Pie OS: From $10m to $350m in 5 Years

Over the span of five years, I led the design of Pie OS, Pie Insurance’s proprietary insurance policy lifecycle system. This foundational effort scaled the company’s operations from $10 million to over $350 million in total written premium, while ensuring Pie retained full control over its technology and processes.

TL;DR

• I designed Pie’s Partner Portal enabling 5-minute quotes, instant binding, and policy servicing—driving a 2900% increase in submission volume and 900% growth in appointed agencies over 5 years.


• I simplified complex linear rating algorithms into a single-page interface for Pie’s Underwriter Rater, boosting underwriter efficiency by 3x.


• I scaled Pie’s Policy Management System with a modular UI architecture, cutting $1m in annual costs and enabling multi-state, multi-line, and multi-carrier growth.

Pie OS Demo Video

From 0 to 1: Parter Portal

Pie was quickly onboarding agents but using the equivalent of a Google Form to intake submissions. Each appointed agency had a unique URL, there was limited auto-decisioning, and once an application was submitted the process migrated to email inboxes. It was a logistical nightmare for everyone. My assignment was to design Pie’s online quoting portal.

Starting in office. Working closely with in-house underwriters and business development managers, I immersed myself in the insurance submission process, uncovering what underwriters prioritize, the agent journey, and the nuances of the competitive landscape. This deep understanding became the foundation for designing the v1 prototype of the agency flow.

Testing with users. Engaging directly with 10+ agents over two weeks, I conducted virtual sessions to test and iterate prototypes, gather insights, and rapidly deepen my understanding of workers' comp insurance. Their invaluable feedback—including references to top competitors and examples of their workflows—shaped the design iterations, ensuring the platform addressed their needs before moving into development.

Shaping the MVP. I collaborated with the Product Manager to translate insights from internal and external testing into a focused feature set, defining fast-follows and a long-term roadmap. The initial release in April 2020 introduced key features: an enhanced quote flow enabling over 50% auto-decisions, an appetite checker to assess submission eligibility, and a dashboard for tracking submissions and policies.

In the years that followed, we introduced critical enhancements, including post-submission application editing, online policy binding, payment capabilities, and multi-line functionality, solidifying Pie’s Partner Portal as one of the most favored quoting systems among agents.

Partner Portal User Testing Readout.

Redefining Risk Assessment: The Underwriter Rater

The next step was addressing inefficiencies in underwriting workflows. At the time, underwriters relied on Insurity, a third-party policy management system that was cumbersome and inadequate for Pie’s needs. The new system needed to meet the needs of underwriters while being able to intake, rate, and decision submissions from both Partner Portal and Pie’s API.

Aligning data models. Unifying previously unaligned data models was crucial for designing an interface that seamlessly handled both manually entered submissions via the Partner Portal and electronic submissions through the Quote API. Working with Product and Engineering, we conducted a thorough audit of the datasets, identifying key discrepancies and prioritizing the necessary changes. While many differences existed, only a handful required remediation, which was completed as pre-work to align the models and enable progress in designing a cohesive interface.

Creating differentiation. Underwriting leadership insisted on a single-page application for the underwriter rater, believing it would offer underwriters the fastest and most intuitive way to review, rate, and decision submissions. They also saw this design as a strategic move to create a competitive edge, ensuring Pie’s technology was positioned for efficiency and scalability—and, if needed, adaptable for a potential SaaS pivot. This ambitious vision demanded overcoming the challenges of simplifying complex, linear rating decisions. Through relentless iteration, detailed data mapping, and countless $6 lattes, I delivered a solution that met both technical demands and user needs, transforming the underwriting review process into a seamless experience.

Iterating based on feedback and data. Similar to the process taken with Partner Portal, user feedback was important to driving enhancements and planning future releases. Immediately the single-page interface was a success with underwriters reporting a 3-5x gain in efficiency, however, they noted the interface could be laid out in a more conducive manner, reducing the amount of vertical scrolling and fixing certain info and controls. V2 addressed these challenges while also providing additional use case coverage, pulling underwriters to the rater for more than 90% of submissions.

Scaling based on user case coverage. Iterating through feedback and data was pivotal in refining the underwriter rater. Following a user-first approach similar to the Partner Portal, we relied heavily on underwriter insights to guide enhancements and shape future releases. The single-page interface was an immediate success, delivering a 3-5x efficiency boost. However, users highlighted areas for improvement, such as reducing vertical scrolling and fixing key information and controls for easier access. Version 2 addressed these concerns and introduced broader use case coverage, further enhancing its functionality and usability.

The Underwriter Rater redefined efficiency, transforming complex tasks into quick, precise decisions. Its success demonstrated the viability of the single-page architecture, paving the way to scale this design philosophy into Pie’s Policy Management System—a solution that would drive the next stage of Pie’s growth.

Scaling with Purpose: Policy Management System

The cornerstone of Pie’s independence—and the key to ending a costly $1 million annual contract with Insurity—was a robust policy management system. This system needed to not only support the complexities of Pie’s current operations but also scale to meet their future ambitions: multi-state, multi-line, and multi-carrier business.

Data-Driven Design Hierarchy. Leveraging the success of the single-page architecture and shared data structures between rating and issued policies, I focused on reusing code and components to develop policy management functionality efficiently. With Pie’s long-term goal of writing multi-line policies across multiple carriers in mind, I conducted extensive data mapping exercises to ensure the architecture could support their vision. My efforts validated the scalability of the single-page design, positioning it to evolve alongside Pie’s growing ambitions.

Modular Design Architecture. Leveraging the single-page design approach from the Underwriter Rater, the system was scalable and adaptable to new territories and lines of business. UI elements such as drawers, toggles, and hover states minimized cognitive load while supporting advanced functionalities like task based user flows, third-party lookups, and SaleForce integrations.

Transforming Pie’s Technology with Pie OS

Over five years, I led the design of Pie OS, transforming Pie Insurance’s operations through proprietary systems. I designed the Partner Portal to streamline agent workflows, enabling instant quotes and policy binding. I designed the Underwriter Rater, introducing a single-page interface that tripled efficiency and redefined risk assessment. Finally, I scaled the design architecture into a robust Policy Management System, replacing a costly third-party system and empowering Pie to scale infinitely.

By leveraging user insights, modular architecture, and data-driven design, I created scalable solutions that modernized Pie’s technology, enabling unprecedented growth and positioning the company for future success.

Left-to-Right: Partner Portal, Underwriter Rater, Policy Management System UIs.
Modular Prototype