This study investigates the feasibility of developing an automated solution that can generate dynamic decision tables from business process model and notation (BPMN)models using agentic artificial intelligence (AI). This work purpose is to reduce human error, inconsistencies and cognitive biases that can be introduced by traditional decision management in BPMN environments, which is often achieved by manually creating decision model and notation (DMN) decision tables (Richter et al., 2025).
A novel AI-based solution is developed to generate dynamic decision tables from BPMN models. The proposed system integrates large language models within an agentic AI framework that autonomously analyses BPMN processes, identifies decision points and produces optimized DMN tables. The system employs agents for BPMN analysis, decision extraction, rule generation and validation, coordinated through a ReAct (Reasoning + Acting) engine with retrieval-augmented generation (RAG) capabilities (Zhang et al., 2025; Braunschweiler et al., 2025).
Experimental evaluation of critical applications demonstrated that the system enhances decision-making by suggesting decision tables with values that humans might not intuitively identify. The system optimizes processes by transforming ambiguous paths into precise decisions. The framework is particularly effective in identifying non-obvious decision criteria and threshold parameters, resulting in significant process automation improvements.
This approach establishes the foundation for intelligent, adaptive decision support systems within mission-critical environments and autonomous decision modeling that can dynamically adapt to evolving business requirements.
This innovative approach represents the implementation of agentic AI specifically designed for automated DMN decision table generation from BPMN models, addressing a gap in the literature.
