Scout InsurTech Rising with GenAirate

Yogesh Joshee, Founder and CEO at GenAirate, and Robert Maycock, Head of Product at GenAirate, were interviewed by Michael Fiedel to discuss the true scale of the unstructured submission problem, what they're hearing from MGAs and Tier 1 carriers, the story behind GenAirate's founding team, and what's next in property risk intelligence. GenAirate's SnapLine platform transforms unstructured broker submissions into clean, decision-ready data for underwriters.

Yogesh, that unstructured submission problem you started with – how big does it actually turn out to be?

Yogesh: When we first started talking to underwriters about processing submissions, it looked like a manageable problem. On the property side, we thought we needed maybe ten fields: construction type, occupancy, protection, exposure. Those ten fields quickly turned into 86 just on property, and today our system tracks over 2,000 fields across the lines of business we support.

The deeper we dug, the more we found. Handwritten submissions, some of them nearly impossible to read even for a person. We built models that can now read handwriting better than we could ourselves. Then came checkboxes, embedded diagrams in PDFs, and every other format underwriters actually work with. What we learned is that unstructured data is not just an insurance problem. Healthcare, manufacturing, banking, they all sit on huge amounts of it. It has built up organically over 25 to 30 years of people storing things in SharePoint and elsewhere. Getting real value out of it requires understanding context and structure, and the scope of that challenge turned out to be much larger than we expected.

Robert: We are taking many sources like applications, emails, engineering reports, third-party data to create a mosaic of an insured or building structure. Structuring this data is essential to decision-making.

This mosaic-creating process is happening at every stage of the value chain. This is how a contractor in Texas is eventually underwritten by retro markets in Bermuda.

Not only each carrier, wholesaler and reinsurer will want the data structured their own way, but also teams within an organization. Underwriting, cat modelling, and pricing have their own preferences for how they want to see the data. The size of the problem explodes when you view it this way.

We are creating the connective tissue between all of these stakeholders. 

While data challenges are nothing new to the industry, I’ve been kicking around a fun analogy that may reframe how we think about them. When I was growing up, myofascia – the body’s connective tissue – was not more than a footnote in my science curriculum. It was viewed as the body’s casing, holding together the important bits like muscles and organs. In the last decade, myofascia has become a serious area of study. We now know that it provides structural integrity and has its own nerve endings. Myofascia work is now an essential component of every elite athlete’s program. 

In a similar way, we help organizations achieve high performance, even if it isn't the most visible part of the system.

What are you learning from the conversations you're having with the MGA and Tier 1 carriers?

Yogesh: We started with an efficiency lens, asking underwriters what their day actually looks like. We found that roughly 60 percent of their time went to copying, pasting, and moving information rather than actually evaluating whether a risk was a good fit for their portfolio.

Coming into insurance, I already knew the industry had a reputation for resistance to change. The attitude was often, we've done it this way for 200 years and our instincts are good. It reminded me of watching Moneyball recently. When Billy Beane brought analytics into scouting for the Oakland A's, the old guard pushed back hard, insisting that gut feel about a player mattered more than the numbers. He proved otherwise, and it changed the game. I think underwriting is heading in a similar direction. The people doing this work have done a great job for decades, but there is real room for improvement, and the industry is slowly coming around to it.

Robert: What we hear consistently is that our clients want to empower their people. More than at any point in my career, everything seems to be on the table: the tools, the workflows, the org structure. There's real openness at the leadership level to rethink how work gets done.

Two themes stand out. First, throughput: handling simple, high-volume submissions efficiently while routing the genuinely complex cases to the humans who are good at operating under uncertainty. Second, there is no single solution that fits every carrier. We meet each client where they are, with whatever combination of tools and workflows they already have. We hear that some of our competitors assume every carrier has a clean standard operating procedure to build around. In our experience, that's rarely true. It's more fragmented than that, and there's often real uncertainty between leadership and the teams doing the work. Our job is to get in there and work directly with people to drive better performance, less like a traditional software sale and more like a hands-on partnership.

Where did your founding team come from, and what made you bring those specific people together?

Yogesh: I started this a couple of years ago after spending nearly 30 years in technology, mostly in financial services and banking. This is my third startup. The first was in software development, the second in sports, and this time the goal was to focus on one industry and one specific problem rather than try to solve everything at once.

Robbie is my co-founder and came from the insurance side. We also built a team of PhD-level AI engineers and technologists, but we were deliberately top-heavy on insurance expertise from the start. One of our earliest collaborators was an underwriter who does peer reviews in the Lloyd's market and has 50 years of experience reading submissions. He's completely non-technical, he wouldn't know how to use an iPhone, but he helped us build our first rule set because he understands what a good submission looks like better than almost anyone.

Robert: I spent the last decade in P&C, including time on the global innovation team at AXA XL, which gave me a rare view across the different silos of a large multinational carrier. That's where I got firsthand exposure to the ingestion problem, how foundational it is to everything that follows, and how even the most sophisticated carriers haven’t really solved it. Meeting Yogesh was a bit of serendipity, and it gave us the chance to go after a problem that I'd seen up close and knew was genuinely meaningful for the industry.

What excites you most about the property risk intelligence work you're pushing into next?

Yogesh: We started by enriching submission data with things like proximity to the nearest forest fire risk or the last recorded earthquake using geolocation. We've since gotten much more sophisticated. Every property has a story: its age, its location, how long a fire engine would take to reach it, its square footage, its valuation, who else owns it, and what other exposures sit nearby. We're now able to piece that story together in real depth, and a lot of that progress comes down to the talent we've been able to attract to this space.

Robert: I'd echo that, the people are what make it possible. Property risk intelligence is a natural extension of what we already do: enriching and verifying submission data using third-party sources, and helping drive decisions. 

Property excites me for a few reasons. On a societal level, something like 60 percent of property risk globally is uninsured. Efficient and informed decision-making can close the gap. On a business level, it's a huge part of the industry where we can have real impact. And it happens to be the area I worked in most directly at AXA XL. On the technical side, what's possible today wasn't possible even a couple of years ago. Traditional machine learning approaches to reading building attributes relied on huge labeled datasets to detect something as simple as a solar panel. Newer technology creates much richer data sources, and when you pair that with agents capable of consuming that information, you get a far more complete picture to base decisions on. It's a logical next step, and one we're super excited about.