Searching by meaning, not exact words.
Classic keyword search (what Google used to be) matches your query against exact words on the page. Semantic search matches against MEANING. Under the hood, it embeds your query and every document, then finds the ones with the closest embeddings. You can search for 'how to fix a slow laptop' and find articles titled 'speeding up your PC' or 'performance troubleshooting', even though they share zero keywords with your query.
Semantic search is the superpower layer underneath most useful AI features you see today: smart support bots, 'similar items' recommendations, better search boxes. When paired with classic keyword search (a 'hybrid' approach), it's the gold standard.