The Birth of Agentic AI for Chemical Processes – Hybrid Models, Process Simulators, and the Future of Autonomous Operations
Jay Yun, Founder & CEO of SIMACRO
Chemical process industries are entering a new phase where AI is moving beyond prediction and assistance into active engineering execution. Jay Yun, Founder & CEO of SIMACRO, explores how agentic AI can become a true operational partner by combining process simulators, hybrid models, and engineering intelligence within safe industrial environments.
Rather than positioning AI as a replacement for physics-based engineering, the presentation shows how SIMACRO combines first-principles models with AI quality layers to improve the prediction of complex operational variables that conventional simulators struggle to capture alone. Using polymer colour prediction as a practical example, SIMACRO demonstrates how hybrid modelling can strengthen product quality forecasting, operational confidence, and decision-making in chemical processes.
A major focus of the session is the operationalisation of industrial AI through SIMACRO’s PMV and AIMIE platforms. PMV acts as the industrial ModelOps layer, connecting plant data, validating and staging models, and creating a controlled runtime environment for operational workflows. AIMIE then uses those validated process and hybrid models as tools for mission-driven reasoning, recommendation, and action workflows.
The presentation also highlights why safe autonomy requires what SIMACRO describes as a “digital handle” between AI and operational technology environments. Rather than allowing unrestricted AI control, the session outlines a staged approach to autonomy built around monitoring, recommendation, simulation, validation, and controlled automation. The result is a vision for agentic AI that is grounded not only in language models, but in engineering rigour, validated process knowledge, and operational trust.
