Recently, I spoke with a chief product officer at a large excess and surplus (E&S) company, and he shared his vision for streamlining intake in this way: “Straight-through processing a third, decline a third, and send the rest to underwriting.”

But adopting that kind of intake strategy naturally leads to a deeper question: How does an insurance company decide what deserves underwriting consideration in the first place?

It would not be unusual to see agent territories, spreadsheets from last quarter, and decades of institutional knowledge from the distribution team coming to bear in that discussion. But increasingly, machine learning projects are surfacing quietly in the background, connecting internal data to real-time risk signals and pointing the way toward smarter intake.

According to a 2024 ICMIF study, nearly two-thirds of mutual and cooperative insurers are already using artificial intelligence (AI) in operations, with machine learning (ML) cited as the most common application. That momentum reflects a strategic shift toward pragmatic tools that can enrich submissions, score leads and reduce friction without needing to rip out legacy systems.

What many insurers still lack, however, is a tightly integrated pipeline that scores submissions at intake, before they ever hit an underwriter’s desk. That gap between strategy and execution creates inefficiency, and for mutuals especially, it can delay profitable decision-making. But mutuals are well-positioned to close that gap. The policyholder alignment, cleaner datasets and lighter tech stacks inherent to many mutuals make them ideal candidates for predictive modeling, especially when the goal is clarity at the top of the funnel, not disruption for its own sake.

__

Additional Context from Fenris

What stands out in this moment for mutual insurers is not experimentation with machine learning, but where it is delivering measurable value.

Across the market, successful ML initiatives tend to focus on improving decision quality at the earliest point in the workflow. Rather than attempting broad transformation, teams are using predictive models to clarify intent and risk at intake, before submissions consume underwriting capacity.

This approach aligns well with how many mutuals already operate. Strong policyholder alignment, established agency relationships, and a preference for transparency over black-box automation make early decision support a natural fit.

In practice, this shift shows up as:

• Earlier clarification of which submissions should flow straight through, be prioritized, or be declined

• Better use of underwriter capacity by reducing low-intent or misaligned submissions

• Predictive signals that support human judgment rather than replace it

The opportunity ahead is less about adopting AI for its own sake and more about closing the gap between strategy and execution. When intake, prioritization, and underwriting decisions are informed by the same real-time signals, insurers can move faster without introducing unnecessary disruption.

This is the transition many mutual insurers are navigating now, moving from gut feel toward guided growth, with machine learning serving as connective tissue between data, intent, and execution.

__

For a deeper look at these themes and how mutual insurers are applying them in practice, read the full article via Carrier Management: https://www.carriermanagement.com/features/2026/02/10/284373.htm?bypass=8b96db6c7b43b5a3d6677a8a32e6db4f