AI-native commercial insurance brokerage for E&S underwriting
Harper is a 2024-founded commercial E&S insurance brokerage built on an AI-native stack—Claude, ClickHouse, Python, TypeScript, plus real-time voice/video (Pipecat, LiveKit, WebRTC)—indicating deep automation across prospecting, quoting, binding, and post-sale service. The project list (outer-loop AI agents, ML models for underwriter load balancing, supervised fine-tuning pipelines) and pain-point focus (slow legacy distribution, human-led workflows) signal a company actively replacing manual insurance processes with agentic systems. Early-stage hiring velocity is decelerating, but engineering and sales headcount remain balanced for a pre-product-market-fit brokerage.
Notable leadership hires: Operations Lead, Content Lead, Chief of Staff
Harper operates as a commercial insurance brokerage for Errors & Omissions (E&S) coverage, a high-touch segment traditionally dominated by human-led underwriting and opaque distribution workflows. The company's differentiation is a proprietary tech stack designed to automate the full lifecycle: lead prospecting, quote generation, policy binding, and customer service. Infrastructure includes vector databases (ClickHouse) for training data, agentic development frameworks (Cursor, Claude via MCP), and real-time communication layers for underwriter and customer interaction. The org is small and lean—11–50 employees across engineering, sales, marketing, ops, and product—operating out of San Francisco.
Claude, ClickHouse, Python, TypeScript, Pipecat, LiveKit, WebRTC, Epic Systems (EHR integration likely for data), plus developer tools Cursor and Windsurf. Adopting MCP for tool integration.
Core projects include outer-loop AI agent development, ML models for underwriter load balancing, supervised fine-tuning pipelines, sales infrastructure, customer experience platform scaling, and full-funnel optimization for E&S distribution automation.
Other companies in the same industry, closest in size