Coverage Coherence Index (CCI)

GMI-INS-019InsuranceInfrastructure-Dependent

Definition

Percentage of AI-bundled multi-policy packages where the component coverages provide coherent, gap-free protection for the customer's risk profile.

The Measurement Gap

Insurance customers increasingly purchase multi-policy packages recommended by AI bundling engines. These engines optimize for price competitiveness, cross-sell conversion, and retention. What they may not optimize for is coverage coordination: whether the component policies in the bundle actually work together to provide comprehensive protection.

A bundling engine can recommend a homeowners policy, an auto policy, an umbrella policy, and a personal articles floater that each individually pass suitability checks while collectively creating coordination failures. The umbrella may not properly attach to the underlying auto limits because the auto limits were set below the umbrella's required minimum. Sublimits across the homeowners and floater may conflict. A coverage gap between the homeowners policy's exclusions and the umbrella's coverage territory may leave the customer exposed.

Legacy distribution metrics report bundle penetration, average policies per household, and retention rates. They do not measure whether the bundled policies coordinate properly. A distribution team can report excellent bundle penetration while building packages that will fail when a large claim tests the coordination between policies.

PSAS (Product Suitability Alignment Score) checks whether each individual product is appropriate for the customer. CCI extends this by checking whether the products work together as a system. A customer can have individually suitable products that collectively leave coverage gaps, creating a coherence failure that no single-product suitability check would catch.

Decision Supported:

Is the AI bundling engine creating packages where policies work together, or is it creating packages that look price-competitive but contain coverage gaps, sublimit conflicts, or coordination failures that the customer discovers at claim time?

Formula

Plain Language: For multi-policy bundles recommended by the AI, have a coverage analyst review whether the component policies coordinate properly: no coverage gaps, no conflicting terms, sublimits stack correctly, and umbrella/excess coverage properly attaches. Score each dimension and take the minimum.

Formula:

CCI = (1/N) x Sum of min(Gap_i, Conflict_i, Stacking_i, Attachment_i)

Variables:

  • N: number of sampled multi-policy bundles
  • Gap_i: 1 if no coverage gaps exist relative to disclosed risk exposures, 0 if gaps exist
  • Conflict_i: 1 if no conflicting terms between policies, 0 if conflicts exist
  • Stacking_i: 1 if sublimits stack correctly across policies, 0 if they do not
  • Attachment_i: 1 if umbrella/excess properly attaches to all underlying policies, 0 if attachment gaps exist
  • Using min() rather than average: any single failure equals an incoherent bundle

Units: 0 to 1. 1.0 means every sampled bundle is coherent. Lower values indicate systematic coordination failures.

Data Requirements:

Data Element Source System Classification
AI-recommended bundles (component policies, limits, endorsements) Recommendation engine / quoting platform NATIVE
Policy form details for each component Policy administration system / forms library NATIVE
Coverage analyst coherence review Structured review program (new process) MANUAL
Customer risk profile (for gap assessment) Application data NATIVE
State-specific coordination rules Compliance / regulatory documentation MANUAL

Proposed formula, pending validation pilots.

At a Glance

Unit
Score (0 to 1)
Trend Direction
Toward 1.0
Reporting Frequency
Quarterly
Category
Distribution Fairness and Suitability
← Triage Complexity Currency Index (TCCI)Non-Renewal Disparate Impact Index (NRDII) →

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