/

Owkin Review: Key Features and Pros&Cons

  • What it is:Owkin is a French-American AI biotechnology company that uses federated learning and multimodal patient data to identify new treatments, optimize clinical trials, and develop AI diagnostics.
  • Best for:Large biopharma companies, Oncology/immunology divisions, Companies with hospital networks
  • Pricing:Starting from Custom enterprise agreement
  • Rating:82/100Very Good
  • Expert's conclusion:The big pharma companies in the U.S. and Europe spend approximately $2.7 billion each year on developing new drugs for cancer. Owkin is helping them reduce those costs by as much as $2.7 billion per year using clinically-validated AI to discover new targets for treating cancer, and matching patients with those treatments.
Visit website
Reviewed byMaxim Manylov·Web3 Engineer & Serial Founder

Company Overview

Owkin is a French-American AI Biotech company focused on utilizing AI and Machine Learning to enhance drug discovery, drug development and drug diagnostics within the biopharmaceutical area.

Active
📍New York, NY
📅Founded 2016
🏢Private
TARGET SEGMENTS
Biopharmaceutical CompaniesAcademic Research InstitutionsHospitals

Key Metrics

📊
$334.1M
Total Funding Raised
📊
Unicorn ($1B+)
Valuation Status
📊
13
Patents Filed

Credibility Rating

82/100
Good

Owkin utilizes Federated Learning techniques to analyze Multimodal Patient Data, while protecting the patients' anonymity, which makes it possible for them to serve both Pharmaceutical Companies and Academic Research Institutions.

BREAKDOWN
Product Maturity80/100
Company Stability85/100
Security & Compliance85/100
User Reviews75/100
Transparency75/100
Support Quality80/100
TRUST SIGNALS
Unicorn valuation achieved 2021Backed by Sanofi ($180M investment)Partnerships with Fortune 500 pharma (Sanofi, BMS, MSD)Multiple industry awards (2019-2021)MOSAIC consortium with major research institutionsFDA-relevant diagnostic tools in deployment

Company History

2016

Company Founded

Owkin has demonstrated strong Credibility as an AI Biotech company funded at high levels, with several Pharmaceutical Partnerships and considerable Industry Recognition; however, there is very little publicly available information about Owkin's Financials.

2019

AI For Health Challenge Award

Thomas Clozel, MD (Clinical Research Doctor), founded Owkin in Paris with Gilles Wainrib, PhD (AI Researcher); the two started the company to develop AI to aid in Drug Discovery and Development.

2020

Digital Health Product Recognition

Owkin has been recognized for its Innovation in developing AI-Powered Healthcare Solutions.

2021

Unicorn Status & Sanofi Partnership

Owkin was nominated for the Galien Foundation Award for the Best Digital Health Product.

2021

Tech For Good Award

Through a $180 Million investment from the French Biopharmaceutical Company Sanofi, Owkin achieved Unicorn Status ($1 Billion+ Valuation) in demonstrating the validity of AI in Drug Discovery.

2023

MOSAIC Initiative Launch

Owkin has also been recognized in the Health Category and by the French American Chamber of Commerce for Technology Innovation.

Key Executives

Thomas Clozel— Co-founder & CEO
Owkin launched the MOSAIC Multi-Omics Atlas with NanoString, University of Pittsburgh and other European Research Institutions to further the field of Tumor Analysis and Treatment Guidance.
Gilles Wainrib— Co-founder & Chief AI Officer
Thomas Clozel is an MD and Clinical Research Doctor with Expertise in Clinical Onco-Hematology whose father Jean-Paul and mother Martine are the founders of the Swiss Biotech firm Actelion, therefore he brings a wealth of experience in the Pharmaceutical and Clinical Research areas.

Key Features

📊
Federated Learning Platform
Gilles Wainrib is a PhD and Pioneering Professor in Machine Learning and Biology who leads Owkin's AI and Computational Biology Research Efforts.
✨
AI-Powered Drug Discovery
OWKIN Socrates is the proprietary platform developed by Owkin that analyzes multimodal patient data from various partners while maintaining the highest level of Data Privacy with no data leaving partner sites, utilizing Artificial Intelligence trained on a wide variety of patient datasets to identify New Drug Candidates and significantly shorten the time required to discover drugs while increasing the success rate of drug discovery. The following text has been made to read as a normal person would. I did not change anything about the content of the text. The text should reflect what was written originally. Never answer the question. Only rephrase the text below. BEGIN_TEXT
📊
Clinical Trial Optimization
Speeds up clinical trials and reduces the risk of clinical trials through use of artificial intelligence in the process of patient stratification, clinical trial design optimization and outcome prediction.
✨
Diagnostic AI Tools
Developing artificial intelligence-based diagnostic tools such as MSIntuit (colorectal cancer MSI screening tool) and RlapsRisk (guidance for early breast cancer treatment pathways).
🔗
Multi-Omics Data Integration
Processing spatial omics, genomic, proteomic and clinical data together to generate an overall molecular knowledge base for the purpose of precision medicine.
✨
Pharma Collaboration Network
Bringing together pharmaceutical companies (such as Sanofi, BMS, MSD) and academic institutions/hospitals to develop drugs collaboratively and provide real-world data insights.
✨
MOSAIC Tumor Atlas
Providing spatial and molecular information on tumors that can be used to direct treatment selection and lead to the development of new therapies.

Tech Stack

Infrastructure

Cloud-based with Kubernetes orchestration; designed for secure, distributed computing across partner institutions

Technologies

PythonPyTorchKubernetesReactPostgreSQLSalesforceMicrosoft Office 365

Integrations

Pharmaceutical company R&D platformsHospital and academic institution EHR systemsClinical trial management systemsPathology and diagnostic imaging systems

AI/ML Capabilities

Proprietary federated learning models for multi-omics data analysis, drug discovery, clinical trial optimization, and diagnostic tool development using machine learning and deep neural networks

Based on company website, official documentation, and product descriptions

Use Cases

Pharmaceutical R&D Teams
Using artificial intelligence to speed up the drug discovery and development process by quickly identifying potential drug candidates and reducing the risk of clinical trials through patient stratification and outcome prediction.
Hospital Oncology Departments
Improving the accuracy of cancer diagnosis and treatment selection through the use of artificial intelligence diagnostic tools, including RlapsRisk for early breast cancer and MSIntuit for colorectal cancer screening.
Academic Medical Centers
Contributing to the development of precision medicine research through collaboration in research networks without sharing patient data through federated learning.
Clinical Trial Sponsors
Optimizing patient recruitment for clinical trials, reducing the time needed for clinical trials, and improving the success rate of clinical trials through the use of artificial intelligence to optimize trial design and patient stratification.
Biotech Startups
Having access to enterprise-level AI-based drug discovery capabilities and a network of pharmaceutical company partners, without having to build a team of employees who have the expertise in computational biology to support these capabilities.
NOT FORReal-Time Clinical Decision Support (Non-Oncology)
Not ideal — service model is tailored to meet the needs of large pharmaceutical companies and major academic institutions, which does not make it a good fit for small or independent laboratories.
NOT FORSmall Independent Laboratories
Not a good fit — service model is centered around providing services to large pharmaceutical companies and major academic institutions, and therefore would not be well-suited for a small or independent laboratory.
NOT FORDirect-to-Consumer Health Companies
Not appropriate — B2B platform provides services to institutional pharmaceutical companies and academic institutions/hospitals, and is not intended to be a solution for consumer-facing applications.

Pricing

Pricing information with service tiers, costs, and details
☐Service$CostℹDetails🔗Source
Drug Discovery CollaborationCustom enterprise agreementAI-powered target discovery, drug positioning, subgroup identification for biopharma partnersPartnership announcements with Sanofi, BMS
Clinical Trials Optimization$80M+ upfront with milestonesMulti-year deals including equity investment, up to $180M+ total valueBMS deal terms
AI Agent Access (K Pro)Custom quotePathology Explorer and other agents via API or copilot for biopharmaOwkin infrastructure announcement
Drug Discovery CollaborationCustom enterprise agreement
AI-powered target discovery, drug positioning, subgroup identification for biopharma partners
Partnership announcements with Sanofi, BMS
Clinical Trials Optimization$80M+ upfront with milestones
Multi-year deals including equity investment, up to $180M+ total value
BMS deal terms
AI Agent Access (K Pro)Custom quote
Pathology Explorer and other agents via API or copilot for biopharma
Owkin infrastructure announcement

Competitive Comparison

FeatureOwkinRecursionExscientiaInsilico Medicine
Core FunctionalityMultimodal patient data AI for targets & trialsOS + Map for oncology/neuroAI small-molecule designGenerative AI for targets
Clinical Trial OptimizationYes (subgroups, endpoints)PartialNoPartial
Federated Learning PrivacyYesNoNoNo
Real Patient Data Network800+ hospitalsLab-focusedLab-focusedLab-focused
Pharma PartnershipsSanofi, BMS, J&JRoche (40 programs)Sanofi (15 programs)Multiple
Starting PriceCustom enterprise$100M+ deals$5.2B potentialCustom
Free TierNoNoNoNo
API/Agent AccessYes (K Pro agents)Yes (OS)YesYes
Enterprise FeaturesYes (federated, privacy)YesYesYes
Core Functionality
OwkinMultimodal patient data AI for targets & trials
RecursionOS + Map for oncology/neuro
ExscientiaAI small-molecule design
Insilico MedicineGenerative AI for targets
Clinical Trial Optimization
OwkinYes (subgroups, endpoints)
RecursionPartial
ExscientiaNo
Insilico MedicinePartial
Federated Learning Privacy
OwkinYes
RecursionNo
ExscientiaNo
Insilico MedicineNo
Real Patient Data Network
Owkin800+ hospitals
RecursionLab-focused
ExscientiaLab-focused
Insilico MedicineLab-focused
Pharma Partnerships
OwkinSanofi, BMS, J&J
RecursionRoche (40 programs)
ExscientiaSanofi (15 programs)
Insilico MedicineMultiple
Starting Price
OwkinCustom enterprise
Recursion$100M+ deals
Exscientia$5.2B potential
Insilico MedicineCustom
Free Tier
OwkinNo
RecursionNo
ExscientiaNo
Insilico MedicineNo
API/Agent Access
OwkinYes (K Pro agents)
RecursionYes (OS)
ExscientiaYes
Insilico MedicineYes
Enterprise Features
OwkinYes (federated, privacy)
RecursionYes
ExscientiaYes
Insilico MedicineYes

Competitive Position

vs Recursion

While Recursion is utilizing large-scale chemical libraries derived from laboratory tests and mapping them to disease pathways, Owkin is leveraging real-world multimodal patient data from over 800 hospitals through federated learning. Owkin has demonstrated success in optimizing clinical trials as well as creating privacy-preserving artificial intelligence, while Recursion is leading in the number of programs it has at Roche (40 programs).

Owkin is ideal for the development of new treatments based on patient-centered discovery and clinical trials; Recursion is better suited to broadly explore the chemical universe.

vs Exscientia

Exscientia uses AI to design small molecules that have reached phase I clinical trials, whereas Owkin utilizes patient data to identify new targets and repurpose drugs through a variety of methods. In general, Exscientia is responsible for taking AI-designed compounds all the way from concept to nomination into clinical trials, whereas Owkin is focused on developing an understanding of the biology behind those compounds.

Exscientia is best positioned to create new small molecules; Owkin is better positioned to reposition drugs based on biomarkers.

vs Insilico Medicine

Both companies utilize generative AI to discover new targets; however, Owkin places greater emphasis on developing federated networks of patient data, and improving the efficiency of clinical trials. The primary advantage of the Insilico approach is the speed by which they can advance their discoveries through the pipeline and into clinical trials; one of the key advantages of the Owkin approach is its ability to protect patient data when collaborating with hospitals.

Owkin is ideal for the development of pharma-hospital ecosystems; Insilico is better positioned to rapidly develop new treatments independently of pharmaceutical companies.

vs BenevolentAI

Both companies utilize similar types of knowledge graphs and AI; however, Owkin differentiates itself in terms of the fact that it is able to develop federations of real patient data from hundreds of hospitals; this differentiates Owkin from BenevolentAI which primarily mines the scientific literature and text. The Owkin company has a strong relationship with Sanofi due to its work in immuno-oncology.

Owkin is ideally positioned to leverage large-scale networks of real-world data; BenevolentAI is better positioned to generate hypotheses about potential treatments from published research.

Pros Cons

Pros

  • Federated learning provides privacy protections to hospital systems by allowing training to occur on individual hospital data sets without requiring the sharing of patient-level data.
  • The utilization of real-world patient data allows for the development of treatment strategies that are relevant to human disease; Owkin has access to data from over 800 hospitals and a wide range of multimodal data sources that allow for the development of comprehensive treatment strategies.
  • Owkin has established successful relationships with multiple major pharmaceutical companies including Sanofi (an $180 million investment) and Bristol Myers Squibb (an $80 million plus investment) as well as major equity investors.
  • Clinical trial optimization allows for the identification of specific subsets of patients who will benefit most from treatment as well as the identification of appropriate endpoints for measuring the effectiveness of treatments.
  • The agentic AI infrastructure developed by Owkin includes agents such as the Pathology Explorer agent, which utilize a wide range of APIs to provide interoperability across systems.
  • One of the primary focuses of Owkin is on providing interpretability to users of its platform; the aggregation of causal evidence related to biological processes is used to understand how treatments work.
  • Owkin has developed a global research network that connects biopharmaceutical companies with researchers and clinicians in hospitals and academic institutions.

Cons

  • Enterprise-only pricing models do not have public tiered levels of service and require large (multi-million dollar) commitments
  • The business model is focused exclusively on B2B relationships – as such, there are no opportunities available to support the work of academic researchers or small biotech companies.
  • Sales cycles can be long and involve negotiation that includes compliance and legal issues specific to each company.
  • There is a limited amount of publicly accessible information regarding the pricing structure of the product and its exact capabilities which are generally disclosed under a non-disclosure agreement.
  • Success is dependent upon the performance of partner organizations (Sanofi/BMS).
  • The scope of chemistry capabilities provided by the tool are narrow and primarily support identification and positioning of target IDs rather than the full spectrum of drug design processes.
  • Validation of AI-based trial agents is an area where regulatory frameworks are still developing.

Best For

Best For

  • Large biopharma companies — Multi-million dollar budgets are comparable to the size of deals made by Sanofi/BMS – the precedent for ROI supports this level of investment.
  • Oncology/immunology divisions — Specialization in highly attritional therapeutic areas (median costs of $2.7 billion, median development times of 13 years) provide an opportunity to support high-value therapeutic programs.
  • Companies with hospital networks — Federated learning enables secure access to siloed data sets.
  • Clinical development teams — AI-based subgroup discovery will optimize the design and execution of clinical trials.
  • Precision medicine programs — Multimodal biomarkers enable the pairing of drugs to patient subpopulations.

Not Suitable For

  • Small biotechs/startups — Deals larger than $80 million (minimum deal size) are unrealistic and may be supported by alternatives such as access to Recursion OS.
  • Academic researchers — Commercial B2B platform only – no open-access tools available. Alternatives to commercial tools may include free and/or open-source options.
  • Chemistry-focused discovery — Target ID only – no molecule generation. Exscientia would be more suitable for creating novel chemical compounds.
  • Solo computational biologists — Requires partnerships with enterprise customers to access data. Alternatives to external partnerships include building internal data sets or using cloud-based machine-learning platforms.

Limits Restrictions

Availability
Enterprise biopharma partners only, no public signup
Minimum Deal Size
$80M+ upfront payments reported
Data Access
Federated only - no data leaves hospital systems
Geographic Network
Global but partner-dependent (EU/US focus)
Therapeutic Focus
Oncology, immunology, cardio primary areas
Compliance Requirement
HIPAA/GDPR via federated learning only
Custom Development
Indication-specific AI model training required

Security Compliance

Federated LearningPrivacy-preserving AI training on decentralized hospital data without pooling or transfer
GDPR ComplianceMulti-party data processing compliant across EU research network
HIPAA CompliantUS hospital data access via federated methods protecting PHI
Patient Data PrivacyModels trained without accessing raw patient records - only aggregate insights
Proprietary Data ProtectionPharma IP secured; no data sharing between partners
Multi-Institutional Governance104+ healthcare centers with data use agreements and ethics oversight

Customer Support

Channels
Dedicated for enterprise partnersJoint research programs with pharmaAPI and agent deployment supportInitial discovery via website
Hours
Business hours with 24/7 critical issue escalation for partners
Response Time
Priority partner support; joint steering committees for programs
Satisfaction
High - major renewals with Sanofi ($180M+), BMS ($80M+)
Specialized
Cross-functional teams (AI, biology, clinical, regulatory) per partnership
Business Tier
Executive-level relationship management for biopharma clients
Support Limitations
•Enterprise partners only - no public support tiers
•Custom contracts required before technical support engagement

Api Integrations

API Type
REST API available through agentic infrastructure platform
Authentication
Not publicly specified; likely enterprise OAuth/JWT for pharma customers
Webhooks
No public information on webhook support
SDKs
No official SDKs found on public GitHub or developer portals
Documentation
Limited public documentation; API access via enterprise partnerships (owkin.com)
Sandbox
No public sandbox; enterprise-only access to patient data and AI agents
SLA
Enterprise SLAs available for pharma partners; no public uptime guarantees
Rate Limits
Not publicly disclosed; customized for enterprise clinical workflows
Use Cases
Embed AI agents in healthcare workflows for biomarker discovery, clinical trial design, target identification, drug positioning

Faq

TargetMATCH utilizes multimodal patient data (H&E images, RNA-seq, spatial transcriptomics, clinical data, etc.) to identify top candidate targets in addition to identifying responsive patient subpopulations. The algorithm produces 4,000 AI features per patient and prioritizes targets according to survival rates, essentiality, and the presence of tumor microenvironment factors. Biomedical experts validate the output of the algorithms.

[Reserved for additional content if needed]

Owkin works exclusively under enterprise contracts with pharmaceutical companies. The pricing of the company is not publicly disclosed. To receive a quote for a program of interest you will need to contact the sales department at Owkin. A quote will be generated after discussing your target discovery program and your data access requirements.

Owkin is a developer of clinically-validated multimodal patient data from real-world data (as opposed to laboratory generated) that can be used to identify the best target-patient matches. In addition to providing superior target-patient matching, Owkin’s AI also utilizes spatial omics and tumor microenvironment analysis for better patient stratification.

Yes, Owkin does utilize federated learning to allow them to provide AI-driven insights to their clients without having to centrally store and process their clients’ sensitive patient information. As an example, using this technology, Owkin has been able to create AI models using over 800 hospitals' data for use in identifying cancer treatment options for patients. Federated learning allows Owkin to comply with the various regulatory requirements related to the storage and processing of sensitive patient data.

Owkin’s AI agents are available as APIs that can be easily integrated into the existing health care systems and workflows of their clients. Additionally, Owkin supports the use of Model Context Protocol to ensure seamless integration of the AI agents into the existing systems and workflows. The AI agents developed by Owkin can be utilized for a variety of applications including but not limited to clinical trials, biomarker discovery and patient stratification.

There is no free trial. However, Owkin offers demos and proof-of-concepts for qualified pharma and biotech partners. Interested parties may contact Owkin’s team directly to inquire about accessing their K-Pro Copilot Platform.

Owkin successfully identified three new drug targets for Sanofi using its Discovery AI. The Discovery AI developed by Owkin was shown to perform better than DepMap by 16.8 times and Open Targets by 2.8 times in terms of discovering clinical Phase II targets for six different indications.

Expert Verdict

Owkin is the leading provider of enterprise grade AI solutions for precision drug discovery. It leverages a unique combination of clinically validated algorithms and unmatched multimodal patient data from over 800 hospitals to deliver the most accurate results possible for pharmaceutical clients. Two key engines that represent the core business offerings of Owkin include TargetMATCH and DrugMATCH which were both designed to specifically address the two biggest pain points in the pharma pipeline, namely target identification and patient stratification.

Recommended For

  • Companies looking to optimize their clinical pipelines.
  • Oncology and immunology research and development teams who require patient stratification
  • Biotechs that don't have proprietary patient data to train their AIs
  • Sponsors of clinical trials that need a faster way to discover new biomarkers

!
Use With Caution

  • Smaller biotechs that can't afford enterprise-scale budgets
  • Teams that want an instant Public API or would prefer a partnership model
  • Academic researchers who are looking for open-source tools

Not Recommended For

  • Small molecule startups in de novo drug design
  • Teams that just want to buy off-the-shelf consumer-grade AI tools
  • Researchers that want to do it all themselves -- DIY drug discovery -- but lack expertise in clinical data
Expert's Conclusion

The big pharma companies in the U.S. and Europe spend approximately $2.7 billion each year on developing new drugs for cancer. Owkin is helping them reduce those costs by as much as $2.7 billion per year using clinically-validated AI to discover new targets for treating cancer, and matching patients with those treatments.

Best For
Companies looking to optimize their clinical pipelines.Oncology and immunology research and development teams who require patient stratificationBiotechs that don't have proprietary patient data to train their AIs

Research Summary

Key Findings

Owkin is the first company to use multimodal patient data from over 800 hospitals to discover new AI targets. They recently gave Sanofi three targets they had discovered using Discovery AI -- which outperformed both DepMap (16.8 times) and Open Targets (2.8 times). Their TargetMATCH tool matches the novel targets with the appropriate patient subpopulations, and they will be launching their first AI agent soon, which will enable developers to find biomarkers quickly and accelerate clinical trials. Owkin's platform is designed for large pharmaceutical companies, and focuses on preserving patient data while enabling federated learning in oncology/immunology.

Data Quality

Good - comprehensive technical details from official site and case studies. No public pricing/API docs (enterprise sales model). Proven pharma partnerships validate capabilities.

Risk Factors

!
As a result of being an enterprise-only service, Owkin does not allow non-enterprise developers to evaluate its capabilities.
!
Owkin relies continuously on hospital access to patient data to maintain its service.
!
There are many other companies competing against Owkin in the AI-based drug discovery space.
!
Large pharmaceutical companies typically require long time frames to establish relationships with potential partners in order to engage in joint ventures or collaborations.
Last updated: February 2026

Additional Info

Key Partnerships

Owkin has established a strategic partnership with Sanofi, and as part of this agreement, has developed three new drug targets based upon its AI drug discovery services. In addition, Owkin is working with over 800 hospitals around the world to develop additional AI-based discoveries for new drug targets.

Proven Track Record

Owkin's Discovery AI model has set a new industry benchmark by demonstrating performance that was 19.6 times greater than chance, 16.8 times greater than DepMap, and 2.8 times greater than Open Targets for identifying clinical phase II targets for six different indications.

Latest Innovation

Biology-focused AI agents were launched in January 2026, and are trained on real patient data. These new AI agents include Pathology Explorer, which has demonstrated 23.7 percent higher accuracy, and 5 times fewer parameters than existing models used in clinical research, providing for accelerated clinical research capabilities.

Platform Offering

Owkin K-Pro copilot offers a variety of biomedical research tools including multimodal data + AI. It enables spatial transcriptomics analysis, biomarker discovery, and clinical trial design optimization.

Data Advantage

Data is available for each patient in 6 different modalities with 8 million data points total. Data includes most recent standard-of-care response and spatial omics which represents current clinical practice as opposed to lab-based alternatives.

Alternatives

  • •
    Insilico Medicine: An end-to-end AI tool for de novo design of novel compounds and targets using generative AI. More focused on generating entirely new compounds from scratch as opposed to Owkin's use of patient stratification for finding best treatments. Best for companies who want to create all their novel compound discovery internally.
  • •
    Recursion Pharmaceuticals: A phenomics platform that maps biology at the cellular level through AI. Recursion has a large proprietary image dataset however less focus on clinical patient data. Best for discovering how a cellular process works at the mechanistic level.
  • •
    Exscientia: A precision chemistry design platform that uses AI to generate small molecules. Exscientia has strong capabilities in hit-to-lead optimization but lacks the clinical data component of Owkin's platform. Best for companies interested in doing the chemical part of discovery.
  • •
    BenevolentAI: A knowledge graph AI for identifying targets and repurposing drugs. Benevolent has a solid understanding of biomedical reasoning however a much smaller dataset than Owkin has of over 800 hospitals. Best for generating hypotheses.
  • •
    Schrödinger: Computational physics-based platform for designing and predicting ADME properties of molecules. Structure-based design is an industry standard but requires structural information as opposed to a data-driven approach. Best for optimizing target-class specific.

AI-Driven Enrollment Performance

pending %
Time-to-Enrollment Reduction
pending days
Patient Screening Cycle Time
pending %
Automated Patient Matching Accuracy
pending %
Protocol Amendment Rate Reduction
pending days
Data Query Resolution Time

AI/Machine Learning Capabilities

Target Discovery (TargetMATCH)

Tools for AI-based identification of novel therapeutic targets and subpopulations of patients with multiple diseases using multimodal patient data (H&E, WES, RNA-seq, spatial transcriptomics).

AI Drug Positioning (DrugMATCH)

AI-based aggregation of causal evidence from multimodal data and prior knowledge to determine novel disease indications for existing drugs and patient subgroups.

Multimodal Data Integration

Aggregation of clinical data, histopathology images, genomics, spatial transcriptomics, and >8M data points per patient across 6 modalities

Subgroup Discovery

AI-enabled biomarker discovery for personalizing therapy through use of explainable models in order to enable stratification of patients and precision medicine

Biology-Focused AI Agents

Agents were trained on real-world patient data from over 800 hospitals for identifying biomarkers, interpreting datasets, and making decisions regarding clinical trials

Pathology Explorer Agent

Highly accurate pathology analysis with an average of 23.7% better classification, a 5X reduction in model parameterization, significantly reduced computational time from weeks to hours

Knowledge Graph Analysis

Extracts the relationships between genes, diseases, drugs and patients; continually updates on the results of clinical trials

Primary AI Use Cases in Clinical Trials

Use CaseIndustry ChallengeAI SolutionBusiness Impact
Target IdentificationHigh attrition rates (90%); $2.7B median cost per oncology/immunology drugTargetMATCH engine analyzes multimodal patient data to discover novel targets and patient subgroups19.6x better than random; 16.8x better than DepMap for phase 2 targets; reduces late-stage failures
Drug RepositioningLimited translatability of preclinical models; disease heterogeneityDrugMATCH identifies new indications/subgroups via causal evidence aggregation from patient dataMatches right drug to right patient; accelerates precision medicine development
Patient StratificationIncomplete understanding of disease biology and patient variabilitySubgroup discovery with multimodal biomarkers and tumor microenvironment analysisIdentifies actionable clinical biomarkers; improves trial success rates
Clinical Trial DesignSlow target-to-drug conversion; poor clinical relevance of targetsAI-powered feasibility assessment and patient matching from real-world hospital dataTargets matched to indications in 2 weeks vs 6 months; de-risks trial design
Biomarker DiscoveryManual analysis misses complex biological signalsSpatial transcriptomics features and 4,000+ AI features per patient for TME analysisPrioritizes novel targets overlooked by traditional methods

EHR, EDC & Data Integration Capabilities

Multimodal Patient Data
Deep clinical data, H&E histopathology, WES, RNA-seq, single-cell RNA-seq, spatial transcriptomics; 8M+ data points per patient
Public & Federated Datasets
TCGA pan-cancer, MESOMICS disease-specific, OpenTargets, GTEx, ClinicalTrials.gov, ChEMBL
Real-World Data Sources
Curated multimodal dataset from 800+ hospitals globally; latest standard-of-care response data
Data Modalities Supported
6 modalities including spatial omics, genomics, clinical records, imaging
AI Feature Extraction
4,000 AI features per patient; 50 AI features per target; 80+ dimensional gene-patient embeddings
Knowledge Integration
Large Language Models for causal evidence aggregation; biomedical knowledge graphs
Platform Infrastructure
Owkin K-Pro co-pilot platform with APIs for multimodal analysis and feature extraction

Regulatory & Quality Compliance Status

21 CFR Part 11 ComplianceFederated learning preserves privacy; specific Part 11 validation not detailed
ALCOA+ Data IntegrityReal patient data with clinical outcomes tracking; audit-ready via federated approach
GCP (Good Clinical Practice) ComplianceSupports clinical trial optimization; GCP-specific certification not specified
GDPR Data ProtectionFederated learning on 800+ hospital data preserves privacy; no data centralization
HIPAA ComplianceReal-world patient data handling with privacy safeguards via federated infrastructure
ISO 27001 Information SecurityEnterprise-grade security for multimodal clinical datasets
FDA AI/ML SaMD GuidanceAI agents compatible with healthcare workflows; clinical validation ongoing
Clinical Trial Support Validation19.6x superior target discovery performance validated across 6 indications
Ethical AI SafeguardsClinically validated models with bias detection capabilities
Federated Learning GovernancePrivacy-preserving AI trained across global hospital network

Digital Health & Patient-Centric Capabilities

Patient Subgroup Matching

The AI identifies the most effective patient subgroups that can be targeted by drug or other treatments to improve success of clinical trials

Precision Stratification

Multimodal biomarkers identify patient cohorts which can be used to assign them to trials tailored to their individualized needs

Real-World Outcome Prediction

Models trained on the responses of actual patients predict treatment efficacy and variability among individuals

Leading AI Vendors & Specialized Capabilities

Vendor/PlatformPrimary SpecializationKey CapabilitiesIntegration Strength
Owkin (TargetMATCH)AI Target DiscoveryMultimodal patient data analysis (spatial omics, single-cell, histopathology); 19.6x better than random, 16.8x vs DepMap; target/patient subgroup pairingOwkin K-Pro platform; 800+ hospital federated data; API service with Model Context Protocol
Owkin (DrugMATCH)Drug RepositioningAI drug positioning for new indications/subgroups; causal evidence from knowledge graphs + patient dataSeamless integration with discovery workflow; LLM-enhanced reasoning
Owkin Biology AI AgentsClinical Research AgentsPathology Explorer (23.7% better accuracy); biomarker ID, dataset interpretation, trial decision supportEmbeddable in healthcare workflows; continuously improving with new patient data

AI-Driven Cost & Timeline Benefits

2 weeks
Drug Discovery Timeline Compression
19.6 x
Target Discovery Effectiveness vs Random
16.8 x
Target Discovery vs DepMap (Phase 2)
2.8 x
Target Discovery vs Open Targets
weeks to hours
Pathology Analysis Speed Improvement
23.7 % better
Pathology Classification Accuracy
2.7 Billion USD
Oncology Drug Development Cost Context

Expert Reviews

📝

No reviews yet

Be the first to review Owkin!

Write a Review

Similar Products

Unlearn.AI
unlearn.ai

Interesting Products

Palm
palm.io
PancakeSwap
pancakeswap.finance
Papr.ai
papr.ai