Skip to content

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?

InsuranceActuarialForecast
Share
AI summary

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

See investigation models →

Continuous backtesting pipeline comparing forecast

Supporting component

Continuous backtesting pipeline comparing forecast to emerged losses each period

See forecast models →

Statistical/actuarial time-series forecasting model (e.g., stochastic chain-ladder or Bornhuetter-Ferguson variants) as the core IBNR engine

Supporting component

See forecast models →

Structured extraction and normalization pipeline

Supporting component

Structured extraction and normalization pipeline for triangle inputs across systems and segments

See extraction models →

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

#378 overall

Catastrophe Exposure

How exposed is the portfolio to catastrophe losses across perils, and where are the concentration hotspots that should inform underwriting, pricing, and reinsurance buying?

Owned by Head of Catastrophe Modeling, Chief Actuary

Insurance

#484 overall

Loss Ratio Forecast

What loss ratio should this book of business or portfolio segment be expected to produce over the coming period, and how should that inform pricing and reserving?

Owned by Chief Actuary, Chief Underwriting Officer

Financial services

#319 overall

Attachment Risk

What is the probability and expected severity that losses on this policy layer will pierce (attach to) a given attachment point?

Owned by Chief Underwriting Officer, Head of Excess and Surplus Lines

Insurance

Reference profile

Oumi doesn't recommend an Oumi-owned specialist model for this engine's core decision.

Atlas of AI. Recommendations require validation against your data and operating constraints.