AI Use Cases
AI Search
Search that ranks results by meaning as well as by keywords, so people can find the right document even when they do not know the words it uses.
Part of Cloud, Data & AI
AI search, often called semantic or vector search, converts both queries and content into numerical vectors (embeddings) and ranks results by how close they are in meaning. A result can match a query even when they share none of the same words.
The problem it addresses
Keyword search struggles when the searcher's words differ from the document's. A search for "cannot log in" may not match an article titled "Authentication failure resolution", even though it is the right result. The gap is largest in internal knowledge bases, support portals and document repositories, where people often search because they do not know the terminology. It suits large collections of unstructured text, where several plausible results are useful.
When keyword search works better
Exact lookups, such as an order number, a product code or a policy reference, are better served by keyword search. A hybrid that combines keyword and semantic ranking usually gives better results than either one alone.
What it takes to run
Content has to be prepared, split into passages and re-indexed whenever it changes, which is ongoing work. Access control must be applied when results are retrieved, so each user only sees documents they are allowed to open. AI search is also the retrieval step behind RAG.