RiskPair

Preparing a commercial insurance submission starts with research: an unfamiliar industry, the right questions, the property, and the markets likely to want the risk.

Brokers do that research risk by risk, and the quality of the submission depends on what they already know.

A complete submission, prepared with the right information up front, saves time for the broker and for the underwriter. It also helps a broker handle a trade they do not usually see.

RiskPair puts the research and the questions in front of the broker before the submission goes out.

The brokerage operating system I built for an employer standardized the operational workflow, including a template-based path for commercial submissions. RiskPair went deeper into the risk itself. It is a separate application, not part of that system, and a broker could use it when a risk needed more than a standard template. See the brokerage case study.

RiskPair is my own project. I defined the broker workflow, designed the product, and built it, drawing on what I know about commercial insurance and underwriting submissions.

It is the project where I combined domain knowledge, AI, research, workflow design, and product development most directly.

• AI-assisted research on unfamiliar industries, including market hardness, using the OpenAI API
• Market guidance: likely carrier appetite, carrier matching, and target-premium guidance
• Dynamic question generation, with follow-up questions
• A step-based broker workflow
• Property and risk enrichment with Google Maps, Places, Street View, and geocoding
• Polished underwriting PDFs: cover page, broker notes, risk summary, photos, property information, questionnaire responses, and supporting risk information
• A dedicated commercial-submission pipeline: tracking submissions, carrier responses, and follow-up reminders
• An internal training system
• Gamification: levels, streaks, and stats

Technical path: first built in PHP, HTML, CSS, and JavaScript, then rebuilt in React.

A broker starts with the risk. RiskPair researches the industry, then asks questions generated for that risk, with follow-ups where an answer needs more detail. It enriches the property with map, place, and street-level data, suggests carriers with likely appetite, and assembles everything into a structured underwriting PDF. The broker then tracks the submission through a commercial-submission pipeline, with carrier follow-up reminders.

A working product, live at riskpair.com and tested with selected commercial brokers.

It has not had a full-market rollout. Carrier matching suggests likely appetite; partnerships with carriers were a future concept, not something that was in place.