Recent advancements in Artificial Intelligence (AI), particularly Large Language Models (LLMs), have enhanced organizations' ability to reengineer business processes by automating knowledge-intensive tasks. This automation drives digital transformation, often through gradual transitions that improve process efficiency and effectiveness. To fully assess the impact of such automation, a data-driven analysis approach is needed — one that examines how traditional and AI-enhanced process variants coexist during this transition. Object-Centric Process Mining (OCPM) has emerged as a valuable method that enables such analysis, yet real-world case studies are still needed to demonstrate its applicability. This paper presents a case study from the insurance sector, where an LLM was deployed in production to automate the identification of claim parts, a task previously performed manually and identified as a bottleneck for scalability. To evaluate this transformation, we apply OCPM to assess the impact of AI-driven automation on process scalability. Our findings indicate that while LLMs significantly enhance operational capacity, they also introduce new process dynamics that require further refinement. This study also demonstrates the practical application of OCPM in a real-world setting, highlighting its advantages and limitations.

Object-centric process mining (OCPM) is an emerging research area that aims to analyze processes involving multiple object types (for instance, orders, items, and deliveries in an order-handling process) with complex intertwined relations captured in a richer format than traditional event logs. The richness of these data, as represented in the Object-Centric Event Log (OCEL) standard, often causes existing discovery algorithms to generate models overloaded with information, exceeding the cognitive limits of users, and reducing their practical usefulness. To address this challenge, we introduce Object-Centric Causal Nets (OCCN) together with an edge-abstraction technique that simplifies the discovered model by merging similar flows across object types. While OCCN provides native support for concurrency and choice, the edge abstraction is essential for reducing visual clutter and producing simpler yet expressive models. A Python implementation is provided, and a comparative evaluation against Object-Centric Petri Nets and Object-Centric Directly-Follows Graphs shows that OCCN with edge abstraction yields models that are easier to understand and more effective in enabling users to identify workflow patterns.

Process mining allows organizations to discover, monitor, and improve processes from event data. Traditional methods focus on discovering models from a single object type perspective, e.g., `orders' in the Order-to-Cash process, thus limiting analysis, with the danger of getting incomplete or misleading findings. Object-Centric Process Mining (OCPM) addresses this problem by analyzing processes from multiple perspectives like `orders' and `delivery'. However, current object-centric algorithms discover complex models that do not deal effectively with noise, concurrency, and choices or are difficult for stakeholders to understand.

This paper introduces a new process mining method to discover Object-Centric Causal Nets (OCCN). This new method extends Causal Nets to enable object-centric analysis, handle concurrency and choice behavior, and produce simpler, more interpretable models by merging redundant paths in process models using the edge-coarse-graining technique. We implemented this method in Python, which is used through a user study, where we compared discovered OCCN and Object-Centric Petri Nets. The result shows that OCCN models are more intuitive and comprehensible, enabling users to recognize patterns better.

Digital transformation allows public organizations to create value for diverse stakeholders. To improve organizational capabilities, internal changes in processes are often needed. However, resistance can pose challenges to such changes, especially when new technology is to be used by staff. We propose a framework that combines low-code tools with a new object-centric process mining technique, called OCCN, to empower staff to identify requirements for such changes and apply them. To evaluate the feasibility of the framework, we developed a prototype, which was evaluated using semi-structured interviews in a tax administration agency. The results show that the framework can enable new technology adoption by public servants for value creation to both internal and external stakeholders.

Object-Centric Process Mining (OCPM) enables business process analysis from multiple perspectives. For example, an educational path can be examined from the viewpoints of students, teachers, and groups. This analysis depends on Object-Centric Event Data (OCED), which captures relationships between events and object types, representing different perspectives. Unlike traditional process mining techniques, extracting OCED minimizes the need for repeated log extractions when shifting the analytical focus. However, recording these complex relationships increases the complexity of the log extraction process. To address this challenge, this paper proposes a methodology for extracting OCED based on PM2, a well-established process mining framework. Our approach introduces a structured framework that guides data analysts and engineers in extracting OCED for process analysis. We validate this framework by applying it in a real-world educational setting, demonstrating its effectiveness in extracting an Object-Centric Event Log (OCEL), which serves as the standard format for recording OCED, from a learning management system and an administrative grading system.

Analyzing process data at varying levels of granularity is important to derive actionable insights and support informed decision-making. Object-Centric Event Data (OCED) enhances process mining by capturing interactions among multiple objects within events, leading to the discovery of more detailed and realistic yet complex process models. The lack of methods to adjust the granularity of the analysis limits users in leveraging the full potential of Object-Centric Process Mining (OCPM). To address this gap, we propose four OnLine Analytical Processing (OLAP) operations: drill-down, roll-up, unfold, and fold, which enable changing the granularity of analysis when working with Object-Centric Event Log (OCEL). These operations allow analysts to seamlessly transition between detailed and aggregated process models, facilitating the discovery of insights that require varying levels of abstraction. We implemented these operations in an open-source Python library, making it available for researchers and practitioners to use in practice. This approach can empower analysts to perform more flexible and comprehensive process exploration, unlocking actionable insights through adaptable granularity adjustments.

Event logs play a fundamental role in enabling data-driven business process analysis. Traditionally, these logs track events related to a single object, known as the case, limiting the scope of analysis. Recent advancements, such as Object-Centric Event Logs (OCEL) and Event Knowledge Graphs (EKG), capture better how events relate to multiple objects. However, attributes of objects can change over time, which was not initially considered in OCEL or EKG. While OCEL 2.0 has addressed some of these limitations, there remains a research gap concerning how attribute changes should be accommodated in EKG and how OCEL 2.0 logs can be transformed into EKG. This paper fills this gap by introducing Temporal Event Knowledge Graphs (tEKG) and defining an algorithm to convert an OCEL~2.0 log to a tEKG.

Methodological attempts to accommodate data from educational technology (EdTech) systems focus, to a significant extent, on the quality of information over the quality of the raw data. Popular data-driven methodologies, such as learning analytics and educational data mining rely on data collected from EdTech systems as their point of departure. However, contemporary research argues that high-quality analytics systems for educational decision-making should begin by assessing the quality of raw data, including efforts to identify - or “datafy” - previously unquantified aspects of the learners and the environments. This paper argues the need to datafy educational processes to optimise the reliance of data for enhancing education. With the assumption that high-quality data leads to high-quality analytics and that the quality of raw educational data can be improved through careful datafication, this scoping review explores three research questions: How is datafication defined in the education domain? What is its purpose? How has datafication been implemented in the literature?

The methodological approach is a scoping review using the PRISMA framework, yielding a selection of 153 articles for analysis. Three researchers conducted the selection process: two independently screened articles, categorizing them as “yes,” “no,” or “maybe,” while the third resolved conflicts. The articles were analyzed using a quantitative approach focussing on the terminology and semantics of the description of datafication in the selected articles.

Findings reveal that articles positions or mostly present, datafication, do not necessarily with the same underlying meaning. Although the terminological point of reference for datafication is Big Data by Schoenberger, et. al., most articles presented the concept closer to the autonomous harbouring of data in EdTech systems instead. An interesting outcome was how little the original description of datafication has changed throughout time. The meaning of the definition was consistence throughout the literature, irrespective of the time, however, the articles heavily focussing on datafication belong to the past 7 years, most of them are after 2023, a decade after its original definition. Thus, this study mapped different interpretations (or points of view) as a summary rather than finding a unified definition. Furthermore, there is no evidence for the validity and completeness of the term qualifying as a concrete definition. However, due to the advancement of technology, where almost any human physical action can be measurable, and recorded through technology, datafication is tapping the ethical boundaries, which may either be considered or included in the definition.

Based on the outcomes, this study brings into the spotlight a gap in the literature regarding the quality of raw data and the role of datafication towards a sustainable and data-driven educational environment. It underscores the potential of datafication to enhance insights into teaching, learning, and education more broadly. Additionally, the outcomes call for deeper investigation into datafication’s impact on user-centricity, privacy, and ethical considerations for preserving data subject privacy when optimising processes through datafication.

Drug-drug interactions (DDIs) pose significant risks to patients, ranging from adverse effects to fatal outcomes. Preventing these issues depends on providing caregivers with timely information on DDIs and offering viable alternative options. Currently, there is a gap in the formal specifications of systems designed to alert caregivers about potential DDIs. This gap hinders the development of further support, such as algorithms that can recommend alternative drugs.

This study adopts the Design Science approach, defining a formal knowledge graph to capture DDIs. Then, algorithms are defined to identify drug interactions and suggest alternative medications with less severe consequences. As a proof of concept, we implemented our approach using Neo4j and Python, transforming data from the Swedish DDIs database.

The implementation was applied to real care session data in the healthcare region of Stockholm for a randomly selected day, focusing on instances where caregivers prescribed drugs with severe DDIs. Validation occurred through expert interviews, discussing the correctness and utility of the approach. Results indicate that our graph-based model effectively supports the development of systems that alert caregivers to potential DDIs and recommend alternative drugs with reduced interactions.

To the best of our knowledge, this paper introduces the first graph-based model serving as a blueprint for developing DDI systems. This model enables systems to i) warn caregivers about the presence of DDIs in prescribed drugs and ii) assess the availability of alternative drugs with less severe interactions, providing recommendations.

In today’s competitive and multifaceted business ecosystem, ensuring the smooth operation of businesses requires choosing appropriate methods and techniques to address the complexity of modern information systems. Small and Medium Enterprises (SMEs) and public organizations often face unique challenges in this regard, such as resource limitations and complying with regulations. To succeed and provide better services in such environments, there is a critical need for research that offers effective strategies for managing these complexities. This issue features articles that provide valuable insights and tools for researchers and practitioners navigating these challenges.