Catastrophe Risk Model Disclaimer

Catastrophes.ai develops artificial intelligence-powered catastrophe risk intelligence technologies designed to support insurance underwriting, portfolio analysis, and risk management.

Our catastrophe risk models combine:

• Machine learning
• Climate analytics
• Historical observations
• Weather and environmental data
• Geographic information
• Property intelligence
• Statistical modeling techniques.

These models are designed to provide forward-looking risk insights and probabilistic analysis for insurance and risk management applications.

Catastrophes.ai catastrophe models are intended to support:

• Insurance underwriting analysis
• Portfolio risk management
• Catastrophe exposure assessment
• Risk selection
• Pricing analysis
• Scenario analysis
• Strategic planning.

Model outputs may include:

• Hazard assessments
• Risk scores
• Probability estimates
• Expected loss estimates
• Loss ratio projections
• Scenario simulations
• Geographic risk analysis.

These outputs are designed to assist professional decision-making.

Catastrophe risk models estimate potential future outcomes using available information and analytical methods.

Natural catastrophe events are inherently uncertain due to factors including:

• Atmospheric variability
• Climate system complexity
• Geographic conditions
• Human and infrastructure factors
• Data limitations
• Emerging environmental changes.

Accordingly, model outputs represent probabilistic estimates and scenarios rather than deterministic predictions.

Catastrophes.ai does not guarantee that:

• A predicted catastrophe event will occur
• A predicted event will not occur
• Estimated losses will exactly match actual losses
• Future catastrophe frequency or severity will follow historical patterns.

Actual events may differ from modeled scenarios due to unforeseen conditions or factors outside model assumptions.

Catastrophes.ai incorporates climate and environmental information into risk analysis.

However:

• Climate systems are complex and continuously evolving
• Historical observations may not fully represent future conditions
• Future climate patterns may differ from modeled assumptions
• Scientific understanding continues to develop.

Users should interpret climate-related risk insights together with additional scientific, operational, and business considerations.

Catastrophes.ai may generate loss-related analytics, including:

• Expected loss
• Loss frequency
• Loss severity
• Loss ratio estimates
• Scenario-based financial impacts.

These estimates are not guarantees of actual insured losses.

Actual losses may be influenced by:

• Policy terms
• Coverage structures
• Deductibles
• Claims handling
• Market conditions
• Construction changes
• Loss mitigation measures
• Event characteristics.

Catastrophe model performance depends on the quality, availability, and suitability of underlying information.

Model inputs may include:

• Historical weather data
• Climate datasets
• Satellite and aerial imagery
• Geographic information
• Property characteristics
• Claims experience
• Publicly available information
• Customer-provided information.

Incomplete, inaccurate, or outdated information may affect model outputs.

Catastrophes.ai uses artificial intelligence and machine learning methods in catastrophe risk analysis.

AI-based models may have limitations, including:

• Limited representation of rare extreme events
• Sensitivity to input data quality
• Uncertainty in extrapolating future conditions
• Changes in environmental or economic conditions
• Model assumptions that may not capture all real-world complexity.

Model outputs should be reviewed using appropriate professional judgment.

Catastrophes.ai catastrophe models may support insurance underwriting workflows.

However:

• Model outputs are not, by themselves, binding underwriting decisions
• Risk scores do not determine coverage eligibility automatically
• Pricing decisions should consider additional underwriting factors
• Insurance professionals remain responsible for final decisions where applicable.

Where Catastrophes.ai operates as a licensed insurance entity or MGA, underwriting activities will be performed according to applicable insurance regulations and authority requirements.

Organizations using Catastrophes.ai for portfolio analysis should consider:

• Model assumptions
• Portfolio characteristics
• Geographic concentration
• Policy structures
• Risk appetite
• Regulatory requirements.

Catastrophe model outputs should be incorporated into broader enterprise risk management processes.

Catastrophes.ai continuously improves its catastrophe risk models through:

• New scientific information
• Improved datasets
• Updated methodologies
• Model validation
• Performance analysis.

Model outputs may change over time as models are updated.

Changes in model results do not necessarily indicate changes in underlying physical risk, but may reflect improved methodologies or updated information.

Extreme catastrophe events are rare and may have limited historical observations.

For such events:

• Statistical uncertainty may be higher
• Model confidence may be lower
• Historical patterns may provide limited guidance.

Users should consider uncertainty when interpreting tail-risk scenarios.

Catastrophes.ai provides technology-enabled risk intelligence.

Information generated by our models does not constitute:

• Legal advice
• Regulatory advice
• Financial advice
• Insurance advice
• Scientific certification.

Users should consult appropriate professionals when making business, underwriting, or risk management decisions.

Users are responsible for:

• Evaluating whether model outputs are appropriate for their intended purpose
• Reviewing assumptions and limitations
• Applying professional judgment
• Maintaining appropriate governance processes.

Catastrophes.ai does not assume responsibility for decisions made solely based on model outputs.

Catastrophes.ai retains ownership of:

• Catastrophe modeling methodologies
• Machine learning models
• Algorithms
• Software systems
• Analytical frameworks.

Use of model outputs does not transfer ownership of underlying technology.