Independent Analysis of AI in Insurance

Insurance professionals are getting paid by the task to teach AI how to underwrite, adjust, and price, and some are training their own replacements without realizing it. Plus: an FT report co-authored with Anthropic and OpenAI researchers finds over 90% of insurers’ AI exposure is unpriced and hiding in conventional policies, Swiss Re puts a number on the AI infrastructure boom creating new insurable risk, veteran account managers describe how AI actually changed their jobs, and IBM’s worst trading day in 115 years ties back to a frontier AI capability release.

The pitch is always the same: AI will finally replace the clunky old AMS. It won’t. What it will do to the interface, and why the foundation underneath only gets more valuable, is the real story.

A new industry AI ranking crowns Allianz the leader, but the real finding is buried in the methodology. Half the sector’s best-documented AI results come from just five companies, and the rest are either doing less than they claim or saying less than they do. Plus: the Bank of England admits its rulebook wasn’t built for autonomous AI agents, Illinois becomes the first state to mandate AI safety audits, and OpenAI offers Washington a 5% stake in itself.

AI has moved from conference slides to agency operations faster than most predicted. This overview maps the current landscape for independent P&C agencies: what is actually available, which categories are mature enough to buy, what is not ready yet, and what to ask before you sign anything.

You didn’t pick your AMS or PMS for its AI roadmap, you picked it years ago for reasons that had nothing to do with artificial intelligence. This is a near-term look at what six major platforms, Guidewire, Duck Creek, Applied Epic, Vertafore, Zywave, and HawkSoft, are actually shipping versus still promising, and what to ask your rep before you assume “coming in 2026” means available to you.

Salesforce ties its AI pricing to actual resolutions, not activity. AWS puts $1 billion behind AI deployments measured in days. McKinsey lays out how AI agents are starting to shop for consumers, a shift with a direct parallel for independent agents. Plus, two very different approaches to a question every insurance AI buyer should be asking: how do you trust an answer from an AI you didn’t build?

The most dangerous thing in insurance right now isn’t AI. It’s the noise surrounding it.
Everywhere you turn for professional information: trade publications, conference main stages, consulting firm white papers, the AI conversation has been captured by two voices that have almost nothing useful to say to each other. One predicts the industry will be hollowed out by machines. The other promises transformation so complete that your legacy systems are already a liability.

Neither one is talking to you. Both are performing.

In this editorial, James W. Moore argues that the real problem isn’t capability. It’s the absence of a relevance test.

Satya Nadella warns that AI power concentration will not be politically tolerated. A Nobel Prize winner leaves Google DeepMind for Anthropic. Progressive ends State Farm’s 84-year reign as America’s largest auto insurer. And what Tony Robbins just did with his expertise should make every carrier rethink what happens when their best underwriters retire.

Most commercial agencies already do advocacy work. They just don’t price it separately. As AI compresses the cost of prospecting, placement, and routine service, that invisibility is becoming a business model problem. The question is whether the rules allow you to charge for it — and the answer varies more than you’d expect. Part 2 of the Distribution Economics Series maps the regulatory landscape across all 50 states, examines what the NAIC PR-70 chart actually shows, and looks at a compensation model that’s been operating legally across every jurisdiction for more than two decades.

The U.S. government pulled Anthropic’s two newest models with hours’ notice. Plus the real numbers on AI productivity gains, Meta’s workforce mistakes, and State Farm’s AI strategy.