Semrush Measures What AI Search Does to SEO Traffic

Semrush has published a study on how AI-powered search affects organic traffic, and the company frames the results as a practical answer to a question the industry has repeated for two years. The report carries a plain title — "We Studied the Impact of AI Search on SEO Traffic. Here's What We Learned." — and it comes from a firm with direct access to ranking and traffic data.

The context matters here. Semrush sells visibility and keyword tooling, so it sits on large volumes of search behavior data. When a vendor of this size runs its own measurement rather than quoting third parties, the methodology is the part worth reading.

What does this change for publishers?

That is the practical question. AI search services summarize content and answer queries directly, which reduces the number of users who click through to the source page. Publishers and e-commerce operators have reported uneven effects: some categories lose clicks, others hold steady or shift toward higher-intent visits.

Semrush's contribution is to put its own traffic data behind that debate. The company says the findings come from what it learned by measuring the effect, rather than from speculation about AI summaries replacing the traditional results page.

Why the study lands now

Search providers keep expanding AI-generated answers inside standard results. Each expansion moves more queries toward zero-click outcomes, where the user never leaves the search engine. SEO teams respond by tracking which of their pages still earn visits and which ones the AI layer now satisfies on its own.

A study from a major analytics vendor gives those teams a benchmark to work against. It also gives Semrush a natural bridge to its own AI-visibility tracking products, a commercial angle readers should weigh when reviewing vendor-led research of this kind.

How to read vendor research like this

Three checks apply to any SEO study published by a toolmaker:

• Sample size and time window — results from a narrow period may reflect algorithm updates, not AI behavior alone.
• Definitions — "AI search" can mean AI Overviews, chatbots with search access, or both, and the traffic effects differ.
• Segment breakdown — aggregate numbers hide the split between informational queries, where AI answers hurt most, and transactional ones.

The full report is available through Semrush's publishing channels. Operators who depend on organic search will find the most value in the segment-level detail rather than any single headline figure.