General Magic Raises $7.2M for AI Insurance Agents

The seed round for General Magic was led by Radical Ventures, with participation from Andreessen Horowitz's (a16z) Speedrun program. Notable angel investors include Brendan O'Driscoll, VP of Product at Figma, and Larry James Erwin from OpenAI, signaling a deep interest from leaders at the intersection of product and foundational AI models. The company's total funding has now reached $8.4 million. General Magic's core product, "Cell," is an AI agent that integrates with existing insurance infrastructure like broker management systems and CRMs. Instead of a "rip and replace" approach, it acts as a reasoning layer on top of legacy systems, allowing insurers to become more AI-native without overhauling their core technology. This architecture is designed to unlock value from data trapped in older systems. The agentic AI is designed to handle asynchronous communication and complex, multi-step insurance workflows. This involves decomposing the claims process into a series of subtasks handled by specialized agents, a pattern known as a multi-agent system (MAS). Such systems are designed for autonomy and collaboration, allowing, for instance, an intake agent to pass structured data to a validation agent, which then routes it to a decision agent. Early deployments have demonstrated significant operational efficiencies, reducing the time-to-quote from 30 minutes to under three minutes. This is achieved by automating routine follow-ups and data collection via SMS. One of the world's largest general insurers reported a 30% reduction in inbound calls and saved over 250 hours a month using General Magic's agents. The founders, Anthony Azrak and Jai Mansukhani, are second-time founders with previous experience selling AI products into legacy industries. Their move into insurance was driven by firsthand frustration with the industry's inefficiencies. The company participated in the a16z Speedrun accelerator, which focuses on how "outsiders" can disrupt traditional industries. For backend and AI engineers, scalable AI agent architecture requires an API-first mindset with clear, role-based authentication and stable response formats. System design must account for high-volume, asynchronous communication, often using message queues like Kafka or RabbitMQ. Observability, including logs, metrics, and tracing, is critical for debugging and maintaining trust in the autonomous agents. The current venture landscape for insurtech shows a flight to quality, with investors concentrating capital into more mature startups with proven models. While overall deal volume has decreased, funding for B2B SaaS solutions in insurance remains strong. There's a clear trend toward funding companies that provide tangible efficiency gains for incumbents rather than those aiming to disrupt them entirely.