Scout InsurTech Interview with Rohan Edwin

Rohan Edwin is Senior Product Manager - Data & Analytics at Ivans, where he helps create connectivity and more value with simple, streamlined workflows at every phase of the insurance lifecycle. Rohan was interviewed by Michael Fiedel, Founder and COO at PolicyFly and Co-Founder at Scout InsurTech.

Rohan, as carriers think about growth in the current market, why are so many prioritizing stronger relationships with existing distribution partners versus adding new ones?

“From what I've seen talking to a lot of carriers, the common message is that deepening an existing relationship carries lower risk than recruiting net new. Many larger carriers already have a significant number of distribution partners, so their mindset tends to shift toward identifying who they're underserving rather than who they should add.

Agencies' appetites change over time, and it's often more profitable to identify existing partners whose appetite is shifting to match your own. Sometimes it's as simple as reaching out to let them know: the business we see you writing is business we also write, so let's talk. That tends to yield better results than pursuing new partnerships, which come with more uncertainty, particularly regarding loss ratios.

There's also a data angle to this. A carrier might consider an agency one of their top partners, then look at the numbers and discover they're ranked seventh out of all the carriers that an agency works with. That's a signal. If you're not in the top three for a given agency, you're generally not getting the most sought-after policies. So rather than looking outward, the smarter move is often to ask: what can we do differently to strengthen this relationship and earn more of that preferred business?”

Everyone in insurance is talking about wanting more data, but what separates data that is truly useful from data that simply creates more complexity?

“The clearest pattern I've seen is that useful data is tied to a specific decision. If you can't connect a data set to a question you're actually trying to answer, it adds noise rather than clarity.

For distribution teams, the relevant question is typically: which agencies should we prioritize, invest in, and pursue? That means data that helps identify existing partners where more opportunity exists, or surfaces new agencies that closely match your appetite.

For product management, the question is around competitive positioning: how do my products compare to the broader market? The challenge there is having confidence in the scale and representativeness of the data. A sample or a predictive estimate is useful, but carriers are right to weigh it differently than true market-level analytics. That's the problem Ivans Insights is designed to solve, giving carriers access to market-level data they can actually trust and act on.

A partner of ours, Honeycomb Insurance, is a good example of useful data in practice. They built what they call an addressable book of business KPI, focused on their niche lines of business. They pull together agency size, policy types, and other available data, then filter it down to identify what portion of each agency's book actually falls within their appetite. A billion-dollar carrier writing only $10 million in your appetite is far less valuable to pursue than an agency with a $100 million book where $50 million aligns with what you write. Once they built out that KPI, they more than doubled their first 90-day onboarding revenue. That's a tangible outcome from getting the data question right.”

As carriers explore both AI and analytics, where do you see the greatest near-term value: helping teams learn, interpret, and act on data more effectively, or automating more of the decision-making itself?

“It really depends on who you're asking. AI companies will naturally emphasize the decision-making capability. Carriers tend to see it differently. They trust their experienced employees to make the final call, particularly in underwriting, and are more comfortable using AI to handle the preparatory work: pulling sources, compiling information, and getting it organized so the underwriter can focus on the actual judgment.

Where I see the clearest short-term win right now is in knowledge transfer. There's a significant workforce transition happening as experienced industry professionals retire, taking a lot of institutional knowledge with them. Insurance is a nuanced, heavily regulated industry, and new employees have always had to absorb an enormous amount of detail quickly, often without much structured mentorship from the people who know it best.

AI can help close that gap. It can compile that knowledge in an accessible way, accelerate onboarding, and get newer employees up to speed more efficiently. That's a lower-risk, higher-return application than full decision automation, and it addresses a real and immediate problem the industry is facing.”

As the industry looks for better ways to anticipate market shifts, how can carriers balance the need for early indicators with the reality that predictive models still require trust, transparency, and human judgment?

“The consistent message I hear from carriers is to start with what's observable. Bound policy data and quoting behavior in a given market segment carry more weight when informing pricing or distribution strategy than a predictive model alone.

The example people keep coming back to is COVID. No predictive model could have anticipated that level of disruption across commercial lines and personal auto. That reality keeps carriers grounded in what's actually happening in the market before they layer in projections about where it might go.

Predictive modeling still has a role, and AI will continue to make those models more capable. But carriers generally want to see the real-world signals first, then use modeling to add context and directional guidance. A product manager who can point to quoting behavior or bound data trends has a much stronger foundation for saying a market is softening, hardening, or holding steady than one relying solely on a model output. That combination of observable data and human judgment is where confidence comes from.”

Scout InsurTech Thanks Our Partners