From AI Decisions to Measurable Risk.
From AI Decisions to Measurable Risk.
Insights from the Assessment of LLM-Assisted Insurance Underwriting
Insights from the Assessment of LLM-Assisted Insurance Underwriting
Underwriting relied on manual and time-consuming processes
Underwriting relied on manual and time-consuming processes
AI transformed it into a fast, efficient workflow
AI transformed it into a fast, efficient workflow
That shift raises new questions around trust, reliability, and compliance
Can we trust the model?
Can we trust the model?
Can we trust the model?
Is the summary correct?
Is the summary correct?
Is the summary correct?
Is this compliant with the EU AI Act?
Is this compliant with the EU AI Act?
Is this compliant with the EU AI Act?
AI assessment turns those questions into systematic tests
This example shows how robustness assessment reveals hidden risks in an AI-driven underwriting workflow
This example shows how robustness assessment reveals hidden risks in an AI-driven underwriting workflow
This example shows how robustness assessment reveals hidden risks in an AI-driven underwriting workflow
Rotated documents led to a sharp increase in extraction and summarization errors. As a result, medical conditions could be incorrectly omitted or added in underwriting summaries.
Rotated documents led to a sharp increase in extraction and summarization errors. As a result, medical conditions could be incorrectly omitted or added in underwriting summaries.
Identified weaknesses become actionable signals for system improvement
Identified weaknesses become actionable signals for system improvement
