AI Overviews are now cutting organic click-through on the average top-ranking result by roughly a third. The instinctive response from a lot of teams is to publish more — more posts, more pages, more coverage. That instinct is understandable and, on its own, wrong. Volume doesn’t fix a visibility problem that’s structural, not quantitative.
More Pages Isn’t More Authority
Adding pages without a coherent topical structure behind them tends to create internal competition — several thin pages splitting relevance for the same query instead of one strong page owning it. That’s the trap semantic SEO vs. traditional SEO addresses directly: traditional SEO measures success in pages published, semantic SEO measures it in topics fully owned.
What Actually Recovers Lost Clicks
If an AI Overview is going to answer the query before the user clicks anything, the only way to still capture that visit — or at minimum the brand impression — is becoming the source the Overview cites. That’s a different optimization target than ranking position, covered in detail in optimizing for Google AI Overviews and AI citation optimization. Consolidating overlapping content into one authoritative page, rather than adding a competing fourth one, is usually the faster path — a distinction covered in macro vs. micro semantic SEO.
Measuring the Right Thing
Click-through rate on its own will keep declining regardless of what you do — that’s the format shift, not a sign of failure. The metric worth tracking is share of citations and impressions within AI answers for your core topics, alongside traditional rankings. Expect this to take longer to show clean movement than classic ranking changes did; realistic timelines are covered in how long semantic SEO takes to show results.
Fixing the Structure, Not the Volume
This is the whole argument for mapping a topic’s entities and attributes before producing content, rather than after — which is what SemanticOS is built to do, and where a consultant engagement usually starts: not “what should we publish next,” but “what’s actually missing from the topic we already claim to own.”