Alex Panait’s first computer solved a personal problem: staying in touch with his father. Growing up in Romania, Panait spent long stretches apart from his dad, an oil-tanker captain whose messages arrived through telegrams collected at the post office. He remembers going home with his mother and deciphering text without spaces. A computer, acquired when he was in fifth grade, offered a better way to communicate and sparked an interest that took him into a computer science-focused high school, where he learned C and C++. Although he later pursued business management and finance, he kept building computers, running a custom-computer business during college. Today, as CEO of Lazarus AI, he is again assembling technology around practical needs, this time for organizations trying to make AI useful in consequential work.
Panait’s route to Lazarus ran through PwC, where he worked in financial services and eventually led digital innovation efforts. He understood the work involved in reviewing documents, assessing risk, and helping large organizations make decisions from incomplete information. After connecting with the Lazarus team through the insurance technology industry, he began testing an early version of its document-understanding technology. One discovery made the relationship more personal: the system understood Romanian. When he called the team to ask why they had not mentioned that capability, he recalls learning that they had not known either. The experience drew him closer to the company, which he joined to build its insurance practice before taking on broader strategy responsibilities and eventually becoming CEO.
That background informs how Panait describes Lazarus’s purpose. An insurer’s expertise, he argues, is underwriting risk; adopting AI should help it perform that work without requiring it to become a software company. Boston-based Lazarus develops systems for document intelligence, cybersecurity, and legacy technology modernization that run on-premises, in air-gapped environments, or inside a customer’s own private cloud—with no per-token charges and every output cited back to its exact source evidence. Its commercial approach largely runs through systems integrators, consultants, and service providers that bring the technology to their customers. Insurance is the exception where Lazarus also works directly with clients, drawing on the team’s industry experience to identify problems worth solving.
One example Panait shared involved a proposed insurance portfolio transaction containing approximately 3,000 claims. Buyers operating under tight deadlines typically examine a sample of the underlying files, he explained, leaving much of the portfolio unreviewed. According to Panait, Lazarus processed the portfolio in less than 24 hours and flagged a claim that appeared close to resolution but contained information suggesting a much larger potential obligation. Specialists investigated the finding and concluded that the exposure could have erased roughly 70% of the deal’s expected profits. The prospective buyer walked away and subsequently became a Lazarus customer, he said. For Panait, the significance was the ability to uncover risk that might otherwise have gone unseen and change a business decision.
Earning trust for that kind of work requires evidence. Panait says Lazarus builds explainability and governance into its systems from the beginning, agreeing with customers on the outcome they need before focusing on technical specifications. The same thinking underpins its approach to model independence: Panait views foundation models as the electricity and the surrounding harness, institutional memory, and verification layer as the appliance—allowing organizations to swap in better models without rebuilding their workflows, context, or controls. As foundation models become more widely available, he sees lasting value in the surrounding systems and expertise that make them dependable in actual operations.
Lazarus announced $43 million in funding in November 2026, with investors including QBE, State Farm Ventures, and Mucker Capital. Panait says its growth priorities include expanding cybersecurity capabilities, modernizing legacy systems, and advancing document and decision intelligence. Across those areas, his advice to businesses remains straightforward: begin with a specific problem and judge the technology by the result. “The boring part is where the value is,” he said. It is a practical philosophy for a company working in places where overlooked information, aging software, and difficult decisions carry real consequences.
