As molecular complexity continues to increase, pharmaceutical companies are increasingly turning to AI in drug development to predict and optimize critical properties such as solubility and bioavailability. By identifying patterns in chemical data, AI can estimate parameters such as pKa and logP, as well as predict formulation-dependent effects on absorption.
These capabilities allow for more informed formulation development, potentially reducing the need for extensive laboratory experiments. However, within the CDMO landscape, this technological momentum is constrained by a significant confidentiality barrier.
The Challenge of Data Confidentiality
While the potential for AI is immense, CDMOs face unique hurdles. Client-owned molecular structures cannot simply be uploaded to public or cloud-based AI platforms without risking intellectual property. This creates a fundamental tension between AI capability and compliance.
Developing secure, in-house AI platforms may appear to be a natural solution, yet this approach introduces additional complexities. High-performance models depend on large, heterogeneous datasets, which in a CDMO environment inevitably consist of multiple client molecules.
This raises challenging questions around data ownership and the risk that proprietary features become embedded within shared models. Even when structures are anonymized, the risk of reverse identification remains.
Moreover, internal AI models must be validated against real molecular data. Without explicit client authorization, rarely granted under standard MSAs, project compounds cannot be used for model training or benchmarking. This constraint slows model maturation and limits the immediate value of internally developed AI tools, despite significant investment.
Explainable AI in Drug Development and the Need for Validation
Even where advanced AI platforms are accessible, their outputs should be treated as hypothesis-generating rather than definitive. Solubility and bioavailability are governed by complex, context-dependent interactions that cannot yet be resolved computationally alone.
Furthermore, the “black-box” nature of many AI systems limits their standalone credibility because predictions lacking mechanistic justification fail to meet established evidentiary standards.
As a result, there is growing interest in explainable AI in drug development (xAI) approaches, which aim to improve transparency and interpretability. While these technologies may strengthen confidence in AI-generated predictions by offering better insight into model logic, they do not eliminate the need for human scientific oversight. Formulation scientists continue to play a critical role in evaluating AI outputs, identifying potential biases, and designing targeted experiments to confirm or refute computational predictions.
A Future Where AI Supports Formulation Science
As we move into the Pharma 4.0 era, AI is expected to play an increasingly important role in pharmaceutical development. Emerging technologies that protect data sovereignty, alongside maturing regulatory expectations, may help overcome existing hurdles.
Ultimately, AI is likely to enhance scientific decision-making by directing scientists toward more promising and sophisticated development strategies. Nevertheless, empirical testing of predicted solubility, bioavailability, and overall product performance remains indispensable.
AI models should be viewed as predictive tools that require validation, rather than a substitute for laboratory science. When applied judiciously, AI has the potential to expedite development timelines by enabling more efficient allocation of formulation resources, thereby reducing time-to-market.
References
- Ann M. Thayer, “Finding solutions: Custom manufacturers take on drug solubility issues to help pharmaceutical firms move products through development” (2010).
- Zeqing Bao et al., “Towards the prediction of drug solubility in binary solvent mixtures at various temperatures using machine learning” (2024). Available at: https://link.springer.com/article/10.1186/s13321-024-00911-3
- Sanjay Konagurthu, Thermo Fisher Scientific, “Revolutionizing drug development: AI-driven solutions for poor solubility and bioavailability” (2024). Available at: https://www.patheon.com/us/en/insights-resources/blog/ai-driven-drug-development-for-poor-soubility-and-bioavailability.html
- Alessio Zoccoli et al., Drug Target Review, “Making sense of AI: bias, trust and transparency in pharma R&D” (2025).
- Jirapornchai Suksaeree, “A review of artificial intelligence (AI)-driven smart and sustainable drug delivery systems: A dual-framework roadmap for the next pharmaceutical paradigm” (2025). Available at: https://www.mdpi.com/2413-4155/7/4/179
- Sanjay Konagurthu, European Pharmaceutical Review, “Applying AI to enhance drug formulation and development” (2026).
- FDA and EMA, “Guiding principles of good AI practice in drug development” (2026). Available at: https://www.fda.gov/about-fda/artificial-intelligence-drug-development/guiding-principles-good-ai-practice-drug-development
Subject Matter Experts
Timothy Pas
AD, Formulation & Production
Dr. Timothy Pas is Associate Director, Formulation Development and Production at Ardena. He leads a team supporting formulation development and GMP manufacturing activities for early clinical programs. Since joining Ardena, Timothy has progressed from Formulation Scientist to Group Leader and currently Associate Director. Throughout his career at Ardena, he has developed extensive expertise in oral and parenteral formulations, formulation development strategies, and GMP manufacturing operations supporting early-stage pharmaceutical development programs. Timothy obtained a Master’s degree in Drug Development and earned his PhD in Pharmaceutical Technology from KU Leuven, specializing in enabling technologies and galenic principles.
Laure Descamps
Senior Formulation Scientist
Laure Descamps is Senior Formulation Scientist at Ardena, specializing in solid dosage development and process understanding. She supports formulation design, process optimisation, and early-phase drug product development for complex pharmaceutical programs. At Ardena, Laure applies her expertise to developing robust formulation strategies that support the efficient advancement of complex molecules toward clinical evaluation. Her work focuses on formulation development and process understanding across early-stage drug product development activities. Laure obtained her Master’s degree in Drug Development and Industrial Pharmacy and earned her PhD from Ghent University’s Laboratory of Pharmaceutical Process Analytical Technology. Her doctoral research focused on tabletting behaviour, material characterisation, and compression challenges in collaboration with industry partners.
