AAPS PharmSci 360 2026: AI Tools, Regulations, and Preparedness

Sanjay Konagurthu, PhD, senior director of Science and Innovation at Thermo Fisher Scientific, spoke with PharmTech at AAPS PharmSci 360 2026 about what goes into building a successful artificial intelligence (AI) strategy for oral drug product development. He also provided insight into how regulators expect AI evidence to be handled, and why data remains the biggest obstacle to the adoption of AI.

Success depends on well-structured access to data and close collaboration across three groups, according to Konagurthu. Formulation scientists design the dosage form, weighing excipient choice, API properties, and unit operations. Analytical scientists characterize the formulation’s physical and chemical attributes. Data scientists turn that data into models that can be validated, he explained.

On regulation, he noted that AI evidence supporting drug product quality or other regulatory decisions should be fit for purpose and backed by an appropriate level of validation. He pointed to the FDA’s recent nonbinding draft guidance and to shared FDA and European Medicines Agency principles emphasizing clear intended use, sound data management and governance, performance assessment, human oversight, and lifecycle management. Sponsors using these tools, he said, should keep an auditable record of data, model versions, validation results, limitations, human review, and any postvalidation changes.

Konagurthu said he expects near-term progress in optimizing the manufacturability and dissolution of tablets and capsules, and in formulation screening for poorly soluble small molecules through platforms such as Thermo Fisher’s Quadrant 2. These platforms produce measurable outcomes that can guide the next experiment. Predicting clinical exposure from an API’s structure and properties remains the “holy grail,” and his team is working on it. To prepare, he advised companies to standardize formulation and dissolution data and connect lab records to process and material attributes.

Data access is still the biggest hurdle, he stressed. In his view, well-organized data are 80% to 90% of the task, with model building and validation coming later. Although models are advancing rapidly, he said that he expects the industry to make real progress on data access within 2 to 5 years.

Konagurthu gave the keynote presentation, 3D Printing: Driving Digital Transformation from Dosage Form Design to Product Development and Manufacturing, at AAPS PharmSci 360, which is being held from October 25-28 in New Orleans.

About the Speaker

Sanjay Konagurthu, PhD, senior director, Science and Innovation at Thermo Fisher Scientific is a chemical engineer with more than 25 years of industry experience. He specializes in small-molecule oral drug delivery, including solubility and bioavailability enhancement, spray drying, hot-melt extrusion, nanotechnology, and modified- and controlled-release dosage forms, from early development through commercialization. He also leads Thermo Fisher’s predictive modeling and simulation team, which applies modeling to solubility enhancement, predictive stability, packaging shelf-life simulation, process development, and pharmacokinetics.