Commercial Insurance Embraces AI While Preserving Human Oversight

Artificial intelligence is fundamentally transforming how commercial insurers evaluate, price, and manage risk—with decisions that once took weeks now completed in minutes—according to research from Lockton examining AI’s growing impact on the P&C industry.

The shift reflects converging industry pressures. Data availability has exploded through advances in IoT sensors, telematics, and aerial imagery. Computing power has accelerated through cloud infrastructure improvements. Meanwhile, escalating losses, talent shortages, and policyholder expectations for speed and transparency have pushed insurers to seek new operational approaches, according to Lockton.

The market is responding with three distinct adoption models:

• “AI Innovators” are deploying the most automated systems, issuing binding quotes in seconds for high-volume, low-complexity lines like personal auto and homeowners.
• “Measured Adopters” are applying AI to data ingestion and risk scoring in middle-market and specialty segments, enabling underwriters to focus on exceptions and complex accounts.
• “Cautious Followers” are using AI primarily for decision support—flagging anomalies and providing pricing benchmarks while keeping final underwriting decisions in human hands.

Promising Opportunities and Mounting Risks

For insurers, AI offers strategic advantages: compressed quote times, improved portfolio segmentation, reduced expense ratios, and expanded risk appetite, according to the report. By analyzing vast datasets, AI can identify patterns and tailor pricing to specific risk profiles, potentially leading to more resilient portfolios and competitive advantages.

Brokers stand to transition from transactional intermediaries to strategic advisors, leveraging benchmarking tools and predictive analytics to guide clients toward optimal programs. Insurance buyers gain access to continuous risk tracking, improved predictive modeling, and enhanced exposure identification—benefits that extend beyond cost savings to include better resource allocation and informed operational controls.

However, significant challenges accompany these opportunities, Lockton said. Bias and discrimination concerns loom as a critical issue, particularly when flawed data or poor model assumptions lead to discriminatory outcomes. The “black box” problem persists—without transparency around data sources and model logic, AI-driven decisions may erode trust.

Evolving regulatory frameworks are escalating scrutiny around privacy, fairness, and explainability. Additionally, reliance on biased or flawed models could cause organizations to underinsure critical exposures or misallocate capital, exposing leadership to potential claims of negligence.

Integrating Human Expertise Into AI Systems

The industry’s success hinges on pairing technological capabilities with human judgment, according to Lockton. Traditional underwriting, actuarial, and analytic skills remain essential, but the workforce must evolve. Insurance and risk professionals need fluency in data interpretation, model logic, and digital workflows.

Underwriters must understand how AI models are trained, what assumptions they rely on, and how to challenge outputs effectively. For brokers and buyers, this means developing capabilities around algorithmic assumptions, data sources, and technological literacy.

View the full report here. &