Owkin

Owkin is an artificial intelligence company focused on applying advanced machine learning, biological reasoning, and multimodal patient data to biomedical research and drug development. Founded in 2016, the company was established to help researchers and healthcare organisations extract useful scientific insights from complex medical information while respecting the privacy and security requirements associated with patient data. Its work brings together artificial intelligence specialists, clinicians, computational biologists, data scientists, hospitals, academic institutions, and pharmaceutical companies. Owkin’s central ambition is to make biological research more efficient by developing AI systems that can analyse complex evidence, generate scientific hypotheses, and support decisions throughout the drug development process.

The company was co-founded by Thomas Clozel, a clinical oncologist, and Gilles Wainrib, a machine learning researcher specialising in biology. Their complementary backgrounds helped shape Owkin’s approach to biomedical AI. Rather than developing general-purpose AI tools, the company concentrates on biological and clinical problems that require specialised scientific knowledge. Its technology is designed to work with different types of biomedical information, including medical images, genomic data, molecular measurements, clinical records, scientific publications, and patient outcomes. Combining these information sources can help researchers investigate relationships that may be difficult to identify when each dataset is examined separately.

Owkin’s current central platform is K Pro, an agentic AI Scientist designed to support pharmaceutical research teams, biotechnology companies, academic researchers, and other life-science professionals. K Pro brings together biological language models, specialised AI capabilities, multimodal patient information, and scientific knowledge in a single environment. Users can ask questions in natural language and receive research-oriented responses without necessarily having to write code or operate several separate analytical applications. The platform is intended to help users move from an initial research question to evidence gathering, hypothesis development, biological analysis, and decisions about potential drug-development opportunities.

K Pro is designed to support work across different stages of pharmaceutical research and development. Researchers can use it to investigate biological targets, examine disease mechanisms, identify possible biomarkers, characterise patient populations, and assess the potential relevance of scientific findings to clinical outcomes. Pharmaceutical teams can also use the platform to examine drug-development portfolios, understand competitive research activity, and evaluate information relevant to clinical programmes. By bringing biological and clinical evidence into a shared environment, K Pro aims to help organisations make more informed decisions about which research questions and development opportunities deserve further investigation.

The platform uses agentic AI to coordinate specialised capabilities rather than relying exclusively on a single conversational model. Its system can select and combine different tools according to the research task. This approach is intended to make it easier to investigate questions that involve several stages of analysis, such as reviewing scientific publications, examining patient data, comparing biological findings, and preparing a structured report. The researcher remains responsible for evaluating the evidence and deciding how to use the results. K Pro is therefore positioned as a scientific research partner and decision-support system rather than a replacement for qualified researchers, clinicians, or pharmaceutical development teams.

K Pro provides three specialised working environments known as Analyze, Activate, and Amplify. These areas support different activities across the research and development process. Analyze focuses on investigating biological questions, generating hypotheses, examining disease mechanisms, and identifying potential therapeutic targets. Activate supports the use of research findings in practical development decisions, including biomarker evaluation and patient subgroup analysis. Amplify helps users extend their findings into wider strategic and development activities, such as portfolio assessment, competitive intelligence, and clinical trial evaluation. Together, these capabilities are intended to connect scientific investigation with decisions that affect drug development.

One of Owkin’s important strengths is its multimodal patient data network. Biomedical research often relies on information collected in different formats and from different institutions. A hospital may hold pathology images, another research centre may have genomic information, and a separate source may contain clinical outcomes or molecular data. These sources can be difficult to analyse together because they differ in structure, quality, availability, and governance requirements. Owkin works with hospitals, academic centres, clinicians, and other data partners to make relevant datasets more suitable for advanced AI analysis. Its network includes information such as histology, imaging, molecular measurements, genomic data, and clinical outcomes. This gives K Pro access to biological and clinical context that can support more detailed research.

Privacy-preserving data collaboration has been part of Owkin’s approach since its early development. The company has used federated learning and other privacy-conscious methods to support research involving sensitive patient information. Federated learning allows participating institutions to contribute to model development without necessarily transferring all their underlying patient data into one central repository. The exact technical and governance arrangements depend on the project, the data involved, and the participating organisations. This approach is important in biomedical research because patient information is subject to strict privacy, security, and regulatory requirements. Owkin’s work in this area is intended to make collaboration between healthcare institutions and research organisations more practical while respecting their responsibilities for sensitive information.

The company also develops specialised AI models and research capabilities for analysing pathology images. One example is Pathology Explorer, an AI agent designed to investigate histology data and identify different cell types and tissue structures in stained pathology images. It can support the detection, segmentation, and classification of cells, helping researchers examine the spatial organisation of tumour tissue and investigate the tumour microenvironment. These capabilities can be useful in cancer research, where the arrangement and characteristics of cells may provide information about disease behaviour and treatment response. Pathology Explorer is also designed to help researchers connect image-based findings with other biological information rather than treating pathology images as isolated visual material.

Owkin’s work in pathology and biological AI is supported by its wider research network. The company collaborates with academic hospitals, research centres, pharmaceutical organisations, and clinical experts to develop and evaluate its technology. These relationships are important because biomedical AI requires more than computational performance. A model may identify a statistical pattern, but researchers must still determine whether that pattern has biological meaning and whether it is relevant to actual patients. Owkin’s approach includes working with scientific and clinical partners to examine model findings, improve research methods, and investigate whether AI-generated hypotheses can be supported by additional evidence.

Scientific literature analysis is another important part of K Pro. Biomedical researchers need to review large amounts of published work to understand what is already known, identify unanswered questions, and determine how new findings relate to existing evidence. K Pro supports literature reviews, research summaries, and gap analysis. Its collaboration with Consensus strengthens its ability to retrieve and synthesise scientific literature. This can help researchers explore a topic, compare findings across publications, and identify areas where additional investigation may be valuable. The platform also supports scientific writing assistance and the preparation of research reports, although users should review generated material and verify scientific claims before using them in formal publications or decisions.

The platform also supports data exploration and visualisation. Researchers can investigate biomedical datasets, examine relationships between variables, and present findings in a more accessible format. K Pro can generate publication-oriented visuals and structured reports to help communicate research results. Its natural-language interface is intended to make some analytical capabilities accessible to both technical and non-technical users. This may reduce the need to move constantly between separate research tools, although advanced studies can still require specialist statistical knowledge, data preparation, and independent validation.

Owkin’s work extends beyond the K Pro platform into a wider ecosystem of biological AI research and pharmaceutical development. The company has developed biological foundation models and specialised AI agents intended to improve how computers interpret biological information. It is also working towards what it describes as Biological Artificial Superintelligence, an ambitious long-term research direction involving AI systems that can reason about biology and carry out increasingly complex scientific tasks. This is a development goal rather than a claim that the company has already achieved autonomous scientific intelligence across all areas of medicine. Its current products and research programmes represent steps towards that broader ambition.

The company has established partnerships with major pharmaceutical organisations to apply its technology to drug discovery and development. Its collaborations have included Sanofi, Bristol Myers Squibb, MSD, AstraZeneca, Boehringer Ingelheim, and Servier. These relationships involve different projects and commercial arrangements, including AI-supported research, data licensing, biomarker investigation, and the development of specialised agents. Owkin also works with hospitals and academic research institutions that contribute scientific expertise and patient-data resources. Such partnerships help the company connect AI development with practical research requirements in the pharmaceutical and healthcare sectors.

Owkin has also developed an ecosystem of technology partnerships to strengthen the capabilities of its AI Scientist. Its work with Anthropic has made Pathology Explorer accessible through Claude for Healthcare and Life Sciences using the Model Context Protocol. This provides a way for compatible Claude workflows to access specialised biological AI capabilities. The company has also announced a partnership with NVIDIA involving AI infrastructure and model-development technologies, including the Nemotron family of models and the NVIDIA NeMo framework. Its collaboration with Consensus strengthens scientific literature intelligence within K Pro. These partnerships demonstrate that Owkin’s platform is designed to work within a wider technology environment rather than depending entirely on a single proprietary model.

The company has a broader research and development ecosystem that includes its patient-data network, biological AI models, laboratory collaborations, and clinical validation activities. Its pipeline includes drug and diagnostic research programmes, although individual projects may be developed through separate organisations or partners. For example, Waiv is an Owkin spinout associated with AI-powered diagnostic testing. Its diagnostic products should be distinguished from K Pro, which is Owkin’s current central AI Scientist platform for biomedical research and biopharmaceutical decision-making. This distinction is useful for organisations evaluating the company because its research platform and its related diagnostic ecosystem do not necessarily have the same users, purchasing arrangements, or regulatory requirements.

K Pro is intended for a range of professional users. Pharmaceutical researchers can use it to investigate drug targets, biomarkers, and patient subgroups. Biotechnology companies can explore biological evidence and assess development opportunities. Academic researchers can use the free version to review scientific literature, test hypotheses, and explore biomedical datasets. Clinical research teams can investigate patient populations and examine information relevant to trial design and development. Executives and portfolio strategists can use the enterprise platform to support decisions involving research programmes, competitive intelligence, and development priorities.

Owkin’s global organisation supports its relationships with pharmaceutical companies, hospitals, research institutions, and technology partners across different markets. The company has offices and teams in locations including Paris, New York, Boston, London, Geneva, and Berlin. Its international network is relevant to its work because biomedical research frequently involves collaboration across countries and institutions. However, access to specific datasets, research capabilities, and enterprise services may depend on the user’s organisation, contractual arrangements, data permissions, and regional requirements.

Owkin is not simply a general-purpose chatbot or an ordinary medical information website. Its main focus is specialised AI for biological research and pharmaceutical development. K Pro combines scientific knowledge, multimodal patient data, and agentic workflows to help users investigate complex biomedical questions. The platform can support research and decision-making, but its outputs should not be treated as automatic proof of a scientific hypothesis or as a substitute for clinical judgement. Researchers and organisations still need to assess the quality of the underlying data, examine the evidence, and validate important findings through appropriate scientific methods.

Pricing

Owkin uses a freemium model for K Pro, with a free version for researchers and a separate enterprise offering for organisations that require advanced capabilities.

• K Pro Free – Free: Provides access to core research features, including literature and gap analysis, hypothesis testing, scientific writing assistance, and interactive data exploration and visualisation. The free version provides access to more than 17 biomedical datasets.
• K Pro Enterprise – Custom pricing: Designed for pharmaceutical companies, biotechnology organisations, and research teams requiring advanced agentic capabilities, additional data access, integrations, organisational deployment, and premium support.
• Enterprise onboarding: Owkin offers personalised demonstrations and Katalyst workshops to help organisations explore the platform and adapt it to their research requirements.

Owkin does not publish a universal price list for its enterprise platform. Organisations interested in advanced access need to contact the company for a demonstration and a quotation. Pricing may depend on the required capabilities, data access, integrations, number of users, and contractual scope.

Review

Owkin is a specialised AI company with a strong focus on biomedical research and pharmaceutical development. Its main strength is the way K Pro combines agentic AI, biological reasoning, scientific literature, and multimodal patient data in one research environment. Instead of functioning as a general-purpose chatbot, the platform is designed to help researchers investigate complex biological questions and connect scientific findings with practical drug-development decisions.

The availability of a free version is a significant advantage for researchers and academics who want to explore its core capabilities. Features such as literature analysis, hypothesis testing, scientific writing assistance, and biomedical data visualisation provide a useful starting point. The enterprise version offers a broader set of capabilities for organisations that need advanced data access, integrations, and support.

Owkin’s patient-data network and scientific partnerships also strengthen its position. Its work with hospitals, pharmaceutical companies, and research institutions provides access to specialised knowledge and data that can support the development and evaluation of biomedical AI. Its technology partnerships, including those involving Anthropic, Consensus, and NVIDIA, further demonstrate its effort to build a connected research ecosystem.

However, Owkin is not a simple tool for casual users. Its strongest capabilities are designed for scientific and pharmaceutical workflows, so users may need a background in biology, medicine, data science, or research to interpret some results effectively. The enterprise offering also has no publicly displayed standard price, which makes it difficult for organisations to estimate costs before contacting the company.

Another consideration is that AI-generated scientific insights require careful verification. Even when a platform provides statistical evidence and references, researchers must evaluate the methods, data quality, assumptions, and relevance of the findings. Access to patient information and advanced datasets may also depend on institutional permissions and contractual arrangements.

Owkin is best suited to biomedical researchers, pharmaceutical companies, biotechnology organisations, and academic institutions that need specialised AI support for scientific investigation and drug development. Its free K Pro version makes the platform more accessible, while its enterprise offering provides a route for organisations requiring deeper integration and advanced capabilities.