Enterprise search makes knowledge across your organisation searchable across wikis, documents, chat, tickets and file storage. Staff find answers faster and, increasingly, get summaries or answers written by a language model. Glean is a well-known non-European reference point.
What is distinctive about this category is that the tool has no content of its own but brings together that of all your other systems. For that it needs connectors with broad read permissions, and it builds an index that is a derived copy of your knowledge. That index must keep following the access rights of the source systems. If it does not, an employee sees through the search bar a document they could not open in the source system. For AI answers, the question is which model provider processes your text. The search queries themselves are sensitive too. Switching is relatively simple because the index can be rebuilt, but that does not automatically hold for your connectors and permission model.
When choosing a European option, look for permission checks at query time, an index and vector store you host yourself, the choice of your own or an EU-hosted language model, coverage of your source systems, and propagation of deletions. Also ask how long search logs are kept.