Infrastructure for Experience Rating & Retrospective Pricing

Experience rating requires explicitly formalized calculation rules: credibility formula parameters, X-Mod calculation methodology per NCCI or independent bureau standards, retrospective rating plan terms, and account-specific modification factors. These rules must be machine-readable and queryable—not just documented in SOPs—because the AI system computes binding premium adjustments with direct financial impact on individual commercial accounts. Regulatory compliance requires that calculation methodology is auditable at the account level, not reconstructed from analyst memory.

Experience modification calculation requires automated capture of individual account loss transactions—paid losses, case reserves, claim counts, closure events—across the full 3-5 year experience window. Each claim event must be timestamped, linked to the correct policy period, and classified by severity for credibility weighting. Automated capture from the claims system ensures the rating algorithm has complete, current account experience without manual data assembly each policy renewal cycle.

Experience rating requires formal ontology mapping Account.PolicyPeriod to Claim.PaidLoss, Claim.CaseReserve, Claim.Severity classification, ExpectedLoss.ByClassification, and CredibilityFactor.ByExposureSize. Without explicit entity relationships, the algorithm can't systematically compute weighted actual-to-expected ratios or apply account-level modification factors to renewal premiums. Machine-readable schema is required for consistent, auditable experience modification calculation across hundreds of commercial accounts.

Experience rating requires API access to claims transaction data, policy administration (exposure and premium by period), bureau expected loss tables (NCCI, state bureaus), and underwriting systems where modification factors are applied. API connectivity enables the rating algorithm to pull current account experience, apply credibility-weighted adjustments, and push modification factors to renewal pricing without manual data transfer. L3 reflects achievable connectivity to core systems.

Experience modification factors must reflect current account loss development—when a large claim closes, reopens, or has a case reserve change, the modification factor for the next renewal period must update accordingly. Event-triggered maintenance ensures material claim developments propagate to the experience rating model between annual renewal cycles. Bureau expected loss tables update annually and must trigger recalibration of classification relativities.

Experience rating integrates the claims system (loss transaction source), policy admin (exposure and premium), bureau data feeds (expected loss tables, credibility parameters), and underwriting platforms (where modification factors drive renewal pricing). API-based connections enable end-to-end experience rating workflows—from claim data ingestion through modification calculation to premium adjustment—without manual data transfer between systems.