- Phlow Corp. and Enveda have completed a joint pilot using AI to accelerate API manufacturing process development.
- The collaboration generated nearly 20,000 reactions in three months, improving AI predictions of reaction kinetics, yields, and purity.
Phlow Corp. has announced progress in AI-driven Active Pharmaceutical Ingredient manufacturing following a pilot collaboration with Enveda. The work focused on rapidly selecting API synthesis routes by training AI models on high-quality, internally generated reaction data, with the aim of shortening development timelines.
The pilot, launched last year, resulted in the generation and analysis of nearly 20,000 unique reactions in three months. According to the companies, the resulting dataset is larger than any publicly available equivalent and comparable in scale to proprietary datasets built over decades by pharmaceutical manufacturers. Models trained on this data delivered more accurate yield predictions than those trained on public datasets alone.
“This is about setting a new standard for pharmaceutical development and manufacturing. By combining Phlow’s advanced API development and manufacturing expertise with Enveda’s leading AI-driven insights and deep learning capabilities, we are creating a first-of-its-kind, innovative platform that predicts the optimal synthesis route the first time, every time.”
Eric S. Edwards, M.D., Ph.D., Co-Founder and CEO of Phlow
The collaboration evaluated traditional machine learning against deep learning approaches, including graph neural networks that model molecules as connected atoms and bonds. This method was reported to better reflect real-world chemistry and deliver stronger predictive performance for reaction optimisation and future retrosynthetic planning.
Phlow said the ability to rapidly generate and integrate reaction data into AI models supports faster route optimisation, improved scalability, and more efficient API development.
