In short. Seznam AI summaries draw on search results, provide source links and support publisher opt-out rules.

Definition: Seznam AI summaries are answers in Seznam Search that combine information from web pages and link to their sources. Seznam describes them as AI sumarizace in its March 2025 explainer.

Key facts

Seznam's March explainer says most summaries at that time draw on three to five sources, with links at individual citations and beneath the summary.
Its January 2025 product article describes extracting relevant content from top search documents and passing it to SeLLMa, Seznam's internal language model.
As of 17 October 2026, Seznam's AI answer controls document a website opt-out through Seznam-Extended in robots.txt or a nosnippet meta tag.

How does Seznam select source material?

Seznam's product explanation starts with a suitability check on the query. The documented focus is non-advertising, highly informational queries for which another form of direct answer is absent.

For an eligible query, Seznam retrieves ordinary search results. It extracts the most relevant content from the top documents and adds that content to a prompt, the input sent to the model. SeLLMa processes the prompt and returns a summary with sources. The article therefore places search retrieval and content extraction before answer generation. It does not publish a formula for selecting individual source pages.

The same article also describes summaries of trending queries, which explain why a query attracts searches, and news summaries, which combine articles about current news topics.

Where do the source links appear?

Seznam's March explanation identifies links beside individual citations, beneath the summary and in the ordinary search results. These locations give readers routes from the condensed answer to the source website. The company describes the displayed content as a sample of the source site's content, with the option to continue reading there.

The same explanation names quality, relevance, source variety and the overall trustworthiness of the summary as priorities. It also describes direct user feedback, user surveys and human evaluators who assess answer quality. Source links and these evaluation methods are separate parts of the documented process.

How can a website opt out?

Seznam's AI answer help page gives this rule for robots.txt, the file used to state crawler access rules:

User-agent: Seznam-Extended
Disallow: /

The page also accepts equivalent blocking rules for the tokens AI or MachineLearning and says one of the listed rules is sufficient. Its alternative is an HTML meta tag placed in the page's head:

<meta name="seznambot" content="nosnippet">

Seznam warns that this meta tag also disables search snippets, the text excerpts displayed with ordinary results. A website using it appears in Seznam Search without a text excerpt.

FAQ

Does every query receive a summary?

Seznam's product article describes a query suitability check, with a focus on highly informational queries that lack another direct answer.

Where can I open the original pages?

Seznam's explainer identifies links at individual citations and beneath the summary. You can follow those links to the source websites.

Can I stop my website being used for AI answers?

Yes. Seznam's help page documents the Seznam-Extended blocking rule above. If you choose the nosnippet alternative, it also removes text excerpts from ordinary search results.

Sources

Check how a page reads to answer engines. Open the GEO score analyzer