F

Frederick Mutual Insurance Company

Frederick Mutual Insurance reduces loss ratio by 26.8% using AI-powered aerial property inspection

57% improvement at product launchDirect Loss Ratio Improvement
35.6%Average Loss & LAE Ratio (5-year)
26.8%Year-over-Year Loss Ratio Reduction

The Challenge

Founded in 1843, Frederick Mutual Insurance Company (FMIC) is one of the oldest P&C insurers in the U.S., serving commercial lines across five states and Washington D.C. Despite its long history, FMIC faced a growing gap between its underwriting practices and the realities of modern property risk. Roof condition — a primary driver of property claims — was impossible to assess without dispatching physical inspection teams, consuming significant time and budget. An uptick in property claims compounded the problem: without a reliable source of truth for pre-existing damage, adjusters struggled to distinguish legitimate losses from pre-existing deterioration or potential fraud, directly eroding loss ratios.

The Solution

In 2019, FMIC adopted the Betterview platform — a property intelligence solution that applies computer vision to high-resolution aerial imagery — to remotely pre-inspect every piece of new business before binding. The deployment standardized underwriting with the Roof Spotlight Index (RSI), a 100-point AI-generated score that surfaces the most material risk drivers. Across FMIC's portfolio, the top three detections were roof staining (47.5%), tree overhang (36.4%), and vents (19.1%). Bulk processing generated detailed reports across the entire book simultaneously, while a historical imagery archive gave claims teams timestamped visual records to verify when damage actually occurred — including isolating CAT-event losses for reinsurance recovery. A customizable flagging engine further automated triage for low-risk straight-through processing.

Results

Following the product and rating structure launch built around Betterview data, FMIC improved its direct loss ratio by 57%. Sustained over time, the platform helped the company maintain an average loss and LAE ratio of 35.6% over five years — a 26.8% year-over-year reduction — even as severe weather frequency increased industry-wide. Operationally, the outcomes included:

  • Straight-through processing of low-risk properties, reducing manual review burden
  • Optimized allocation of physical inspections to genuinely ambiguous cases
  • Measurably reduced fraudulent claim payouts through historical imagery verification

CEO Nancy Newmister cited the platform as transformative across both underwriting and claims functions.

Key Takeaways

  • Pre-inspecting every property at new business submission — before binding — eliminates on-site inspection costs while raising underwriting accuracy from day one.
  • Historical aerial imagery functions as an objective claims record, enabling adjusters to determine whether damage predates the policy term or qualifies as a CAT-event loss for reinsurance.
  • A standardized, quantitative roof score (like RSI) creates consistency in underwriting decisions and gives agents a concrete basis for policyholder conversations.
  • Bulk processing capability is essential for portfolio-level deployment; point-in-time lookups alone will not deliver the loss ratio improvements seen here.
  • Early adoption of property intelligence platforms provides compounding advantages — FMIC's five-year average ratio reflects sustained, not just launch-year, gains.

Share:

Details

Industry
Reinsurance
AI Technology
Computer Vision
Company Size
SME
Quality
Verified

Have a similar implementation?

Share your customer's AI results and link it to your vendor profile.

Submit a case study →