Semantic + lexical
Meaning-based retrieval is fused with exact-term matching instead of forcing one method.
ServAI searches organizational knowledge with semantic, lexical, metadata and structured context while preserving source locations and enforcing live scope boundaries.
ServAI combines semantic and lexical retrieval with metadata and entity context, then reranks and diversifies the result set. Authorized scope is enforced before anything reaches the user.
Meaning-based retrieval is fused with exact-term matching instead of forcing one method.
Source, date, classification, scope and structured fields can narrow the evidence set.
Knowledge can carry organization, account, role, product, KPI and relationship projections.
Page, slide, sheet/cell, section, line, speaker and timestamp locators survive into results.
Candidates are rechecked after backend retrieval; a permissive backend cannot widen access.
Diagnostics can show why accessible results ranked without exposing inaccessible candidates.
The Knowledge surface is designed to say when organizational evidence is missing instead of filling the gap with a confident but ungrounded answer.
ServAI can carry an evidence-backed finding into a governed Mission or specialist workflow while preserving human authority.