A quantitative estimate of how strongly a model’s output is supported by specific retrieved documents or data sources.
Why this matters:
Enterprises want high grounding confidence on their own assets so that AI systems reliably attribute statements to them, reducing misinformation and strengthening perceived authority.