Otto

Overview

Commercial insurance in the US is a massive $400B+ market. For the last decade, most startups seeking a share of this market have tried to disintermediate independent agencies and sidestep agents by developing DTC insurtech offerings.

Despite this pressure, independent insurance agencies and agents remain remarkably resilient, accounting for roughly 87% of all commercial written premium in the US.

However, the resiliency of independent insurance agencies depends on organic growth, and that growth depends on agents who can identify the right-fit businesses, diagnose risks effectively, and quickly place them with the right carriers. Today, though, the process remains complex, cumbersome, and knowledge-intensive.

Otto is an AI sales solution that automates prospecting, risk research, and carrier matching for commercial insurance agents: it compresses the days and weeks of manual effort between "I have a lead" and "I have a submission an underwriter will take seriously."

I led Otto from problem identification through closing the first 25 agencies to the seed commitment.

The Challenge

The insurance industry pays for its sales complexity in people, and the numbers tell a brutal story. Between 50% and 80% of new commercial agents fail within their first few years, crushed by a role that demands both technical skill in assessing complex business risks and relentless prospecting.

Industry research puts the average time for a new producer to become "validated" (to earn more in commission than they cost the agency in salary) at almost 3 years. That’s a long, hard road for new agents, and for an agency with under $2.5 million in commission income, which describes 85% of all agencies, the investment in a new producer can be risky; a producer who fails can have a serious financial impact on the agency.

Sales complexity and steep learning curves don’t just hurt new agents. Established agents stall, too, when they try to expand into unfamiliar business classes, keep up with changing carrier appetites, and balance client servicing with new-business prospecting. Without a solution to reduce complexity, streamline knowledge transfer, and cut administrative tasks, agencies will keep struggling with producer success, organic growth, and, ultimately, long-term viability.

The Solution

Otto is pioneering the sales intelligence category within the commercial insurance industry. Unlike agency management systems, which are systems of record for policies, Otto is a system of action, an agentic sales solution for the pre-binding workflow: where a producer finds a business, understands its risk, decides which carriers will write it, gathers what the underwriter needs, and tells the client’s story. That is where producers have the greatest impact on agency growth, where the messy-inbox problem lives, and where today, the least software exists.

Otto works the way a producer works. It finds businesses that match the producer's appetite and expertise, and keeps the pipeline full through demanding service cycles. It assembles each prospect into a complete profile of the business, the property, and its exposures, with an estimated premium and the math behind it. It matches the risk to the carriers likely to write it, with the reasons attached, and helps draft the story the underwriter needs to hear, so commercial producers place business faster, across more classes, and with a higher likelihood of carrier acceptance.

The Approach

We didn't build Otto in a straight line. We moved through the process in a deliberate loop, figuring out what decision the next stretch of work needed to produce, doing the smallest thing that could produce it, learning from what came back, and going again. Some loops closed in a few days. Others took weeks. What mattered was that we kept moving forward toward the next decision.

Discovery ran for four weeks and started with desktop research. We gathered qualitative and quantitative market information to define the problem space, including dozens of industry reports, articles, and insurance industry documentation and training materials.

We delved deeply into the industry to understand the broader solution landscape and identify areas where the market is fragmented, overly reliant on legacy systems, or where new tech could improve the cost structure or customer experience. We also established early hypotheses on what a future solution would need to displace, i.e., Excel, people, and incumbent solutions, by building out a coverage and gap map that identified areas across the workflow where different solutions played and fell short.

We conducted 25 interviews with subject-matter experts and producers across personal and commercial lines to understand their contexts, problems, and desired outcomes. We then mapped the commercial agent's workflows and experience stage by stage to identify pain points and opportunities.

In parallel, we conducted multiple technical explorations. We created early agent architectures that focused on collecting business and carrier appetite data to determine what data existed and where it was available. Knowing early on that the data existed and was messy let us make product bets with more confidence.

Once we synthesized our learnings, we ran a two-day business sprint and used it as a forcing function to sharpen our thinking on the broader business model, who to build for and why, and what to test and learn to de-risk the idea and make decisions. The goal moving forward wasn’t to test and learn; it was to gain insights and decide. We came out of those two days with an early shape of the product we could test quickly and effectively.

From there, we ran as fast as we could, running one-week cycles of building and testing with customers. At the same time, we were actively developing a group of design partners. We built a group of excited early users that tightly matched our ICP and provided product feedback. These relationships fed our early sales pipeline by providing introductions, credibility, and our first paid customers.

Decisions

Decisions we made with high levels of ambiguity and risk that helped us learn, ship fast, and move forward

Build for the producer, not the back office.

The back office had obvious automation targets. However, the producer's complex, messy sales process was where agencies made or lost money, where tooling was thinnest, and starting at the top of the funnel allowed us to build a valuable wedge and potentially earn the right to claim space down funnel over time.

Ride the frontier wave, let go of old patterns.

We believed early on that it was better to build toward where the models were going than hold on to old building patterns. We used AI to accelerate the front-end and back-end build, and over very short periods, we saw improvements in our ability to lean on the models rather than data partners to improve the product's research performance. Building with model momentum helped us capitalize on intelligence, flexibility, and speed.

Build Trust, and identify where accuracy actually matters.

Trust is non-negotiable in highly regulated industries, but accuracy is more situationally dependent, and that’s an important difference to remember. We allowed customers to run head-to-head tests against real premiums, provided cited sources, editable recommendations, and human approvals, which built trust. However, first, we dug deep to understand where accuracy mattered and where it didn’t, and it turns out there were more places than we thought it didn’t, which saved us time and resources.

Empower agents, don’t replace them.

The commercial insurance business is built on trust and connection. We quickly learned that agencies have unique workflows, and agents see their value as trusted advisors. We decided early on that Otto should keep the producer at the center of everything we built, and adapt to an agency's needs rather than force standardization. This focus established credibility and advocacy with users.

Validate with practitioners, but sell early.

Building a design partner advisory group helped us develop accuracy benchmarks, a prioritized roadmap, and a group of excited advocates. Closing the first twenty-five agencies personally meant gaining the benefit of hearing every objection, and each one influenced a positioning decision or product call.

Results

Commitment of $3M seed round