Building an AI Underwriter with MongoDB Atlas Vector Search and Voyage AI
Source code and setup: https://github.com/mongodb-industry-solutions/leafy-insurance-underwriter Check out the solutions library: https://www.mongodb.com/docs/atlas/architecture/current/solutions-library/insurance-underwriter/ Learn how you can improve the efficiency of insurance underwriting by using MongoDB and Voyage AI. Manual insurance underwriting can take up to 7 hours per submission and most carriers only quote half of them. By leveraging Voyage AI and MongoDB, it is possible to automate the tasks related to insurance underwriting in the same platform. See the agentic underwriter workflow in action—from fetching and reviewing auto and home insurance quotes to generating an explainable underwriting assessment. The walkthrough also explores how Voyage AI embeddings and MongoDB Atlas Vector Search support domain-specific retrieval within the MongoDB platform. Chapters 00:00 Human underwriting can take 3–7 hours 00:40 The manual underwriting bottleneck 01:40 Why AI agents need the right foundation 02:14 MongoDB + Voyage AI architecture 03:15 Live demo: Fetch and review quotes 04:29 Generate an underwriting assessment 05:16 Three lessons for AI underwriting 06:03 The evolving role of the underwriter 06:55 Source code and setup instructions Visit Mongodb.com → https://mdb.link/MongoDB Read the MongoDB Blog → https://mdb.link/Blog Read the Developer Blog → https://mdb.link/developerblog MongoDB for Developers YouTube Channel → https://www.youtube.com/@MongoDBDevelopers