What the study found
The study found that an agentic artificial intelligence (AI) system can generate dynamic decision tables from business process model and notation (BPMN) models. The findings indicate that the system can suggest decision-table values that humans might not intuitively identify and can turn ambiguous process paths into more precise decisions.
Why the authors say this matters
The authors conclude that this approach could reduce human error, inconsistencies, and cognitive biases in manual decision management. They also say it may help build intelligent, adaptive decision support systems for mission-critical environments and autonomous decision modeling that can adjust to changing business requirements.
What the researchers tested
The researchers developed a novel AI-based system that uses large language models within an agentic AI framework to analyze BPMN processes, identify decision points, and produce optimized decision models and notation (DMN) tables. The system used agents for BPMN analysis, decision extraction, rule generation, and validation, coordinated through a ReAct engine with retrieval-augmented generation (RAG) capabilities.
What worked and what didn't
In experimental evaluation of critical applications, the system enhanced decision-making by suggesting decision tables with values that humans might not intuitively identify. It was reported to optimize processes by transforming ambiguous paths into precise decisions, and it was particularly effective at identifying non-obvious decision criteria and threshold parameters.
What to keep in mind
The abstract does not describe detailed limitations, sample size, or performance metrics. It also does not provide enough information to compare this system directly with other approaches beyond the general claim that it addresses a gap in the literature.
Key points
- An agentic AI system was developed to generate dynamic decision tables from BPMN models.
- The system uses large language models, ReAct coordination, and retrieval-augmented generation.
- Experimental evaluation in critical applications found it could suggest values humans might not intuitively identify.
- The authors say the approach may reduce human error, inconsistencies, and cognitive biases.
- The abstract does not report detailed limitations or performance metrics.
Disclosure
- Research title:
- Agentic AI generated decision tables from BPMN models
- Authors:
- Sourour Meddeb, Selma Batti, Habib Fathallah
- Institutions:
- University of Carthage
- Publication date:
- 2026-01-21
- OpenAlex record:
- View
- Image credit:
- Photo by RDNE Stock project on Pexels · Pexels License
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.