Change Attribution Success Rate (CASR)

GMI-MLOPS-022MLOps / AI Platform OperationsInfrastructure-DependentPattern O-32: Change Attribution Integrity

Definition

The share of material output changes for which the causal change (model, prompt, index, data, guardrail, routing policy) can be identified from platform records within a bounded investigation time.

The Measurement Gap

There is no metric that tells you whether an output change can be attributed to its causal change, which means incident response and rollback targeting rest on tribal memory.

Formula

CASR = attributed_within_timebox_and_confirmed / investigated_output_changes

Proposed formula, pending validation pilots.

At a Glance

Category
Estate and Capacity Governance

Decision Supported

Whether incident response and rollback targeting rest on records or tribal memory, and where the change ledger has holes.

← Training Data Recursion Fraction (TDRF)Eval Workload Segregation Index (EWSI) →

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