Claim Reserve

Claim reserves are informal estimates in adjuster heads or handwritten notes in claim files. No standardized process exists for setting initial reserves or updating them as claims develop. Reserve amounts may not be documented at all until final payment. Each adjuster uses their own judgment with no explicit methodology or criteria.

None — AI cannot access unrecorded mental estimates or handwritten notes. Reserve adequacy analysis, trend forecasting, and portfolio management are impossible. Actuarial reserving relies on aggregated paid loss data with no insight into case-by-case reserve judgment.

Adjusters enter initial reserve estimates into the claims system with separate indemnity and expense components. However, reserve rationale is informal — adjusters set amounts based on experience without documenting methodology. Reserve changes are recorded but without explanation of why reserves increased or decreased. No standardized reserve adequacy criteria exist.

Basic reserve tracking and reporting are possible. AI can analyze reserve development patterns but cannot recommend appropriate reserves or detect inadequate reserves because the factors influencing reserve judgment (injury severity, liability assessment, jurisdiction) aren't captured in structured form.

Claim reserves include structured reserve methodology: injury severity ratings (soft tissue, fracture, permanent disability), liability percentage assessments (0%, 25%, 50%, 100%), jurisdiction-specific settlement value ranges, and expected treatment cost estimates. Reserve changes are documented with explicit rationale referencing which factors changed (new medical information, liability determination, settlement offer received).

AI can analyze reserve adequacy by comparing reserves to similar historical claims and recommend reserve adjustments when factors change. However, AI cannot fully automate reserve setting because reserve judgment incorporates nuanced factors (adjuster assessment of claimant credibility, attorney aggressiveness, jury bias in specific venues) that aren't formalized.

Claim reserves follow formalized criteria. Reserve ranges are defined by injury severity and liability percentage (soft tissue injury at 100% liability: $15k-$25k). Reserve adjustment triggers are explicit (new MRI showing herniated disc: increase reserve by $30k-$50k). Reserve authority limits are specified by adjuster experience and claim complexity. Every reserve decision or adjustment references the criteria and triggers applied.

AI can recommend initial reserves based on formalized criteria and suggest adjustments when triggers are met. Complex reserve judgment involving subjective factors (claimant likeability, attorney reputation, venue-specific jury tendencies) still requires adjuster expertise. However, AI cannot learn from actual claim outcomes to improve reserve accuracy because final settlement amounts aren't systematically linked back to initial reserve estimates.

Claim reserve outcomes feed back to reserve models. When claims close, the system records actual settlement amounts vs. reserves at FNOL, 60 days, 180 days, and pre-settlement, analyzing reserve accuracy by injury type, jurisdiction, and adjuster. This feedback continuously refines reserve prediction models, improving initial reserve accuracy and identifying leading indicators of reserve inadequacy. AI learns from every claim outcome, adapting to evolving settlement patterns.

AI reserve recommendations improve continuously through closed-loop learning, producing accurate reserves that reflect actual settlement outcomes. However, AI operates reactively — reserves are set after losses occur. Proactive loss cost forecasting (predicting ultimate claim costs at underwriting to inform pricing) isn't possible because reserve intelligence doesn't integrate with underwriting.

Reserve intelligence operates across the policy lifecycle. At underwriting, AI forecasts expected claim severities for new policies based on risk characteristics and historical reserve outcomes. At claims, reserves are set using formalized criteria refined by continuous outcome learning. Post-settlement, actual costs update severity forecasting models. Reserve methodology is formalized, proactive, and continuously learning across all insurance touchpoints.

Fully autonomous reserve setting and management optimized by continuous learning from claim outcomes. AI sets accurate initial reserves, adjusts reserves proactively as claims develop, and informs underwriting pricing with refined severity forecasts, improving reserve adequacy and pricing accuracy.

Ceiling of the CMC framework for this dimension.