What is Grounding?

Explained

Grounding ensures AI responses are based on verifiable sources—documents, databases, or APIs—rather than model memory. RAG is a primary technique for grounding; it reduces hallucinations.

Decision guides that use this term

• RAG vs fine-tuning: which one does your enterprise use case need?
• RAG vs enterprise search: do your people need answers or documents?
• LLM evals vs human review: how should you assure the quality of enterprise AI?

Further reading

• Fusion RAG: Boosting Accuracy and Relevance in Generative AI
• Taming LLM Hallucinations: A Confidence-Driven Approach for Accurate Customer Experiences

Related terms

• RAG Retrieve relevant documents and use them as context to generate the answer.
• Agent Agenda The agent’s live plan: ordered steps, open subtasks, and checkpoints toward a goal.
• Agent Planner Component or prompt pattern that proposes multi-step plans before tools run.
• Agent Skill Packaged capability for an agent: instructions, allowed tools, and I/O schemas for one coherent task.
• Agentic AI AI systems that act autonomously, make decisions, and execute tasks in sequence.
• AI Agent Software entity that perceives its environment, makes decisions, and executes actions to achieve goals.