Enterprise AI Glossary

How to use this glossary

Every term here maps to a decision you'll actually make building enterprise AI. If you're picking between RAG and fine-tuning, you need to know what each one changes. If you're writing prompts that sometimes hallucinate, you need to know what grounding does. The definitions are short. The "In Practice" sections name real vendors and real numbers. The "Worked Example" sections show one concrete end-to-end scenario.

Terms cluster into four groups. Model behavior: context window, hallucination, grounding, prompt engineering, fine-tuning. Retrieval: embeddings, vector databases, chunking, semantic search, RAG. Agents: AI agent, agentic AI, multi-agent systems, tool use, MCP. Operations: LLM orchestration, guardrails, observability, governance.