Serial entrepreneur Martin Varsavsky has emerged from stealth mode with his latest venture, Certuma, fueled by $10 million in seed funding led by 8VC. The company’s ambitious goal is to develop an Artificial Intelligence system that is safe and sophisticated enough to officially diagnose medical conditions and prescribe necessary treatments.

The Mission: Fixing the "I Am Not a Doctor" Problem

Varsavsky, known for launching over a dozen successful ventures, believes current AI tools fall short when patients seek actionable medical steps. He notes that while AI can offer accurate diagnoses, it currently must disclaim, “I am not a doctor,” preventing prescriptions or scheduling necessary tests.

“I want to fix the ‘I am not a doctor’ problem by building AI that is recognized by the FDA and recognized by the states,” Varsavsky stated. This innovation seeks to address the critical shortage of physicians, particularly in underserved rural areas, mirroring the essential role telemedicine played during the COVID-19 pandemic.

Targeting Initial Conditions and Regulatory Hurdles

Certuma has identified an initial target list of 25 common medical conditions, such as sore throats and urinary tract infections. These conditions were chosen because they carry low medical risk and adhere to well-defined management protocols, according to Dr. Armando Cuesta, the startup’s Chief Medical Officer.

The most significant obstacles involve ensuring patient safety against inaccurate advice and navigating the complex regulatory landscape. Varsavsky acknowledges that gaining regulatory approval is as crucial as perfecting the technology itself.

Funding, Strategy, and Global Operations

Austin, Texas-based Certuma launched with a $60 million valuation, securing its $10 million seed round from 8VC. The company is strategically pursuing development simultaneously in the United States and in Varsavsky’s native Argentina, where clinical trials are significantly less costly.

Sebastian Caliri, head of 8VC’s healthcare team, recognized Varsavsky’s experience in navigating regulatory conversations necessary for introducing such technology. Caliri emphasized that pioneering a clear path to FDA approval will be essential.

The Long Road to Approval

Achieving approval for an AI “doctor” capable of ordering tests or writing prescriptions will likely require a case-by-case approach. This might mean securing approval for treating urinary tract infections before moving to the next condition.

Further complicating matters is that physician licensing resides with individual states, potentially creating jurisdictional conflicts between state and federal oversight, especially concerning prescribing authority. Varsavsky anticipates this will be a “painful, long, difficult process,” but one that promises significant payoff.

Varsavsky’s Entrepreneurial History and Vision

Varsavsky has a history of success in technology and healthcare, including building the largest chain of fertility clinics in North America through Prelude, which merged with Inception Fertility in 2019.

His inspiration for Certuma struck when he experienced difficulty accessing care for a dermatological issue in Miami. He became focused on creating an always-available AI physician.

Dr. Robert Wachter of UCSF cautions that while AI is needed for the primary care crisis, the standards must remain high. He pointed out that even seemingly simple cases, like a sore throat, can mask serious conditions like throat cancer, requiring robust AI triage capabilities.

Building Trust and Safety into the AI

Varsavsky aims for an AI that adheres to the Hippocratic oath: “do no harm.” To mitigate risks like AI hallucinations, Certuma is implementing a dual safety layer. This includes a deterministic safeguard based on explicit clinical data rules to detect red flags, working alongside multiple Large Language Models (LLMs).

Initially, every AI-generated decision will be reviewed by a licensed physician. The long-term vision is for the AI to handle routine cases autonomously, freeing human doctors for complex issues. Varsavsky suggests AI doctors could be available for certain primary care tasks in as little as six months, noting that other pilots, like Doctronic in Utah, are already testing limited AI prescription renewals.