Petri net technique enables cross-organisational process mining from message logs

Pieter Kwantes and Jetty Kleijn from the University of Amsterdam introduce a novel approach to cross-organisational process mining that relies solely on message logs—collections of recorded output and input events between two processes. Their work, published on arXiv (2607.26917), builds on the Enterprise net and Industry net framework, where individual business processes are modeled as Enterprise nets and their interactions as Industry nets. Previous methods required access to event logs (finite sequences of executed actions) from each organization, which is often impractical due to privacy or accessibility constraints.

This paper addresses that gap by showing how message logs can be transformed into an event log that preserves the causal order of message events. This transformed log is then fed into existing algorithms for single-process discovery to synthesize an Industry net—a model of asynchronously communicating processes. The technique was presented as an extended abstract at the third International Workshop on Collaboration Mining for Distributed Systems (COMINDS 2024), co-located with ICPM 2024 in Copenhagen. The work leverages established Petri net theory to ensure the synthesized model accurately reflects the underlying communication patterns.

By eliminating the dependency on internal event logs, the method opens up process mining to scenarios where organizations wish to collaborate but cannot share proprietary execution data. Potential applications include supply chain analysis, multi-party financial transactions, and healthcare coordination—any domain where inter-organizational message exchanges are recorded. The paper provides a formal transformation and demonstrates its feasibility, paving the way for practical tools that can analyze cross-company workflows without compromising data sovereignty.