Agent Grounding

Agent grounding is the practice of anchoring an agent's responses and decisions in verified, authoritative data rather than relying solely on the model's parametric knowledge. Grounding reduces hallucination and improves factual accuracy.