proof that every AI Overview, device, country, or query now behaves this way. Google describes it as dynamic behavior for some queries, which means teams need to observe their own priority searches instead of treating one screenshot as a universal result.

The practical consequence is simple: visibility now has two layers. A page can still rank organically, yet be less immediately visible if a long answer appears first. A page can also appear as a supporting link inside that answer, where the user sees the brand before reaching the blue links. That is why Citation Share matters alongside rank tracking. It measures whether a brand is cited in relevant AI answers, not merely where a URL appears in a traditional list.

Google's current documentation says AI Overviews and AI Mode can use query fan-out, issuing related searches across subtopics and sources to develop a response. That gives a content team a clearer brief than chasing an interface trick. A strong page needs to answer its core question plainly, support its claims, and cover the legitimate next questions that a buyer may ask after the first answer.

1. Track the observed result state: no AI Overview, compact AI Overview, or expanded AI Overview.
2. Separate a supporting-link citation from a traditional organic position in reports.
3. Record device, location, query wording, date, and result state for every priority observation.
4. Treat the change as a tested, dynamic interface behavior unless your own query set shows broader exposure.

Who does Expanded AI Overviews affect first?

Expanded AI Overviews affect publishers and growth teams most when their demand depends on informational searches that Google can answer directly. Definitions, explanations, comparisons, planning questions, and research-led queries have the clearest exposure because an expanded answer can address the initial question before a user reaches an organic listing.

The risk is not confined to high-ranking sites. A business at position one may lose first-screen prominence when the query displays a large answer first. A business at position six may lose even more practical visibility because the user must pass the AI response and several organic results to reach it. For B2B SaaS teams, this changes how they should evaluate category education, alternative pages, integration questions, and buyer research. For local and service businesses, it raises the importance of observing service-plus-location searches rather than assuming the desktop result page looks the same everywhere.

Existing evidence on AI Overviews should be used carefully. Ahrefs reported on April 17, 2025 that its study of 300,000 keywords found a 34.5% lower estimated click-through rate for position-one pages on informational keywords with an AI Overview, compared with a counterfactual forecast. That study predates the reported expanded interface, so it does not measure the incremental impact of expansion. It does establish why click assumptions should be checked rather than carried forward from a pre-AI-results model.

The opportunity is also uneven. Google says that AI has surface relevant links and may provide opportunities for more types of sites to appear. Google also says that an eligible page must be indexed and eligible to be shown with a snippet in Google Search. Expansion does not create a separate technical shortcut. It makes durable technical eligibility, reliable content, and visible citation coverage more important.

1. Publishers with definition and explainer traffic should inspect their highest-impression informational queries first.
2. B2B teams should test category, comparison, cost, implementation, and alternative questions that shape shortlist decisions.
3. Local businesses should test service and location combinations on the devices their buyers use.
4. Teams should not assume that a single AI Overview observation applies to every market or query class.

Why is rank tracking no longer enough?

Rank tracking is no longer enough because it can report a URL position without showing the result surface that appeared above it. A position-one ranking below a compact answer, an expanded answer, a shopping unit, or a local pack is not the same reader experience. The ranking is still real, but the visibility context changed.

Search Console also has a measurement boundary. Google says traffic from AI has is included in the Performance report under the Web search type. Google does not present AI Overview or AI Mode activity as a separate search type in that report. That means a fall or rise in aggregate web clicks cannot, by itself, prove whether an expanded AI Overview caused the outcome.

Build a query-level observation layer beside Search Console. For each priority question, record whether an AI Overview appears, whether it is compact or expanded, which domains appear as supporting links, whether a link opens an external page or continues an AI interaction, and whether your brand is visible. This turns an anecdotal interface change into a repeatable measurement routine.

Use the results to distinguish traffic from answer presence. Traffic still matters because it leads to signups, calls, and revenue. Yet a buyer may see a cited brand, return later through another path, or compare brands inside an AI conversation without producing a clean last-click story. Answer Presence captures whether the brand shows up across the question universe. Citation Share captures how often it is cited among relevant answers. Neither should replace conversion measurement, but both explain exposure that rank alone misses.

1. Keep traditional rank, impressions, clicks, and conversions in the scorecard.
2. Add AI result state and cited-domain observations for the same priority queries.
3. Report Citation Share and Answer Presence separately from Citation Count per day.
4. Review changes by query intent before drawing conclusions from aggregate traffic.

How should content change for expanded answers?

Content should become easier to cite, not merely longer. Start each page with a direct answer to the question it owns. Follow that answer with evidence, scope, definitions, and the decision details a reader needs next. A page that postpones its answer may be harder for a user and an AI system to understand, especially when the search result itself already has a broad summary.

Expanded answers make coverage gaps easier to expose. A buyer who asks about a product category may continue with cost, timing, alternatives, implementation, fit, risks, or local availability. Do not force every possible question onto one page. Create a clear pillar-and-spoke structure, then link related pages together so the site can support a connected research journey. The goal is not to manipulate a model. It is to publish useful, verifiable coverage that deserves to be cited.

Keep claims auditable. State who a product fits, what an approach requires, and where the evidence came from. Use dated primary sources for statistics and changeable platform behavior. Add tables when a reader must compare methods or options. Explain uncertainty when a platform has not published a metric. A citation-grade page gives both people and systems a concrete fact pattern to evaluate.

Do not add fictional AI files or special markup in response to this change. Google states that no additional technical requirements, machine-readable AI text files, or special schema.org markup are needed for inclusion in AI features. Standard search fundamentals remain the foundation: indexing eligibility, crawlable content, helpful information, internal links, accurate structured data where applicable, and strong page experience.

1. Make the answer to the page's main question clear in the opening.
2. Support changing claims with dated, first-party sources where available.
3. Map natural follow-up questions to connected pages instead of repeating thin variations.
4. Check that important pages remain indexable and eligible for normal Google Search snippets.

What should an SEO team do this month?

The first move is to establish a before-and-after baseline for the exact queries that create demand. Capture the current result state now, then compare it with a consistent set of observations over the next several weeks. The relevant before-and-after is not a generic old SERP versus a generic new SERP. It is the buyer question, device, location, date, result state, visible citations, organic placement, and business result.

Prioritize questions by commercial importance rather than by search volume alone. A category-defining query, a high-intent comparison, or a local service question may matter more than a large set of generic terms. For each group, identify the pages that should earn citations and the factual gaps that prevent them from is a dependable source. Then decide whether to improve an existing page, create a missing spoke, or replace unsupported claims with evidence.

Create a reporting cadence that makes the shift visible to decision-makers. Show the observed rate of expanded results in the monitored set, Citation Share by topic, Answer Presence, the domains cited alongside you, conventional rank, and downstream conversions. Include the limitations. A result sample is not a census of all Google users, and interface behavior can vary. Clear caveats make the report more credible, not less useful.

The strategic response is not to abandon SEO. Google says the same foundational SEO practices remain relevant for AI features. The work is to update the definition of winning. Rankings helped a site get found. Citations help a brand get chosen when the first screen is an answer. That is the operating logic behind Citation Engineering.

1. Choose a monitored query set tied to revenue, pipeline, bookings, or qualified research demand.
2. Capture result-state evidence twice each week with a consistent device and location method.
3. Audit pages that should support citations, beginning with weak answers and unsupported claims.
4. Report visibility, citations, traffic, and conversion outcomes together.

What should leaders avoid doing?

Leaders should avoid declaring that organic search is dead because a subset of results expands. Google continues to show conventional search results and says AI Overviews appear only when its systems determine they are additive to classic Search. The more useful question is which buyer questions now produce an expanded result, and whether the brand appears in that answer or only below it.

They should also avoid treating every visible link as equivalent. Search Engine Journal documented that Google displays different link types in AI experiences and that some interactions can continue within Google rather than sending a user to an external page. A citation may create awareness, an external click may create a visit, and a follow-up interaction may create neither immediately. Report each signal honestly.

Finally, do not react with tactics that weaken the site. Do not publish pages with invented data, hide content behind unhelpful technical controls, or produce repetitive pages for every phrasing variation. Google says standard search requirements and people-first content practices apply. The durable response is evidence-backed coverage, clear information architecture, and consistent observation of the search experience your buyers actually receive.

Expanded AI Overviews raise the cost of vague reporting. A weekly ranking report that ignores the answer surface may tell a comforting but incomplete story. A citation visibility report, tied to the questions that matter, gives the team something it can improve.

1. Do not generalize a limited test into a universal traffic forecast.
2. Do not claim a causal traffic loss without query-level evidence.
3. Do not confuse an in-answer citation with a verified external click.
4. Do not replace useful SEO work with unsupported AI-specific tactics.

Sources

Source: Search Engine Roundtable

Google says AI Overviews may dynamically expand for some queries and that the experience can show a fuller answer with an AI Mode follow-up prompt by default. Search Engine Roundtable, 2026-08-28

Source: Google Search Central

Google says AI has can use query fan-out, surface supporting links, require normal Search eligibility, and do not require special AI files or markup. Google Search Central, 2026-06-03

Source: Ahrefs

Ahrefs reported a 34.5% estimated reduction in position-one click-through rate for informational keywords with AI Overviews in its 300,000-keyword study. Ahrefs, 2025-04-17

Source: Search Engine Journal

Search Engine Journal documented multiple AI Overview link displays and the distinction between external-page clicks and links that run a new Google query. Search Engine Journal, 2026-09-25