AI Benefits Cost Projection Agent

How Cassidy automates this using AI

Step 1: Trigger on data refresh

The Workflow activates when new claims, enrollment, or financial data lands in your systems—whether from monthly ledger closes, eligibility feeds, or pharmacy rebate reconciliations.

Step 2: Validate and reconcile data

Cassidy applies ASOP 23 data quality checks, reconciles incurred claims to financial ledgers, screens for outliers, and flags incomplete runout periods requiring completion factor adjustments.

Step 3: Calculate base period PMPM

The agent computes allowed and paid PMPM by service category, applies large claim pooling and normalization, and blends with manual rates where credibility is limited.

Step 4: Decompose and project trends

Cassidy separates unit cost from utilization trends, adjusts for leverage effects and risk score normalization, layers in known contract and formulary changes, and generates monthly PMPM projections by line of business.

Step 5: Run scenarios and validate

The Workflow executes sensitivity tests, back-tests against holdout periods, calculates confidence intervals, and compares results to external benchmarks and prior model runs.

Step 6: Assemble deliverables

Cassidy produces PMPM forecast tables, assumption inventories, bridge exhibits, tornado charts, and ASOP-compliant actuarial memoranda—ready for peer review, regulatory filing, or executive presentation.