Domain Health

Domain Health gives you a quantitative picture of your corpus quality. Every domain gets a composite health score built from four dimensions: embedding coverage, chunk distribution, duplicate detection, and freshness.

Gap analysis surfaces what is missing. When users or agents query topics that return low-confidence results, those patterns accumulate. Domain Health shows them as gap signals: "users asked about deployment pipelines 14 times this week, but the engineering domain has no documents about deployment." This tells you what to ingest next, before retrieval quality becomes a problem.

The freshness dimension combines document age with query frequency. A domain where the newest document is six months old is not necessarily unhealthy if nobody queries it. But a domain with old documents and high query volume is a risk. The staleness risk score accounts for how actively the domain is being used.

The portfolio view shows all domains side by side. Identify which domains need investment (high query volume, low health), which are stable (low volume, high health), and which are trending down. This turns corpus management from a reactive activity into a proactive one.