Review: Multi-agent systems for chemical engineering

Our review "Multi-agent systems for chemical engineering: a review and perspective" by Sophia Rupprecht, Qinghe Gao, Tanuj Karia and Artur M. Schweidtmann was published in Current Opinion in Chemical Engineering.

Multi-agent systems based on large language models (LLMs) split complex workflows into teams of collaborating agents, each with specialized knowledge and tools. This mirrors how chemical engineering work is organized: in teams of experts who work across scales, from molecules and thermodynamics to process units, whole plants and supply chains.

The review surveys the current state of multi-agent systems in chemical engineering. Early studies show promising results, but open challenges remain: architectures tailored to engineering tasks, the integration of heterogeneous data such as flowsheets, P&IDs and simulation files, foundation models with domain-specific modalities, and strategies to ensure transparency, safety and low environmental impact. In our vision, human engineers stay at the center and supervise a team of digital agents.

More on this topic on our multi-agent systems research page.

Paper: S. Rupprecht, Q. Gao, T. Karia, A. M. Schweidtmann (2026). Multi-agent systems for chemical engineering: a review and perspective. Current Opinion in Chemical Engineering, 51, 101209. https://doi.org/10.1016/j.coche.2025.101209 (preprint: arXiv:2508.07880)