Google is rolling out AI Overviews, blocks with neural-network answers that appear above organic results. Semrush reported an 80% figure. That sounds high, but the number needs context.
Semrush measured 400 commercial queries on desktop in the United States. The narrow sample produced a high result. Other studies found 30% and 18%. The spread shows there is no universal figure — it depends on vertical, region and query type.
Why the figures do not transfer to Runet
Data from the US says little about Russian projects. Yandex does not disclose what share of queries triggers an Alice AI answer. Relying on 80% from foreign reports is pointless.
Each study describes a specific sample: a set of keywords, GEO, device. The results cannot be extrapolated to the entire market. They reflect the dataset of a particular service, not the objective state of search.
How to assess the real situation in your vertical
Instead of relying on external figures, check your own vertical. Collect the keywords you rank for. Run them through Yandex and see where Alice AI appears.
That is the only way to measure how much neural answers affect your niche. Some verticals see high coverage, others minimal. There are no universal forecasts.
What to do with neural answers in SEO strategy
Neural answers take clicks from organic results. If the user gets an answer immediately, they do not visit the site. That affects traffic and conversions.
Optimisation for neural answers is not yet standardised. Google and Yandex pull data from different sources: snippets, structured data, page text. There is no clear algorithm for getting into AI Overviews.
Monitoring and adaptation
Track how the results page changes for your queries. Record where neural answers appear and what their sources are. If your content makes it into AI Overview, that is a plus. If not, analyse where Google is pulling information from.
Structure content with headings, lists and short paragraphs. Neural networks parse that format more easily. Add schema.org markup, answer questions directly.
Runet vs foreign markets
The difference between Yandex and Google in their approach to neural answers is significant. Alice AI is integrated into search differently than AI Overviews. The algorithms for selecting sources differ.
Applying Western cases to Runet directly will not work. Test on your own data, watch traffic dynamics, adjust strategy for the specific search engine and audience.
