Knowledge Base Contamination Rate (KBCR)

GMI-SOLO-016Solo AI OperationsPilot-ReadyPattern O-3: Twin Fidelity

Definition

Share of sampled retrievals that return stale, wrong, or duplicate-conflicting content, measured against a standing panel of representative queries at a fixed retrieval depth.

The Measurement Gap

There is no metric for the staleness and contamination rate of the operator's knowledge base as a share of what retrieval actually returns. Retrieval quality decays while retrieval volume looks healthy, and corpus growth is mistaken for capability.

Formula

KBCR = R_contaminated / R_sampled

Proposed formula, pending validation pilots.

At a Glance

Unit
Ratio (0 to 1)
Trend Direction
Lower
Reporting Frequency
Monthly
Category
Knowledge and Context Integrity

Decision Supported

What enters the knowledge base, when to prune it, and how much to trust retrieval for production writing.

← Cross-Store Divergence Rate (CSDR)Asset Fix Ratio (AFR) →

Assess Your Organization

The AI-GPS Assessment evaluates your AI governance maturity and recommends KPIs like this one for your implementation roadmap.

Take the Free AI-GPS Assessment