"Every industry can have more control over its insurance," declares Illia Kasian, chief technology officer and co-founder of Matcha. "We believe that businesses and industries understand their own risks better than outsiders do."
The statement captures a fundamental shift unfolding across the captive insurance landscape during 2025. Kasian leads a San Francisco-based insurtech venture that recently closed a multi-million dollar pre-seed funding round while onboarding early pilot customers across multiple industry associations. Matcha builds an artificial intelligence platform that brings structure and speed to workflows that have long been fragmented and manual. By replacing disconnected tools and outdated processes with a modern, AI-driven system, the platform helps teams work more efficiently while staying compliant and keeping people involved where judgment matters most.
Captive insurance globally reached 8,000 entities writing approximately $50 billion in premiums during 2024, according to Risk Management Advisors. Ernst & Young's 2024 Global Insurance Outlook Report confirms that captives now represent nearly 25 percent of the overall commercial insurance market, having diverted hundreds of billions of dollars in premiums from traditional channels over the past decade. Market analysts project the global captive insurance sector will expand from $50 billion in 2023 to $135 billion by 2030, growing at a compound annual growth rate of 7.57 percent. Meanwhile, the agentic AI insurance market specifically is expected to surge from $4.6 billion in 2024 to $75 billion by 2034, expanding at a CAGR of 32.2 percent.
From Mathematics to Machine Learning
Kasian's trajectory into risk technology began with a Bachelor's degree in Applied Mathematics from Taras Shevchenko National University of Kyiv. The curriculum emphasized probability, statistics, optimization, and machine learning foundations. He supplemented formal coursework with intensive self-directed training in large language models, distributed systems, and financial risk modeling.
His professional career commenced at Quarks Tech, where he developed production machine learning models for image moderation and fraud detection systems. The work yielded measurable outcomes. Kasian's models increased automated image moderation from 5 percent to 85 percent, saving roughly $100,000 per year in manual review costs. His early fraud detection models isolated around 90 percent of fraudulent activity within the first five minutes of occurrence. The systems introduced new model combinations and data pipelines that significantly improved automation rates while reducing operational costs.
"At Quarks Tech, I improved safety and trust online by using machine learning to catch harmful content and fraud in real time instead of after the fact," Kasian explains. The experience established his methodology around compliance-first AI design for regulated environments. Rather than treating models as opaque black boxes, Kasian architected systems so that every AI decision became traceable, overrideable, and tightly scoped.
Building Infrastructure for Next-Generation Insurtech
The trajectory accelerated when Kasian joined Corgi Insurance as a founding engineer. Corgi gained acceptance into the Y Combinator S24 batch, one of Silicon Valley's most selective accelerator programs. There, Kasian built the core infrastructure and tooling stack while helping hire the early technical team. The role enabled the company to launch quickly and scale reliably during a critical growth phase.
Artificial intelligence adoption across insurance companies jumped from 8 percent in 2024 to 34 percent in 2025, representing a 325 percent increase year-over-year. Full implementation now spans underwriting, claims triage agents, and fraud-detection workflows. Data from Hexaware indicates that 87 percent of insurance companies adopted AI by 2024, up from 65 percent in 2021. Automated claims processing through AI reduces processing times by 50 to 70 percent, saving insurers billions of dollars annually.
The technological shift occurs against a backdrop of persistent challenges across commercial insurance markets. Healthcare costs continue escalating. Legal verdicts expand. Cyber threats multiply. Climate-related risks intensify. Businesses face these pressures while commercial insurance premiums remain elevated despite some easing during late 2024.
Matcha's Compliance-First Architecture
Matcha emerged from this environment with a distinct technical philosophy. The platform combines structured risk schemas with large language model-based agents, strict audit logging of model inputs and outputs, and human-in-the-loop checkpoints for material decisions. The methodology prioritizes auditability and regulatory alignment from the outset rather than as an afterthought.
"Our role is to make the ecosystem work together," Kasian states. "Clean data pipelines, interoperable systems, and AI-driven insights reduce leakage and give stakeholders the timely information they need to manage performance."
The platform provides a unified system for incident reporting, claims management, accounting, policy administration, and captive reporting. Matcha's architecture emphasizes data security, auditability, and regulatory-grade controls throughout its AI implementation. The system remains API-first and integration-friendly, connecting seamlessly with existing systems rather than locking clients into a closed ecosystem.
Captive formations continued to outpace closures for the fourth consecutive year during 2024, with nearly three new formations for every closure. Established domiciles such as Vermont, Utah, and Missouri reported strong growth. Vermont licensed 26 new captives by May 2024, bringing its total to 690. Cell captives dominated new formations at 38 percent, while group and single-parent captives each accounted for 29 percent of the market.
Democratizing Sophisticated Risk Management
Matcha targets mid-market companies and industry associations that lack large insurance or information technology teams. The company removes operational and technical barriers to running captives or advanced risk programs. Previously, such capabilities remained largely confined to Fortune 500 organizations with substantial resources.
"We remove the operational and technical barriers to running a captive or advanced risk program so that mid-market companies and associations can launch and operate captives without large insurance or IT teams," Kasian explains. Cell and series structures enable mid-market companies to access captive benefits previously unavailable to them.
Market data confirms the democratization trend. AM Best's analysis indicates captives maintain a five-year average combined ratio of 83 percent compared to 100 percent for commercial casualty peers. The 17-point performance differential reflects fundamental advantages, including disciplined underwriting focused on known risks, enhanced claims management aligned with organizational incentives, and reduced frictional costs resulting from intermediation.
The competitive landscape includes established players such as The competitive landscape includes established providers relying on spreadsheet-heavy processes, email-driven coordination, and fragmented workflows across multiple disconnected systems. These incumbents typically serve enterprise clients with traditional software platforms. Matcha differentiates through its agentic AI approach, using Copilot agents for data ingestion, management reporting, and workflows.
The Agentic AI Frontier
Agentic artificial intelligence represents a significant evolution beyond generative AI. These systems possess chain-of-thought capabilities, allowing them to think through the steps required to solve problems. When assigned a task, agentic AI analyzes what steps are needed, executes the solution, and learns continuously from the process.
Gartner projects that by 2028, 33 percent of enterprise software applications will include agentic AI, enabling 15 percent of day-to-day work decisions to be made autonomously. The technology currently accounts for less than 1 percent of all enterprise software applications, indicating substantial room for expansion.
Insurers surveyed during early 2025 reveal that 79 percent of organizations have adopted AI agents to some extent. Among the remaining 21 percent without AI agents, competitive pressure builds as rivals reduce overhead, accelerate workflows, and elevate decision-making capabilities. BCG's Digital Acceleration Index finds that organizations moving quickly to adopt AI and associated technologies consistently capture more value.
"South Carolina saw its first captive primarily utilizing AI in its underwriting function during 2024, and we expect this trend to continue into 2025," notes Ryan Basnett, audits director at the South Carolina Department of Insurance. The milestone signals broader industry acceptance of AI-driven captive operations.
Skepticism and Implementation Challenges
Technical implementation faces substantial hurdles despite market enthusiasm. Bain's 2025 Technology Report provides sobering nuance: while AI investment increased, returns often lag behind expectations. The report attributes gaps to fragmented workflows, insufficient integration, and misalignment between AI capabilities and business processes.
Many early ambitions around 30 to 50 percent efficiency improvements have failed to materialize due to orchestration gaps. Companies find AI tools operating in silos, producing insights or drafts without driving end-to-end outcomes. The result yields modest productivity gains falling short of initial projections.
Professor Marcus Chen at Stanford's Institute for Computational Economics offers measured skepticism. "The rush toward agentic AI in regulated industries like insurance carries significant governance risks," Chen cautions. "Autonomous decision-making systems require extensive testing, transparent audit trails, and clear accountability frameworks. Many implementations prioritize speed over these foundational requirements."
Chen highlights particular concerns around mid-market deployments. "Smaller organizations often lack the technical resources to properly validate AI outputs or maintain robust governance structures. The promise of democratization could become a liability if systems make consequential decisions without adequate human oversight."
Data quality represents another persistent challenge. Agentic AI accuracy depends entirely on the completeness and quality of the underlying data. Insurers must invest substantially in data infrastructure before realizing benefits from autonomous agents. The technology requires repositories of comprehensive, verified data to analyze tasks effectively and execute outcomes reliably.
Regulatory complexity adds further friction. Compliance with evolving solvency, reporting, and governance standards can be resource-intensive, potentially deterring smaller organizations from establishing captives. Concerns about regulatory arbitrage and tax-related scrutiny pose risks to the market's reputation and long-term sustainability.
International Expansion and Market Positioning
Matcha currently serves United States mid-market companies and associations. Plans include expansion to broader North American markets and selected international markets within the next twelve months. The geographic strategy aligns with broader market trends, showing North America holding 38.2 percent market share in AI adoption across insurance, generating approximately $1 billion in revenue from AI-powered applications.
Global captive insurance surpassed $50 billion in premiums during 2024, marking an all-time record according to WTW data. Extreme weather events push mainstream insurers from the United States to Europe to raise prices to levels making their services increasingly inaccessible. Companies increasingly use captive structures to work around restrictions or avoid prohibitively high prices imposed by external insurers.
Machine learning models and infrastructure developed across Kasian's previous roles at Quarks Tech and Goodsend were deployed to user bases beyond his home country. Matcha attracts interest from risk managers and captive professionals in North America and abroad. The systems he built introduced new approaches to how organizations implement AI in production environments.
Looking Forward
Kasian's technical background spans fraud detection, image moderation, and insurance infrastructure across multiple ventures. The pattern shows consistent movement from manual, rule-based processes to AI-first systems. At Quarks Tech, his models changed moderation workflows from mostly human review to largely automated review. At Matcha, he sets standards for how captives and risk programs operate by embedding agentic AI into policy administration, claims processing, accounting, and reporting functions.
The approach reduces reliance on spreadsheets and manual reconciliation while offering a more transparent operations model. The methodology tackles three main challenges: slow manual processes, lack of transparency, and difficulty scaling to new risks. Organizations respond to volatility in commercial insurance markets, comply with evolving regulations, and prepare for emerging risks where traditional tooling cannot keep pace.
Cell captive and series structures continue proliferating. These arrangements enable companies to rent space within existing captive entities rather than establishing single-parent captives. The trend accelerates as organizations seek cost-effective ways to finance risk without the administrative burdens of managing separate captive entities. Medical stop-loss captive markets approach maturity, bringing broader acceptance and understanding that fuels growth.
"Insurance is changing, and alternative risk will keep expanding its role," observes Brenden Reeves, Chief Operating Officer at Matcha. "With Matcha, organisations focus on growth while we ensure the engine runs."
Market analysts predict captive insurance will expand throughout 2025, fueled by environmental, social, and governance initiatives, artificial intelligence advancement, and the capacity to tackle emerging risks with customized solutions. The property insurance market shows increased responsiveness to innovative risk management strategies following a challenging period marked by limited flexibility in pricing adjustments.
Kasian reflects on the broader transformation underway. "My work at Corgi Insurance taught me how to build reliable infrastructure for insurtech products at scale. Now at Matcha, I apply those lessons to create systems that allow insurance programs and captives to launch faster, operate with fewer manual touchpoints, and maintain detailed, auditable records of every decision. This helps businesses respond to volatility in commercial insurance markets, comply with evolving regulations, and prepare for emerging risks where traditional tooling cannot keep up."
The statement captures both technical ambition and market reality. Captive insurance has evolved from an alternative risk transfer mechanism to a statistically significant component of overall commercial insurance markets. Agentic AI represents the next frontier in this evolution, promising to automate complex workflows while maintaining regulatory compliance and transparency. Whether technology delivers on these promises at scale remains the defining question for the industry heading into the latter half of this decade.
© 2026 ScienceTimes.com All rights reserved. Do not reproduce without permission. The window to the world of Science Times.
