Insurers Pivot AI Investment From Claims to Underwriting

Artificial intelligence in insurance is entering a second phase. After years of automating claims workflows, carriers are now deploying AI systems that assess and monitor risk continuously throughout a policy's duration, according to research reported by Finextra.

The shift reflects both the maturity of claims automation and the industry's recognition that preventing losses delivers greater value than processing them faster.

Claims automation delivered early wins

Claims became AI's first major foothold in insurance because the process involved structured data, visible bottlenecks, and measurable outcomes. Carriers used computer vision to analyze damage photos, deployed fraud detection algorithms across claim portfolios, and built mobile-first notice-of-loss systems that eliminated phone queues.

A 2025 National Association of Insurance Commissioners survey of 93 health insurers found that 84% already use AI or machine learning, with fraud detection among the most common applications, Finextra reported.

Insurers that invested heavily in claims automation increased straight-through processing rates from roughly 10–15% to 70–90%, according to the report.

Risk automation targets underwriting and monitoring

While claims AI responds after a loss occurs, risk automation evaluates exposure before and during coverage. Instead of a single underwriting snapshot at policy inception, AI systems now ingest ongoing data streams to track changing conditions.

These agentic underwriting systems pull from IoT sensors, satellite imagery, telematics devices, and third-party databases to maintain continuously updated risk profiles, Finextra reported. Vehicle condition, building maintenance, and driver behavior all evolve over time, and AI is designed to flag material changes.

Computer vision converts photos and videos into structured underwriting data. Pre-policy vehicle inspections, roof scans, and fleet images produce documented condition reports within minutes. Predictive analytics combine historical loss patterns with current conditions to generate forward-looking risk scores, with industry case studies reporting underwriting accuracy improvements up to 54% in certain applications, according to the report.

Telematics and IoT sensors provide the continuous data streams that make ongoing monitoring feasible at scale.

Underwriting cycle times compress sharply

AI improved underwriting data intake accuracy from roughly 75% to more than 90% in certain workflows, Finextra reported. For commercial fleets, continuous vehicle monitoring delivers ongoing risk signals. Research cited in the report indicated that AI-driven underwriting improvements contributed to an approximately 3-percentage-point reduction in loss ratios for business lines incorporating unstructured vehicle condition data into pricing.

AI reduced underwriting cycle times by 31% and improved risk assessment accuracy by 43% for complex policies, according to the report. In some automated workflows, underwriting timelines dropped from three days to three minutes.

Insurers using advanced analytics achieved combined ratios approximately six points lower than slower adopters between 2022 and 2024, Finextra reported. Insurance AI deployments increased 87% year over year, with agentic AI systems accounting for roughly one in five public deployments.

Why it matters

The migration of AI investment from claims to underwriting signals a strategic evolution. Claims automation improves customer experience and operational efficiency, but risk automation directly affects loss ratios and pricing accuracy. Continuous monitoring enables dynamic pricing models and early intervention, transforming insurance from a reactive product into a risk management service. Carriers that master continuous risk assessment gain competitive advantages in both underwriting profit and customer retention.

These findings were first reported by Finextra.