Abstract
Process models are the basis for a wide range of critical activities within an organisation. It is not surprising then that process models, and the act of process modelling, have been the focus of much research over the last two decades. Recent research indicates, however, that common process modelling notations lack sufficient representation for capturing business rules. Although the need for business processes and business rules to be modelled in an integrated manner is well established, the body of knowledge on integrated modelling of the two is limited. In this paper our aim is to review the state of related research and develop a research agenda, based on a systematic review of related literature, to advance research in this field. We present a consolidated view of the benefits of rule and process model integration, together with an overview of current related approaches, and a research agenda going forward.
The Challenge: Separating the 'How' from the 'What'
In many organizations, business processes and business rules are treated as separate entities. Process models (like BPMN) excel at showing the flow of activities—the "how"—but often fail to capture the complex logic and constraints—the business rules—that govern those activities.
When these rules are "hidden" in program code, separate databases, or manual handbooks, several problems arise:
- Incompleteness: Models don't reflect the full operational reality, leading to non-compliant execution.
- Rigidity: Changing a rule might require a complete overhaul of a hard-coded process.
- Communication Gaps: Designers and developers lack a "shared truth," leading to errors during system implementation.
The authors tackle the lack of a consolidated body of knowledge regarding how to best merge these two worlds.
Systematic Literature Review Methodology
The authors conducted a Systematic Literature Review (SLR) following a rigorous multi-stage protocol to analyze the academic corpus.
- Search Strategy: They screened over 34,000 articles from core Information Systems (IS) and Computer Science (CS) journals and conferences published between 1990 and 2014.
- Filtering: Using automated keyword indexing and manual screening, they narrowed the field to 66 core papers.
- Coding: They used NVivo 10 to systematically code these papers based on motivations, integration approaches, and research contribution types.
Evolution of the Field and Integration Paradigms
The review revealed that research activity spiked following the release of major standards: BPMN in 2004 and SBVR (Semantics of Business Vocabulary and Business Rules) in early 2008.
The authors identified four primary technical "flavors" of integration, each with distinct trade-offs:
- Extending and Merging: Creating new notations or combining meta-elements from different languages (e.g., rBPMN). This is the most common approach.
- Embedding: Hiding complexity by encapsulating rules within a single process element, such as a "Decision Table" task.
- Annotation: Attaching textual or formal metadata (like SBVR or OCL) to existing process symbols.
- Transformation: Using algorithms to automatically convert a procedural model into a rule-based one (or vice versa).
Key Results and the Evaluation Gap
The paper identifies five core drivers for integration: Process Flexibility, Model Completeness, Improved Governance/Risk/Compliance (GRC), Better Understandability, and Interoperability. While "Extending and Merging" is the dominant research trend (18 papers), the field shows a significant maturity imbalance.
A major finding is the Evaluation Gap: while many researchers propose new methods (17 papers) or implementations (14 papers), very few (only 8 out of 66) actually evaluate whether these integrated models are easier for humans to understand or more effective in practice.
A Research Agenda for Integrated Modeling
Based on the identified gaps, the authors propose a research agenda to move the field forward:
- Empirical Evidence: Studies are needed to prove that integrated models actually improve "ease of use" and "understandability" for business users.
- Theoretical Foundations: Defining what "model completeness" actually looks like in a mathematical or ontological sense.
- Change Management: Investigating how to maintain consistency when a single rule change affects hundreds of integrated processes.
- Decision Frameworks: Developing frameworks to help architects decide which rules to integrate and which approach (embedding vs. annotation, etc.) to use based on the organizational context.
Summary updated