Meet the AI Underwriter

Three years ago, Sixfold launched the first AI product built for insurance underwriting.

Since then, we’ve learned what it takes for AI to earn a place in an underwriter’s workflow. It has to do more than read and summarize a submission. It has to support judgment in context.

We learned this by processing more than 1.5 million submissions across more than 50 lines of business for carriers including Zurich, Skyward Specialty, Generali GC&C, Guardian, AXIS, and New York Life in North America, Europe, South America, and Australia.

A great underwriter holds an enormous amount in their head. The risk. The carrier’s appetite and guidelines. The broker’s history. The shape of the portfolio. Similar cases they have seen before. Where the carrier is willing to bet.

Even the best underwriters hit two limits.

First, no one can hold their entire book in memory. Every interaction with a broker. Every similar risk. Every detail of the underwriting guidelines. It’s too much information to carry and recall for every decision.

Second, underwriters mostly learn from the cases on their own desk. What the underwriter next to them knows often stays there. Institutional knowledge that should compound across a team doesn’t.

Today, we're launching something new to solve both. The AI Underwriter.

From Submission to Recommendation

The first version of Sixfold was about providing relevant risk insights. Since our launch, our AI has been exceptional at ingesting submission data straight from an inbox, enriching it, filling in the gaps, and pointing to what matters based on appetite fit. But, to solve the deeper problems preventing underwriters from binding more and better business, we needed to go further.

The AI Underwriter does exactly that. It’s an agentic colleague that receives a submission and produces real underwriting work product as the output, one that understands the risk, knows the broker, holds the portfolio in memory, and tells the underwriter what to do next.

The agent receives a submission, extracts and cleans the data, identifies what's missing, and evaluates appetite and portfolio fit. It reasons about the risk, but also about whether this risk, from this broker, belongs in this book right now. It draws on everything the carrier's team has ever learned about that broker and that risk class, surfaces portfolio context, and produces a recommendation and the rationale to back it up. The next step is always clear, whether that's a referral, a declination, or a broker follow-up drafted and ready to send.

At the center of the AI Underwriter is what we call the Underwriting Brain. It brings functional knowledge across lines, reasoning patterns from experienced underwriters, and a ground-truth library of what good underwriting looks like. It doesn't need months of ramp time. It already knows how to underwrite on day one.

Most AI Tools Digitize The Submission. This One Does the Work.

How is it different from other AI tools out there? The reality is that a great underwriter doesn't need a digitized version of the application or a dashboard with data that was current last week. They need to understand the risk in the full context of their current book, the broker relationship, and their guidelines, and what the next best action on the case is.

That's what the AI Underwriter does. And it's something no text ingestion, triage, or workflow tool was built to do, because none of them were designed to learn from the work itself.

After a short time with a carrier, the AI Underwriter holds institutional memory that compounds with every submission. The broker's history. Prior outcomes on similar risks. Portfolio context. Everything every colleague has ever learned, available the moment the next submission lands.

The Impact on Your Book

Sixfold customers have historically seen 50% to 97% faster processing, hit ratios up 15%, and 30% more GWP per underwriter. One carrier went from pilot with a handful of users to 500+ underwriters, reaching go/no-go decisions on submissions in 5 minutes and more than 2 hours saved per submission. Another one cut days-to-quote from 16 to 7 days, with more than 100 underwriters using Sixfold today.

Matthew Richardson, Global Head of Operations & IT, from Generali GC&C, summarized the impact well:

“Sixfold helps us quote risks in minutes instead of days or, in some cases, weeks. The work gets better for our underwriters. What sets Sixfold apart is how they build their product, alongside underwriters, shaped by how the job really works. That is why we believe they are building the future of underwriting.”

The AI Underwriter moves insurers further toward a better combined ratio and more profitable growth. Expense ratios fall as underwriters focus on judgment instead of analysis. Over time, loss ratios improve as fewer good risks get missed and appetite alignment tightens.

For a startup MGA writing complex specialty risks with a small team, it makes growth possible without hiring ahead of revenue. For a global carrier, it means every submission is evaluated the same way, with more context, and with better insights from your best-performing distribution channels.

What We're Hearing from Users

Before today's launch, a select group of insurers worked with the AI Underwriter on live submissions. At Skyward Specialty, early results show that the AI Underwriter performed the majority of the underwriting analysis needed to evaluate submissions, helping underwriters quote better risks, faster, with more consistent judgment across the book.

The feedback we’ve received speaks for itself. Melissa Butt, Vice President E&S Brokerage, Property at Skyward Specialty, said it best.

"The AI Underwriter does not just summarize what you are looking at, it gives you a point of view and a clear path forward. For more junior underwriters, it’s a powerful companion that gives them the perspectives of a more experienced underwriter and accelerates their development. Even veteran underwriters who are a little more skeptical of new technology see the AI as an added layer of intelligence. Plus, the ability to fully examine a submission and make better, more confident decisions without even leaving the tool is a huge benefit. The entire team is excited to work with it."

AI Colleague. The Underwriter in Charge.

I want to be very clear. Underwriters are not going anywhere.

But, we’re not automating the existing underwriting process either. We’ve invented (and patented) a new way to run a risk-taking shop.

The best underwriting teams will not remove human judgment. They will focus human judgment where it matters most. Complex risks. Exceptions. Broker relationships. Negotiation. Accountability. Where to take bets.

We recently surveyed more than 500 underwriters and underwriting leaders across North America and Europe, and nearly 80% said they wanted more time for exactly that work.

The AI Underwriter creates that time.

It does the analytical work. It surfaces what is missing. It explains its reasoning. It recommends the next action. The human underwriter directs it, corrects it, decides, and remains accountable.

That is the partnership we believe will define underwriting over the next five years.

And today, it is here.

That’s better #funderwriting.

This post was originally shared on LinkedIn.