Competing Attention Zones
A modern Google results page can present an answer before it presents results. An AI Overview appears above the organic listings for eligible queries, and AI Mode goes further, offering a conversational experience built on the same underlying systems.
Attention follows a predictable order: the generated answer first, then the links, if the reader still needs them. For SEO this creates a two-track problem, because being the best blue link and being a source the answer draws on are related but separate outcomes.
What Has Not Changed
More than people expect. Google's May 2026 guidance is explicit that its generative features are rooted in its core Search ranking and quality systems, and that there are no additional requirements to appear in them. So the foundations still hold:
- Crawlability and indexing. A page must be indexed and snippet-eligible to be used as a supporting link. That is the technical bar, and there is nothing beyond it.
- Page experience. Still expected, still worth the engineering time.
- Genuine relevance and depth. Still what the quality systems are built to reward.
- Authority. Still influences where you land in the organic results.
Abandoning traditional SEO in favour of AI-specific tactics is the wrong trade. The same index and the same quality signals feed both surfaces.
What Has Changed
The unit of competition widened. Google's AI features use query fan-out, issuing multiple related searches across subtopics before synthesizing an answer. A page can be surfaced for a query it never targeted because it answered one of the subqueries well. Breadth across a topic now has a mechanical payoff, not just a reputational one.
Exposure decoupled from clicks. A reader who sees you cited and does not click has still encountered you. Click-based analytics will not record that, so a citation-heavy strategy can look flat in a traffic report while working. Branded search volume and assistant referral traffic are the compensating signals.
Attribution got harder. Visits arriving from AI assistants land in GA4 as ordinary referrals, spread across individual source rows such as chatgpt.com and perplexity.ai rather than grouped as a channel. Without deliberate grouping, AI referral traffic is present in your analytics but effectively invisible.
Measurement improved. Since June 2026, Search Console has reported site visibility within generative AI features on Search. Much of what previously required inference from CTR patterns is now reported directly.
What Has Not Changed But Is Widely Claimed To Have
It is worth stating plainly, because a great deal of 2025-era advice is still circulating: AI Overviews did not introduce a requirement for special schema, an llms.txt file, chunked content, or a particular word count. Google has addressed each of those directly and said they are not needed. Time spent on them is time not spent on coverage and originality, which are what the guidance actually asks for.
The Integrated Strategy
Differentiate by query type rather than applying one approach everywhere:
- Queries returning a generated answer. Pursue both the citation and the ranking. Monitor citation status alongside position, and accept that the click rate for these will behave differently from the rest of your set.
- Queries not returning one. Apply conventional SEO. Commercial and navigational intent frequently falls here, and click behaviour on these queries remains familiar.
- Queries where you are already cited. Treat these as exposure wins and protect them. Measure via branded search and assistant referrals rather than direct clicks alone, or you will conclude a working page is failing.
AIORadar's AI Overview Tracker classifies a keyword set by whether an AI Overview appears and whether you are cited, which is what lets you sort your queries into those three buckets in the first place.