WaterStreet Company

05/25/2026

Today we pause to honor the men and women who gave everything in service to this country. Wishing you and yours a meaningful Memorial Day.

05/22/2026

A hallucinating doesn't fail visibly. It fails confidently.

No error message. No asterisk. Just a well-formed answer that happens to be wrong — and in P&C insurance, wrong looks a lot like a coverage dispute, a miscalculated premium, or a regulatory gap you can't explain to an examiner.

47% of enterprise AI users made at least one major business decision based on hallucinated content in 2024. That number doesn't get smaller when you add insurance-specific complexity to the mix.

We wrote a post for carriers and MGAs who want a clear-eyed look at what hallucinations actually are, how they surface in policy admin and underwriting workflows, and what governance architecture keeps them from becoming a liability.

Read the Post Here: https://hubs.li/Q04hK2Gs0

AI hallucination poses real risk in P&C insurance. Learn how they surface in policy admin, underwriting, and document workflows, and what governance stops them.

05/11/2026

Most P&C carriers and MGAs have heard both terms. Far fewer can explain the actual difference.

A system manages the full lifecycle of every policy you write. A calculates the premium. They work together, but they are built differently, owned by different teams, and break down in different ways when something goes wrong.

Confusing them is one of the most common and expensive mistakes carriers make when evaluating core technology. You can end up replacing a perfectly good PAS to solve a rating problem, or adding a standalone rating engine when the real issue is your policy workflow.

We broke down exactly how they differ, how they interact, and what it means for small and mid-size carriers deciding between embedded and standalone rating architectures.

A policy administration system and a rating engine are not the same thing. Confusing them leads to costly technology decisions for P&C carriers.

05/08/2026

June 1 isn't just a date on the reinsurance calendar.

It's the moment that determines whether a Florida homeowners carrier can survive a major storm season.

We broke down what's actually driving the 6/1 renewal in 2025: FHCF retention shifts, ILS capacity hitting $114B, cat bond issuance up 39% year over year, and why small and mid-size carriers now represent 35% of the cat bond sponsor market.

If you write wind-exposed business, this is your pre-season read.

https://hubs.li/Q04f-tP90

The June 1 reinsurance renewal defines hurricane season for U.S. property carriers. Here's what's driving pricing, capacity, and cat program structure in 2026.

05/05/2026

Outsourcing your to a vendor does not outsource your compliance obligation.

That is one of the harder truths buried in the NAIC's Model Bulletin and one of the most overlooked by carriers relying on third-party models for , pricing, and .

We broke down the full documentation picture in our latest post: what regulators are looking for, what you need to have ready, and why building it now is significantly cheaper than building it under exam pressure.

Insurance AI regulation is accelerating. Learn what documentation your carrier needs in place before the next market conduct exam puts your governance on trial.

05/04/2026

What does it take to absorb 60,000 policies in a week?

That's what some Florida takeout carriers did in 2025 as dropped from 1.4 million policies to under 400,000 in two years. The carriers that thrived had more than capital. They had the operational infrastructure to back it up.

We broke down what actually separates the prepared from the unprepared: tech stack, staffing, and compliance readiness for takeout volume.

If you're watching Florida and thinking about what it means for your market, this one is worth the read:
https://hubs.li/Q04f7PXP0

Florida depopulation pushed 546K+ policies into the private market. Is your carrier operationally ready? Learn what takeout carriers need to succeed.

05/01/2026

Florida's P&C market just posted its first profit since 2016. Litigation is down 35%. Rates are dropping. And 16 new entities were approved by in January and February alone.

The window is open.

The carriers moving fastest will need the infrastructure to back it up.

New post on what that actually looks like:

The Florida insurance market is rebounding fast. Learn what the data says about new carriers, Citizens depopulation, and what it means for your infrastructure.

04/17/2026

"We'll figure out AI governance once the federal picture clears up."

That's the wrong bet.

State AI laws are in force now. The NAIC AI examination tool is live across 12 states in 2026. Nearly 1 in 3 insurers still aren't testing their models for bias. Examiners know it.

Federal preemption won't protect carriers from class action exposure. And it's years away at best.
The carriers building governance programs now aren't checking a box. They're building advantages their competitors don't have yet.

We wrapped up our 6-part series on insurance AI regulation with a look at what to build regardless of how the federal debate resolves.

Read it here: https://hubs.li/Q04cz8Gk0

Insurance AI governance isn't waiting on federal clarity. Here's what carriers must build now to stay compliant across state and NAIC frameworks.

04/16/2026

Over half of U.S. states have adopted the NAIC Model Bulletin on insurance AI and the examination infrastructure to enforce it is being built right now.

A 12-state pilot of the NAIC's AI Systems Evaluation Tool launched in January 2026. When it becomes standard practice, examiners will arrive with a structured checklist. Carriers without documented AI governance programs will feel that gap.

Read the full post here: https://hubs.li/Q04cqp080

The NAIC Model Bulletin on insurance AI is live in 25+ states and examiners are coming. Here's what carriers need in place before they do.

04/13/2026

New York just raised the bar for in .

DFS Circular Letter 2024-7 does not just ban protected characteristics in AI models. It requires insurers to prove their models are not producing discriminatory outcomes through indirect variables either.

Our latest article breaks down exactly what the test requires and what compliance actually looks like in practice:

DFS Circular Letter 2024-7 sets strict rules for insurance AI use in underwriting. Learn what the proxy test requires and how to stay compliant in New York.