Insurability & Community Development
      Back to home
      1. InnSure Innovation Hub
      2. Insurability & Community Development

      What is the InnSure Total Cost of Risk Simulator?

      An AI-driven municipal planning tool that community leaders can use to inform proactive community-wide risk management and investment planning.

      Technology Overview

      InnSure is currently in the process of designing, developing, and implementing an AI-driven Total Cost of Risk (TCOR) Simulator aimed at enhancing community planning and promoting climate resilience initiatives. This innovative technology is being developed concurrently with several pilot and demonstration projects being conducted with partner municipalities. Each pilot targets a specific community and addresses a particular hazard area, allowing for the testing and demonstration of the simulator's effectiveness in supporting economic development and resilience planning efforts.

      The TCOR Simulator is being designed to analyze, distill, and effectively convey complex risk management concepts to community leaders and other stakeholders. The simulator will leverage large language model (LLM) technology to create a user-friendly interface and calculate the expected direct and indirect economic loss costs in the pilot community based on different scenarios including adjustments in climate change models. 

      The simulator will model the following:

      • Economic Loss Costs
      • Climate Adaptation costs and benfits
      • Risk Transfer (Insurance) Costs & response & recovery Payments

      Modeling Economic Loss Costs

      Representative Inputs (Datasets and Assessment Models)

      The solution will utilize commercially available, publicly available open-source and locally curated datasets and models. 

      • Asset Values and Vulnerability Data
      • Population Demographics
      • Peril Specific Cat Models
      • Scenario Based Recovery Data
      • Climate Adjustment Models
      • Inflation adjustment models

      Representative Outputs

      Where possible the system will output data in Oasis Loss Modeling formats. 

      • Direct Loss Costs
        • Physical Damages to Assets
        • Life and Health Impacts to People
        • Displacement of Households and Businesses
        • Loss of Economic Activities
          • Business interruptions
          • Wage impairments
        • Tax Revenue Impacts
      • Indirect Loss Costs
        • Extra expenses for repair/rebuild/recover efforts
        • Cascading community impacts
          • Asset devaluation
          • Tax revenue impacts 
          • Increased cost of financing including bond market impacts
      • Climate Adjustments
        • For forecasting 5+ years into future
        • Based on 5 IPCC scenarios
        • Based on inflation expectations

      For further details, please refer to the Total Cost of Risk Analysis Framework article and our prototype mockup.

      Modeling Climate Adaptation Costs and Benefits 

      The simulator will also calculate the costs of climate adaptation investments and their associated expected loss cost reductions. Climate adaptation investments bend the Uninsured Average Annual loss (UAAL) down, by reducing the average severity and frequency. Given that some of these investments also have other co-benefits to a community (i.e., if investment is a community amenity), the TCOR Simulator will also attempt to consider these co-benefits while assessing the net cost effectiveness of climate adaptation investments. 

      Representative Inputs

      • Scenario based costs details
      • Loss reduction models
      Representative Outputs
      • Risk Management Expenditures
      • Recovery and Emergency Management Costs
      • Infrastructure Investments
        • Technical Solutions
        • Nature Based Solutions
      • Expected Loss-reduction Impacts Resulting From Changes in Adaptation Investment

      Modeling Risk Transfer (Insurance) Costs & Recovery Payments

      Representative Inputs

      • Insurance Uptake/Gap Data (to identify areas/households with protection)
      • Catastrophe models - by peril and geography
      • Premium data projection models for new and existing insurance products
      • Relief reimbursement estimation models (amount, timing and effort to obtain) 

      Representative Outputs

      • Insurance costs & expected recovery
        • Expected premium costs
        • Administration costs
        • Expected recovery payments
      • Disaster relief (Gov't & Philanthropic funded)
        • Pre-loss funding 
          • Reserves
          • Administrative
        • Expected recovery payments

      Modular Agent-Based Design & Components Breakdown

      The system is designed to be LLM-driven, with a structured Prompt Library, Data Library, and an Agent Interface that allows users to ask risk-related queries.

      Agent Interface

      • The primary user interface where users can query the model.
      • Accepts structured and unstructured queries (e.g., "What would happen to 45 Wharf St, {City} in a 3ft storm surge?").
      • Routes user inputs to the Prompt Execution Engine.

      Prompt Execution Engine (LLM Core)

      • Manages prompt selection based on user queries.
      • Calls relevant prompts from the Prompt Library.
      • Interacts with the Data Library to fetch raw data before performing calculations.
      • Returns structured risk assessments & resilience recommendations.

      Prompt Library: Structured Prompt Flow

      Each step in answering a user query is handled by a specific prompt within the Prompt Library, leveraging different datasets from the Data Library.