Semantic Search
Semantic search retrieves results by meaning rather than exact keyword matches, converting the query and candidate documents into embeddings (numeric vectors) and ranking by how close those vectors sit in that space, so a page can match a question even when it shares no words with it.
Older keyword search counted term overlap: a page ranked because it repeated the searched words. Semantic search instead measures conceptual similarity, so a query like "shoes that survive trail running" can surface a product described as "rugged off-road sneakers" with no shared terms. This is why keyword stuffing has lost its force; padding a page with repeated phrases does little, while clear, specific writing that states the concept plainly tends to land closer to the relevant queries.
The practical lesson is to write the way a person would ask, and to define the subject in direct language near the top of the content. Specificity helps: "waterproof to 50 metres" embeds more usefully than "great for water", because it carries a concrete concept the model can locate.
Semantic retrieval also underpins most AI answers. When an assistant pulls sources before responding, it usually retrieves by embedding similarity rather than keywords, then summarises what it finds. A caveat worth stating: similarity is not the same as accuracy, so a page can be retrieved for being topically close while still being wrong, which is one reason corroboration between independent sources matters for what gets cited.