US juries awarded more than $14.5 billion in nuclear verdicts in 2023, a 15-year high, according to research firm Marathon Strategies. For brokers, carriers, and claims professionals, the cost of an incorrect coverage determination has rarely been higher.
Qumis, a Chicago-based coverage intelligence platform, launched attorney-certified AI agents Monday to address that risk. The platform analyzes entire insurance programs and returns citation-backed coverage conclusions across 16 lines of business.
The system deploys what Qumis calls a "bench" of specialist agents, from cyber and directors and officers liability to marine and workers' compensation. Each specialist analyzes its own line in parallel before the platform synthesizes findings into a single deliverable with citations linked to the underlying policy language.
When evidence is incomplete, the system surfaces uncertainty rather than forces a conclusion. Coverage errors that go undetected until a claim can carry E&O liability and reputational consequences for brokers and carriers.
The platform is built for brokers, carriers, MGAs, TPAs, claims professionals, and coverage counsel. Users upload a full insurance program, with policies, endorsements, and carrier forms processed in full. The system returns a finished work product rather than a summary.
Dan Schuleman, co-founder and CEO of Qumis, said the company built the platform to improve the quality of coverage decisions. "AI in insurance has largely been about helping people do the same work faster," Schuleman said. "We built Qumis to make them more intelligent, so coverage professionals can see the whole program more clearly and make better decisions."
The launch comes as AI adoption in insurance expands but coverage-specific analysis remains a gap. A Gallagher survey from earlier in 2026 found that 63% of insurance respondents had fully or partially operationalized AI, up from 45% in 2025. Only a small minority of carriers have scaled beyond pilots.
Beyond coverage analysis, the platform produces branded proposals, coverage checklists, Excel workbooks, and live interactive dashboards that can be published as hosted microsites. The company said this approach replaces static PDF proposals that brokers have relied on for decades.
Each session runs on what Qumis calls Private Compute, a dedicated virtual environment isolated from other customers' data. Co-founder and CTO Shiv Sinha said the architecture allows the platform to return the work product itself rather than a file summary.
Every specialist agent is built on a documented, line-specific coverage framework reviewed by Qumis coverage counsel and validated against representative scenarios before deployment. Qumis said five of the 15 largest US insurance brokers already use the platform, along with specialty carriers and law firms.
The certified agents are available to existing customers and through self-serve sign-up.
