Insurance is complicated. Customers have questions before they quote, need guidance after, and expect clarity when they file a claim. But the work of answering those questions, collecting documents, and following up still runs on calls, emails, and portals stitched together by manual effort. For brokers and carriers, this coordination overhead is one of the most operationally expensive and taxing parts of the business.
General Magic is building AI agents to solve this problem.
Today, the company announced a US$7.2M seed funding round led by Radical Ventures, with significant participation from a16z Speedrun and new investment from Figma VP of Product Brendan O’Driscoll and Larry James Erwin from OpenAI. The company has raised $8.4M to date, backed by Radical Ventures, a16z Speedrun, and Comma Capital, along with operators who have built foundational AI and product platforms, including Aidan Gomez, CEO of Cohere, as well as the executive team at Braze, including Kevin Wang, Chief Product Officer, and Spencer Burke, SVP of Growth.
General Magic builds AI agents that take over the work insurance teams spend the most time on: answering routine questions, collecting documents, and following up with customers when clarity matters. These agents work across the full insurance lifecycle, covering pre-quote eligibility, post quote engagement, and claims coordination. They do all of this while connecting directly to broker management systems, quoting platforms, and CRMs.
Early deployments show what’s possible. Working with one of the world’s largest general insurers, General Magic has reduced time-to-quote from roughly 30 minutes down to under 3 minutes via its SMS-based agent.
The company’s agentic offerings are centered around a product called Cell, a proactive AI agent that connects directly to the systems insurance teams already use. Cell integrates with broker management systems, quoting and rating platforms and CRMs to support teams. It can be deployed across SMS, iMessage and RCS, and can extend into policy, billing and claims workflows as needed.
When a customer has a question, they can text Cell over SMS, or the insurance team can proactively deploy it to the customer. The agent responds using real system data, asks for missing information, follows up automatically, and updates records as workflows progress. Conversations stay in one thread, context is preserved, and customers move forward at their own pace without being chased or dropped.
Early deployments point to the scale of the opportunity. In early rollouts with large personal lines insurers, General Magic is reducing the time required to generate and finalize quotes from roughly 30 minutes to about 3 minutes by automating routine clarification and follow ups over SMS across auto and life insurance workflows. This increase in speed expands effective quoting capacity while keeping customers engaged through the most failure prone part of the journey after a quote is issued. By handling frequent questions and coordination over text, the agent reduces delays and prevents conversations from stalling. General Magic is currently supporting deployments with carriers across auto and life insurance, where post quote follow through and customer coordination are most critical.
In parallel, the team is focused on building agents that understand the realities of insurance distribution, including licensing and regulatory frameworks such as RIBO, OTL, and other broker and advisor exams. By specializing agents around how licensed professionals are trained to communicate, General Magic aims to ensure conversations feel accurate, compliant, and aligned with how insurance teams actually explain coverage to customers.
General Magic was founded by Anthony Azrak and Jai Mansukhani, second-time founders who previously sold AI products into legacy industries. The company’s move into insurance came from firsthand frustration. After a water leak spiraled into weeks of calls, delays, and higher premiums, the founders began exploring how common this experience really was. What they found was an industry that technically works, but often fails customers and intermediaries in the moments that matter most. That insight shaped General Magic’s decision to go deep into insurance rather than remain a horizontal AI platform.
The broader industry context underscores the urgency. Retention rates in insurance lag behind other sectors, and acquiring new customers costs significantly more than keeping existing ones. As digital distribution accelerates and customers shop more aggressively at renewal, both carriers and brokers that fail to improve post-quote engagement risk losing revenue they already worked to win.
Looking ahead, General Magic plans to expand across insurance lines and workflows, staying focused on moments where customer intent is high and coordination most often breaks down. The platform is being built to support high impact workflows across the insurance stack, prioritizing areas where follow through fails today and where fixing it creates meaningful value for customers, brokers, and carriers.
The long term vision is simple but ambitious: make follow through automatic, reliable, and invisible. By removing the need for manual chasing and fragmented handoffs, teams can spend less time managing processes and more time serving end customers. The team is motivated by solving complex, real world problems that sit at the center of insurance operations, with the goal of delivering tangible improvements to how people experience insurance when it matters most.
Deepgram and IBM Introduce Advanced Voice Capabilities for Enterprise AI
Posted in Commentary with tags Deepgram on February 24, 2026 by itnerdIBM and Deepgram today announced a collaboration to integrate Deepgram’s industry-leading speech-to-text and text-to-speech capabilities into IBM’s watsonx Orchestrate generative AI solution.
To address client needs for highly performant, enterprise-grade transcription and real-time captioning, IBM will embed Deepgram’s capabilities into watsonx Orchestrate. This collaboration makes Deepgram IBM’s first voice partner, bringing voice AI technology that helps enterprises automate their operations and meet the growing demand for conversational AI technology, including advanced speech-to-text voice recognition so users can interact with digital agents using natural speech.
Many organizations are adopting AI-powered speech-to-text systems to automate transcription while handling real-world audio conditions, including background noise, diverse accents, and real-life dialog. This integration addresses these challenges by offering a wider range of languages and dialects, including dozens of Arabic and Indian variants, along with voices that reflect regional accents. It also adds options for custom tuning, real-time captioning and natural-sounding speech.
These technologies open new possibilities for enhanced automated customer care and support, call analysis, and voice-driven data entry in fields like healthcare and finance.
Voice interfaces are quickly becoming essential for enterprise AI, and this collaboration strengthens IBM’s role in delivering modern, flexible solutions to its clients. For Deepgram, it expands access to new customers through a trusted enterprise partner and reinforces its position as a reliable, real-time voice platform built for large-scale use.
Leave a comment »