Owkin K: the AI co-pilot set to transform biomedical research

In the fast-moving world of biomedical research, time is everything. The process of moving from an initial hypothesis to validated insights can take weeks or even months – a pace at odds with the urgency of drug discovery and clinical development. 

K Navigator, the first of Owkin’s new AI-powered co-pilots, promises to change that, helping scientists move from question to insight in minutes.

Developed by the Franco-American biotech company, Owkin – known for its pioneering use of AI to advance precision medicine – K Navigator builds on years of work at the intersection of data science, machine learning and oncology. Founded in 2016, Owkin partners with leading academic centres, hospitals and pharmaceutical companies to unlock the potential of multimodal patient data while keeping privacy and ethics front of mind.

At the heart of K Navigator’s development is Dr Eric Durand, Owkin’s Chief Research & Technology Officer, who leads the multidisciplinary team behind the tool. With a career spanning machine-learning research at UC Berkeley, biotech in Silicon Valley, and oncology data science at Novartis, Durand has spent years grappling with the challenges K Navigator is designed to solve.

Tackling fragmentation and slow cycles

According to Durand, two long-standing challenges have held back biomedical research: data fragmentation and the slow, iterative nature of biology.

“Data sits in silos across the biomedical community – clinical trial datasets here, research hospital biobanks there – and scientists often struggle to connect the dots,” he explains. “On top of that, biology is inherently iterative. It can take weeks to move from hypothesis generation to validation, especially for biologists who rely on data scientists to run analyses on their behalf. That slows everything down.”

K Navigator aims to change this by putting powerful tools directly into the hands of biologists. Acting as an AI co-scientist, it unifies data from millions of research articles, large databases, and rich multimodal patient datasets – everything from genomics and transcriptomics to medical images and clinical records. Crucially, researchers can interact with the system in plain English.

K Navigator, Owkin’s no-code AI copilot, lets biologists test and refine hypotheses in minutes. Watch the film to learn more.

The real breakthrough, Durand explains, is speed. “With K Navigator, a translational scientist in pharma could ask: ‘I have a drug targeting this gene – what is the best patient population to deploy it in?’ Instead of waiting weeks, they get a structured answer in minutes, grounded in literature, existing databases and patient data.”

This capability has the potential to accelerate not just academic research but also translational decision-making in drug development. By reasoning across different data modalities, the AI helps researchers rapidly refine hypotheses and identify the most promising clinical trial pathways.

With the explosion of large language models, such as ChatGPT, comparisons are inevitable. But Durand is quick to clarify the distinction.

“While ChatGPT is a generalist tool, K Navigator is purpose-built for biology and drug discovery,” he says. “Owkin has spent more than eight years developing proprietary AI models trained specifically on multimodal biomedical data. That means K Navigator doesn’t just read text, it integrates genomics, clinical data, and medical imaging, and reasons across them. Generalist models simply can’t do that.”

So, what does it take to build a game-changing AI tool like K Navigator? Behind the scenes, building a tool of this complexity has required more than just cutting-edge AI.

“The biggest challenge was assembling the right mix of expertise,” Durand notes. “You need world-class AI engineers, but also product managers who understand researchers’ needs, and biologists and medical doctors – the end users themselves – integrated into the development process.”

The second critical ingredient is data. Owkin’s extensive partnerships with university hospitals and academic centres across Europe, the UK and the US provide access to rich patient datasets – the essential fuel for training and validating the AI.

From Navigator to Pro: the next step

Currently in beta, K Navigator launched on 6 May at AWS Reinvent in New York and is being rolled out to selected users. But for Owkin, it is only the beginning.

K Navigator serves as the free entry point to K Pro, the enterprise-grade version designed for pharma and biotech. Whereas Navigator connects to a subset of data and tools, K Pro will open access to much deeper datasets and advanced AI capabilities. This includes the ability to discover prognostic biomarkers, automatically identify patient subgroups for clinical trials, and securely connect a company’s proprietary data to Owkin’s AI.

Pilot programmes for K Pro are being run throughout the autumn, with a beta launch scheduled for October.

Looking further ahead, Durand sees K Pro evolving from a co-pilot to a more autonomous research partner. “In five years, we believe K Pro will not just make human researchers more productive – it will begin generating novel hypotheses itself, developing a deeper, autonomous understanding of biology,” he says.

This aligns with broader advances in AI. While frontier models are demonstrating remarkable reasoning in maths and coding, Owkin’s focus is on pushing that frontier into biology. “We’re building AI that can reason through complex biological data and questions. That’s where the real innovation is needed, and where we’re determined to lead.”

For now, K Navigator offers the research community a glimpse of what’s to come. By collapsing the time it takes to move from question to insight, it could fundamentally reshape the pace of discovery.

Durand sums it up simply: “The main value of K is to supercharge human researchers. It’s not replacing them – it’s making them vastly more productive.”

Head to Owkin.com to learn more.