Reserve Forecast Engine reference profile.
What is the projected incurred-but-not-reported (IBNR) and total reserve requirement for this book of business over the forecast horizon?
Overview
Improves accuracy of portfolio-level IBNR and total reserve projections used in financial reporting.
Who owns this decision
Chief Actuary, Chief Financial Officer
What this Specialist Model actually does
Decision: What is the projected incurred-but-not-reported (IBNR) and total reserve requirement for this book of business over the forecast horizon?
The final taxonomy records 6 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.
Who uses this Specialist Model
- Chief Actuary
- Chief Financial Officer
- Chief Risk Officer
- VP Actuarial Reserving
Where the specialist model sits
The final taxonomy records 6 supporting or external components, but it doesn't promote an Oumi-owned specialist model for this engine.
Anomaly detection on segment-level development trend shifts
Supporting component
Continuous backtesting pipeline comparing forecast
Supporting component
Continuous backtesting pipeline comparing forecast to emerged losses each period
Statistical/actuarial time-series forecasting model (e.g., stochastic chain-ladder or Bornhuetter-Ferguson variants) as the core IBNR engine
Supporting component
Structured extraction and normalization pipeline
Supporting component
Structured extraction and normalization pipeline for triangle inputs across systems and segments
Specialist LLM
Supporting component
Specialist LLM for narrative explanation of key drivers and trend flags surfaced to actuaries
How Oumi connects to your stack →
Actuarial review workflow
Supporting component
Actuarial review workflow for overriding or confirming model-selected development factors
How Oumi connects to your stack →
How to evaluate and improve it
Evaluation
- Forecast accuracy of IBNR versus actual emerged losses (mean absolute percentage error)
- Calibration of expected range (actual development falls within stated range)
- Segment-level forecast accuracy versus aggregate-only methods
Feedback signals
- Actual paid and incurred loss development versus forecast, updated each reporting period
- Actuarial opinion adjustments and selected-versus-model-recommended factor overrides
- External audit or regulatory examination findings on carried reserves
- Reinsurance recoverable disputes tied to reserve estimates
- Emergence of new large losses or claim types not reflected in historical triangles
- Portfolio mix changes that shift development patterns going forward


