We Didn't Start The Fire
Library  /  Insights

Commercial Lines Claims: The Reckoning Has Already Begun

14 consecutive years of Underwriting Losses demand more than Rate Action - the carriers who survive the next cycle will be those who transformed their claim operating model with AI-design at the centre!

Arijit Sarkar10 April 2026 · 2 min read

Commercial lines claims is no longer an operational function. It is a strategic battleground — and the carriers that have invested in AI, early resolution infrastructure, and IoT-led intelligence are building compound advantages that late movers will find structurally difficult to close.

$4.9B

Commercial auto underwriting loss 2024 14th consecutive year

103.5

Commercial auto combined ratio 2025 S&P Global

$45B

General liability losses 2024 — up from $18B in 2015

~7%

Annual liability cost growth — highest in two decades · Swiss Re

THE CRISIS

For fourteen consecutive years, commercial auto has posted underwriting losses. Social inflation is growing at 7% annually — the highest rate in two decades per Swiss Re Institute — with claim severity rising at 8% per year, nearly triple the rate of general economic inflation. General liability losses grew from $18B in 2015 to $45B in 2024. Commercial multi-peril liability posted a combined ratio of 114.9 in 2025. Products liability: 107.9. These are not isolated line-of-business problems. They are the simultaneous readout of a system under structural stress.

The commercial property market compounds the picture. Business interruption claim durations run from 30 days at best-in-class Tier 1 carriers to 365 days at Tier 3 industry average. Construction cost inflation at 7–9% annually, skilled trades shortages, and nuclear verdict proliferation are all running simultaneously. Rate increases have bought time. They have not bought a solution.

THE DIVIDE

The benchmarking data across 62 metrics tells the same story in every category. The gap between best-in-class and industry-average carriers is not one of degree. It is one of kind — and it widens every year.

Metric

Best-in-Class (Tier 1)

Industry Average (Tier 3)

AI utilisation in claims

70 – 85%

10 – 20%

Litigation rate

3 – 6%

20 – 30%

Litigation cost per claim

$5,000 – $25,000

up to $100,000

FNOL to auto damage payment

5 – 10 days

17 – 35 days

Business interruption duration

30 – 90 days

120 – 365 days

Post-claim client retention

95 – 100%

70 – 75%

Loss ratio

50 – 55%

65 – 85%

Reserve accuracy

95 – 100%

85 – 90%

THE IMPERATIVES

Five investments separate the carriers that will lead this market from those that will follow it at increasing cost:

1 Industrialise AI across the full claims lifecycle — Target 70%+ AI utilisation by 2027. Deploy GenAI workbenches, AI legal bill review, and 40–60% straight-through processing for commercial auto by 2029.

2 Build early resolution infrastructure before litigation funding colonises your portfolio — FNOL risk-scoring, pre-attorney settlement protocols for BI and GL, and structured mediation pipelines. The single highest-return investment in claims today.

3 Develop commercial EV and IoT claims infrastructure now, not in 2027 — OEM-certified EV repair networks and telematics integration at 70–80% (Tier 1 best-in-class) versus 20–30% (Tier 3 average) is the most actionable capability gap in the dataset.

4 Close the reserve accuracy and BI valuation gap — AM Best identifies $4–5B in commercial auto reserve deficiencies industry-wide. Forensic accounting partnerships and AI-assisted duration modelling are the difference between 30-day and 365-day BI claims.

5 Invest in post-claim retention before the policyholder makes the decision for you — A 20–25 percentage point retention gap between Tier 1 best-in-class and Tier 3 industry average is the premium revenue consequence of claims excellence — or its absence.

“The carriers who emerge from this era as structural winners will not be those who responded most dramatically to any single force. They will be those who saw the system changing most clearly — and invested with the most discipline.”
Follow on LinkedInDiscuss this with ArijitSee the working demos

More from the library