Goal driven approach improves process variant grouping with large language models
Goal-driven Variant Categorization
Artificial IntelligenceDatabasesSoftware Engineering
Summary
Companies often analyze different versions of their business processes but struggle to group them in meaningful ways that reflect their goals. This paper presents a method where a company first creates a model of its goals. Then, each process version is described in text, and a large language model (LLM) uses that description plus the goals to categorize the process variant. This technique helps connect detailed process actions to business priorities, making grouping more purposeful and easier to adjust. The authors tested this on multiple datasets and found it created different and goal-aligned groups compared to traditional automatic clustering.
What this means in practice
- •For business process analysts: Generate process variant groups aligned with specific organizational goals to simplify interpretation and decision-making.
- •For enterprise software developers: Integrate goal-driven process variant categorization using LLMs to enhance process mining analytics tools with more meaningful classifications.$Commercial implications: This method enables developers to create smarter process mining software that sells better business insight features to enterprises.
Authors
Daniel Calegari, Daniel Amyot
Abstract
Process discovery rarely yields a single coherent process structure. For analysis, a common step is to cluster process variants based on structural similarity and then assign business meaning to the resulting groups. Since these partitions are not derived from the organization's goals, analysts must manually interpret and consolidate variants into business-meaningful categories. This judgment-intensive step becomes increasingly difficult as the number and complexity of variants grow. In this paper, we propose a goal-driven approach to variant categorization that reverses this workflow. We first author an organization's goal model that predefines the categorization axis. Each variant is transformed into a textual narrative describing its behavior, and a Large Language Model (LLM) interprets it in the context of the goal model and assigns the variant to the most appropriate category. LLM-based semantic reasoning connects low-level process behavior with analyst-defined business goals. We instantiate this approach end-to-end and evaluate it on three public logs differing substantially in scale and behavioral diversity. Goal-model guidance yields partitions that differ from those produced by unguided induction and respond to controlled edits to the declared alternatives, at the cost of authoring a goal model.