How the founder of Delegance vibe-coded an AI-native insurance brokerage on Weaviate

Summary

Delegance Brokerage is replacing the career insurance broker with agentic AI that reads, remembers, and acts on a client's entire policy history. Founder Alex Ledbetter built the platform on Weaviate in nine months and has grown it to $20M in premium under management in under a year, as a team of one.

The Challenge

Commercial insurance brokerages are often aging, paperwork-heavy businesses built on slow, manual workflows. Alex Ledbetter saw the gap while studying economics and English and interning at a global commercial insurance brokerage, where he underwrote political risk and credit. He set out to build Delegance, an agentic AI system that could act as a customer's broker of record by sourcing coverage, handling renewals, and answering client questions, while passing the cost savings on to customers by cutting broker commissions 60%.

The catch: Ledbetter is not an engineer, and he would be building it alone. After raising $200K from angel investors in February 2025 and obtaining 190 state insurance licenses in 30 days, he spent the next nine months standing up his first backend, database, and DevOps pipeline. He chose AWS and learned largely by feeding console screenshots to ChatGPT because the coding agents he used seemed to know it best.

His solution needed to classify, extract, and structure dozens of documents from an insurance binder in seconds; maintain a single, conversational memory of every client across an MCP server, web, mobile, Slack, voice, email, and text; run autonomous workflows for policy changes, quote generation, and portfolio reviews; and remain buildable and operable by a solo, non-technical founder working with AI coding agents.

Why Weaviate?

Developers Ledbetter trusted steered him toward a vector database for Delegance's agentic and RAG backend. He looked at AWS OpenSearch first but found it a nightmare to work with, and chose Weaviate for the developer experience it offered his AI coding agents.

• Self-serve signup: Ledbetter could stand up a cluster on a startup budget without navigating a sales process.
• Headless-first workflow: He could do everything through the API, MCP, and coding agents like Cursor, Claude Code, and Codex rather than digging around in consoles.
• Schema flexibility: Weaviate let him capture the same relational structure a knowledge graph would, on top of vectorized insurance binders.

What's Next?

Delegance's next phase is distribution: embedding its AI broker-of-record service inside other financial services businesses. Ledbetter is targeting banks with in-house brokerages, private-equity portfolio companies, and fintech platforms as partners, with a goal of surpassing $100M in annual premium and 10 live distribution partnerships within 18 months. Weaviate stays the memory layer underneath as that footprint grows.