Enterprise Search Glossary

Every term you need to understand enterprise search, knowledge management, and AI-powered retrieval — explained in plain language.

RAG (Retrieval-Augmented Generation)

An AI architecture that combines vector-based semantic retrieval with large language model generation to answer questions over proprietary data. Enterprise RAG platforms include connectors, permission controls, and audit logging beyond the basic retriever-and-LLM pattern.

Learn more →

Vector Embeddings

Numerical representations of text meaning. When documents are converted into vector embeddings, semantically similar content gets similar numerical representations — enabling search by meaning rather than keywords.

Learn more →

Vector Database

A specialized database that stores and queries vector embeddings efficiently. Vector databases enable fast similarity search across millions of document chunks, forming the retrieval layer of semantic search and RAG systems.

Workspace

A logical container in enterprise search that scopes search to specific data sources and restricts access to specific users. Workspaces enable multi-tenant isolation — a consulting firm can create separate workspaces per client, each with its own connectors and members.

Learn more →

Connector

An integration module that connects an enterprise search platform to a specific SaaS application via OAuth. Connectors handle authentication, data sync, incremental indexing, and permission metadata extraction.

Learn more →

OAuth

An open standard for access delegation. In enterprise search, OAuth enables one-click connection to SaaS applications like Gmail, Google Drive, and Confluence without sharing passwords or API keys.

Chunking

The process of splitting documents into smaller segments optimized for embedding and retrieval. Chunk size, overlap, and boundary detection directly affect search quality in semantic search systems.

Citation

A reference linking an AI-generated answer back to the source document. Citations include the source system, author, timestamp, and a direct link, enabling users to verify AI responses against original content.

AI Synthesis

The process of generating a direct answer from multiple retrieved documents rather than returning a list of links. AI synthesis combines information from across platforms into a coherent, cited response.

Knowledge Management

The practice of creating, sharing, using, and managing organizational knowledge. Modern knowledge management focuses on making existing knowledge findable rather than creating new documentation systems.

Learn more →

Institutional Knowledge

The accumulated experience, decisions, processes, and context that an organization develops over time. Research shows 42% of institutional knowledge exists only in employees' heads — disappearing when they leave.

Learn more →

Tribal Knowledge

Unwritten expertise, workflows, and informal practices that exist only in the minds of experienced employees. Tribal knowledge is the most valuable and most vulnerable form of organizational knowledge.

SaaS Sprawl

The proliferation of SaaS applications across an organization. The average company uses 112 SaaS apps, each creating a separate data silo with its own search bar. SaaS sprawl is the root cause of the enterprise search problem.

Learn more →

Data Silo

A repository of information that is isolated from other systems. Each SaaS application creates a data silo — Confluence search cannot see Gmail, SharePoint search cannot see Jira. Enterprise search breaks down silos by searching across all of them.

E-Discovery

The process of identifying and collecting electronically stored information for legal proceedings. Enterprise search accelerates e-discovery by finding relevant documents across all platforms in a single query.

Learn more →

MCP (Model Context Protocol)

A protocol that enables AI agents and tools to query external knowledge sources. An MCP server for enterprise search lets AI assistants like Claude access organizational knowledge directly during conversations.

Learn more →

Audit Trail

A chronological log of all search queries, results returned, and AI-generated answers within an enterprise search system. Audit trails are required for compliance in regulated industries like healthcare, finance, and legal.

Learn more →

MTTR (Mean Time to Resolve)

The average time from incident detection to complete resolution. Enterprise search reduces MTTR by enabling support and engineering teams to find past incident resolutions, runbooks, and troubleshooting docs in seconds instead of hours.

Learn more →

See Enterprise Search in Action

RetrieveIT connects to your existing tools and makes every document, email, and conversation searchable from a single interface.