Biopharmaceutical artificial intelligence startup Owkin Inc. has announced the launch of a suite of autonomous agents aimed at drug discovery and research, following partnerships with industry leaders Nvidia Corp. and Anthropic PBC. This initiative marks a significant step for the decade-old AI company as it seeks to provide tools that leverage real-world patient data for the healthcare sector.
Owkin is releasing its comprehensive agentic infrastructure platform to clients, which includes a meticulously curated multimodal patient dataset collected from over 800 hospitals worldwide. Unlike many AI firms constrained to lab-generated data, Owkin claims its agents are trained on clinically relevant insights derived from a diverse cohort of real patients, allowing for a more accurate reflection of actual diseases.
The company envisions that its AI agents will assist drug companies and healthcare organizations in identifying biomarkers, interpreting complex medical datasets, and making informed decisions during clinical trials. By utilizing real patient data, Owkin believes it can enhance clinical success rates and expedite drug discovery, ensuring that new medications reach those in need more swiftly, all while preserving patient privacy.
Co-founder and Chief Executive Thomas Clozel described the launch of the agentic infrastructure as a pivotal milestone for the healthcare industry. He emphasized that it allows researchers unprecedented access to advanced AI tools for drug development, stating, “For the first time, healthcare companies can use specialized agents to quickly and efficiently assist in research, analysis and clinical trial decision making, from their platform of choice.”
Owkin positions itself as a key connector that integrates real-world patient data from hospitals, laboratories, and academic institutions through its AI agents into a pharmaceutical industry that is increasingly reliant on such data. Clozel envisions the agentic infrastructure as the groundwork for what he refers to as “biological artificial super intelligence,” an AI that can reason through the intricate complexities of biological systems. He hopes that, over time, Owkin can fully automate scientific research and drug discovery, paving the way for innovative treatments and diagnostics.
The first AI agent released by Owkin is named Pathology Explorer, which will be available through Anthropic’s Claude for Healthcare Life Sciences offering, showcasing the interoperability of the agentic infrastructure. Pathology Explorer is based on Owkin’s Histo Interpretability Prediction Engine model and is designed to identify various cell types, their locations, and biomarkers using only digital pathology images.
Clozel noted that this capability has the potential to significantly accelerate drug discovery and clinical trial design. The agent has been trained on Owkin’s extensive patient dataset, achieving a remarkable 23.7% improvement in classification accuracy on key benchmarks, while utilizing five times fewer parameters than existing models for similar research. This efficiency allows Pathology Explorer to reduce computational times from weeks to mere hours.
Owkin’s AI agents will be accessible via an application programming interface (API) and are compatible with the Model Context Protocol, enabling seamless integration into existing healthcare workflows with minimal disruption. The agents are designed to evolve continually, improving over time as they are refined using fresh patient data and lab-in-the-loop testing.
The partnership with Nvidia targets enhancements to Owkin’s foundational model, OwkinZero. The company intends to refine its biological large reasoning model through Nvidia’s tools and services, including the Nemo framework and cuCIM, an open-source library for accelerated multidimensional image processing. Clozel stated that this collaboration aims to bolster the performance and scalability of OwkinZero’s biological reasoning capabilities, enhancing its robustness.
“By combining Nvidia’s world-class AI ecosystem with our domain-specific biological reasoning capabilities and multimodal patient data, we’ll be able to create AI agents that can truly understand and reason about the complexity of biology at a scale that was previously impossible,” Clozel said.
As Owkin deploys its agentic infrastructure, the healthcare sector stands at the cusp of a transformation that could redefine drug discovery and clinical trials, potentially leading to faster and more effective treatments for patients.
See also
NVIDIA Expands BioNeMo Platform with $1B AI Lab Partnership for Drug Discovery Advances
Alibaba’s AI Division Surges as Downloads Exceed 700M Amid Favorable Regulatory Shift
Chalmers University Researchers Accelerate High Entropy Oxide Discovery with Active Learning
Microsoft Acquires Bonsai to Enhance Azure’s AI and Reinforcement Learning Capabilities
