Autonomous Underwriting, Generational Industry Shifts, and $82M in Regulatory Penalties: AI's Growing Footprint in Insurance

Legacy Systems and Retiring Experts: Insurance’s Twin Roadblocks to an AI Future

The Insurance Journal article examines how carriers can prepare for an AI-native future by addressing workforce transitions, legacy system barriers, and project-scoping challenges. Panelists noted that AI can distribute expertise across organizations, helping newer underwriters compensate for the retirement of senior decision-makers. However, outdated infrastructure remains a major obstacle, as legacy systems create data silos and consume half of IT budgets just for maintenance, leaving little room for innovation.

— Kimberly Tallon at Insurance Journal

URL: https://www.insurancejournal.com/magazines/mag-features/2026/03/23/862411.htm

$42M Bet on Insuring the Physical Backbone of the AI Boom

Shepherd, an AI-native commercial insurance platform, raised $42 million in Series B funding led by Intact Private Capital, bringing total funding to $67 million. The company has grown revenue more than sevenfold over the past 24 months, insuring over $400 billion in project value across more than 1,500 policies. Shepherd focuses on insuring AI infrastructure — data centers, energy assets, and semiconductor facilities — using real-time construction data to underwrite risk in minutes rather than weeks.

— Shepherd via PR Newswire

URL: https://www.prnewswire.com/news-releases/shepherd-raises-42m-series-b-to-power-the-insurance-behind-ai-infrastructure-boom-302723122.html

AI Without Explainability Is a Lawsuit Waiting to Happen

The Digital Insurance piece warns that insurers deploying AI without explainability are creating legal liabilities, not innovation. The regulatory landscape is fast-moving and fragmented: the NAIC issued its Model Bulletin on AI governance in December 2023, but only 24 of 50 states have adopted it, often with varying modifications. States like Colorado require insurers to test AI systems for unfair discrimination and report results annually. The article argues that carriers need governance infrastructure built into AI deployments from the start, not bolted on after the fact.

— Marcin Nowak at Digital Insurance

URL: https://www.dig-in.com/opinion/legal-issues-and-ai-compliance

This Week’s TL;DR

This week’s coverage centers on a familiar tension: AI adoption in insurance is accelerating on the capital and product side, even as the operational and regulatory foundations struggle to keep pace. Shepherd’s $42 million raise, and its sevenfold revenue growth over two years, shows real investor conviction that AI-native underwriting for physical AI infrastructure is a durable category, not a passing trend. That capital is chasing a genuine gap: traditional carriers still underwrite data centers and energy assets through processes built for a slower-moving world.

Inside incumbent carriers, the Insurance Journal panel’s warning is a useful counterweight. AI can distribute expertise and help newer underwriters compensate for retiring senior talent, but legacy systems that consume half of IT budgets just on maintenance are a hard ceiling on how quickly that promise can be realized. Data silos don’t resolve themselves because a new model gets deployed on top of them.

The regulatory picture adds a third dimension. With only 24 of 50 states having adopted the NAIC’s Model Bulletin, and meaningful variation even among adopters, carriers operating across state lines are working against a genuinely fragmented compliance landscape rather than a single national standard. States like Colorado, which require formal bias testing and annual reporting, are setting a bar that other jurisdictions haven’t yet matched. For risk managers and compliance teams, that unevenness is itself the operational risk: governance infrastructure built for the strictest state is the safer bet, even where it isn’t yet legally required elsewhere.