Given the increasing amount of data available in organizational systems, there is an opportunity for early requirements engineering (RE) activities to be better based on evidence than ever before. Process mining (PM) has been used for over two decades to discover and analyze as-is process models from event logs extracted from such data, with outputs often in the form of Petri Nets, directly-follows graphs, or BPMN models. This paper aims to make Use Case Map (UCM) models, from ITU-T's User Requirements Notation (URN), a first-class output of process discovery, so that mined behavior can be used in URN-based modeling, analysis, and management activities. This paper contributes and illustrates PM4Py-UCM, an open-source extension to the existing PM4Py Python library. This new tool contributes 1) a UCM discovery pipeline, 2) hierarchical decomposition strategies producing nested UCM models, 3) configurable performer mappings for UCM and BPMN visualizations, and 4) an exporter to a URN tool (jUCMNav) that preserves the mined model under round-trip. Using public and synthetic event logs, the paper showcases how the same behavior is rendered under different performer abstractions and decomposition strategies, and discusses how PM can become a practical instrument for model-driven RE.

Care pathways constitute evidence-based “best-practices” to coordinate the activities and roles for diagnosing and treating illnesses. A patient’s actual series of clinical activities, their care trajectory, may differ from these pathways for multiple reasons—differing knowledge and experiences of clinicians, local practices, or multiple patient morbidities. Checking the compliance of care trajectories with clinical pathways not only tests their alignment with evidence-based practices, but may also identify opportunities for improving pathways; such as assessing their effectiveness in practice, and finding non-standardized yet valuable pathways for complex cases. To conduct compliance checking on a meaningful scale, we can use process (and decision) mining to extract care trajectories as process models from clinical data. In turn, when considering care pathways as “reference” process models, the latter can be extracted from guidelines using process extraction techniques. While challenges remain in mining care trajectories, a major barrier involves automated process extraction from clinical text. In this regard, novel Large Language Models (LLMs) frameworks offer promise. This paper outlines a methodology for checking pathway compliance using process mining and process extraction. To assess the feasibility of the latter, we present initial evaluation results using multiple state-of-the-art LLM architectures for Quality-Based Procedures from Ontario (Canada).

Enterprise-modeling (EM) tools are often complex and hard to extend. Yet, users may want to explore new EM features and capabilities that currently do not exist. AI coding agents can help here by enabling the development of new capabilities and entire tools, but whether we can trust a modeling-language tool an LLM largely wrote remains a question. This paper reports on the AI-assisted construction of PM4Py-UCM, an open-source tool that mines Use Case Map (UCM) models from event logs. PM4Py-UCM's capabilities include some expected from process mining tools (e.g., performance heat-maps and dashboards) and distinctive ones (e.g., mined executable scenarios/variants, and model decomposition). We mined the development record itself, composed of 18 agent sessions (374 human turns and 10,328 tool actions over 65 hours), 151 commits, 20 releases, and a test suite grown from 108 to 691 test functions, in order to characterize, in a single in-depth case, how the tool was built with an agent (Claude Code), complemented by an independent static assessment of the resulting code (coverage, complexity, maintainability, security, architecture). We contribute a reproducible, privacy-preserving toolkit and taxonomy that classify human turns and flag cross-cutting consistency work, agent corrections, and retracted requests. Up to version 0.7.4, fixes outnumber features 2.3:1, with 18

Within the field of Process Mining, process discovery is a data-driven method for discovering the control flow of processes—the how of a process. However, control flow alone does not explain why cases follow different branches at decision points. Decision Mining (DM) addresses this limitation by leveraging Machine Learning (ML) techniques to discover decisional criteria, also known as guards, from event logs. While two decades have passed since DM’s introduction, there is still no systematic and comprehensive overview of research in this field. This study presents a systematic literature review of DM, conducted across digital libraries using predefined inclusion and exclusion criteria, resulting in 30 primary research studies. We examine these articles from multiple perspectives, including investigated problems and proposed solutions, employed algorithms and available implementations, relevant application domains and datasets, and challenges that remain in the literature. Our findings identify four categories of DM research problems being investigated: problems related to event logs, types of decisions, evaluation practices, and model types. Across these categories, incomplete event logs—where key contextual data required to extract guards is missing—were the most prevalent problem. We further identified multiple challenges for the practical adoption of DM in process-aware information systems. The reviewed studies have mainly focused on decision trees, with limited uptake of more advanced ML algorithms or other types of techniques (e.g., local search) that may improve performance. We also found the literature fragmented, as studies rarely build upon earlier work or leverage advancements in the Process Mining field.

Context: Laws and regulations increasingly influence software design, development, and quality assurance in regulated domains; however, the technology-neutral formulation of legal provisions complicates the derivation of concrete specifications, requirements, and acceptance criteria needed to verify software compliance. Producing these artifacts manually is labour-intensive and error-prone. Recent advances in generative AI, particularly large language models (LLMs), offer the potential for automated assistance in deriving software engineering artifacts from legal texts. Objective: Following a quasi-experimental design, we present the first systematic human-subject evaluation of LLMs' ability to automatically derive Gherkin behavioral specifications from legal texts. Gherkin is a domain-specific language for specifying system behaviors through scenario-based descriptions written in the Given--When--Then format. Due to their structured and machine-readable nature, Gherkin specifications lend themselves more readily to automation within software-development processes. Methods: We recruited 10 participants to evaluate Gherkin specifications generated from food-safety regulations by two LLMs, Claude and Llama. Sixty specifications were generated. Each participant independently assessed 12 specifications across five quality criteria: relevance, clarity, completeness, singularity, and time savings. Each specification was evaluated by two participants, yielding 120 assessments with quantitative ratings and qualitative feedback. Results: Ratings were uniformly high (top-two categories): relevance 95%, clarity 100%, completeness 94.2%, singularity 93.4%, and time savings 91.7%. No statistically reliable differences were observed across participants or between LLMs. Qualitative feedback noted occasional omissions, hallucinations, and mixed intents; the first two, in particular, underscore the importance of human oversight, especially in safety-critical domains where non-compliance can have severe consequences. Conclusion: Our results suggest that, in the context of food safety, LLMs can assist in deriving Gherkin specifications from legal texts; however, observed omissions and hallucinations necessitate systematic human review.

Process discovery is a data-driven method for discovering the control flow of processes – the how of a process. Complementary to this, Decision Mining (DM) uses machine learning techniques to discover the decisional criteria used in the process – the why of a process. Such decisional criteria, also known as guards, reflect the reasons why cases are routed along branches at a decision point. We performed a literature review on existing DM approaches and their utilized algorithms, datasets, and tools. We provide an interim report of our review in this paper. Additionally, to make our review results more accessible to DM practitioners and researchers, we present a structured, machine-readable, queryable and interactive representation of our review results. The Systematic Literature Review Ontology on DM (SLRO-DM) allows capturing work in DM as a Knowledge Graph (KG), from multiple perspectives, including research problems, proposed solutions, related works, and decision types. The SLRO part of the ontology is extendable and reusable, so it may be adapted for creating literature KG in other domains. The goal of the DM-KG is to support (a) practitioners in the practical application of DM, and (b) researchers in their exploration of the research landscape and identification of gaps. We evaluated our ontology using competency questions that target both these stakeholders.

This research commentary explores how requirements engineering (RE) can contribute to achieving digital humanism in AI-enabled healthcare, ensuring that technological innovation is balanced with human values. Drawing on insights from a panel discussion at the REWBAH’25 (RE for Well-Being, Aging, and Health) workshop, we identify four key themes: the necessity of interdisciplinary collaboration, the nuances of cultural and individual differences, human-AI collaborative decision-making, and the challenge of balancing technological advancement with sustainability and emotional considerations. We argue that RE is well suited for addressing these issues because it brings together diverse stakeholder perspectives to define what systems should achieve and how they should be used. In this paper, we build on digital humanism principles, on RE-relevant frameworks for AI, and on the above themes to frame the role of RE in supporting human-centred healthcare. We highlight open challenges, and outline five promising research directions and related research questions to guide future research and practice in RE, aiming to ensure that AI-enabled healthcare systems better reflect the values of digital humanism.

This paper presents a manifesto that articulates the conceptual foundations of Agentic Business Process Management (APM), an extension of Business Process Management (BPM) for governing autonomous agents executing processes in organizations. From a management perspective, APM represents a paradigm shift from the traditional view on business processes. This shift is driven by the realization of process awareness by agent-oriented abstractions: software and human agents act as primary functional entities that perceive, reason, and act within explicit process frames. Thus, APM moves away from automation-oriented BPM towards systems in which autonomy is constrained, aligned, and made operational through process aware agents.We introduce the core abstractions and architectural elements required to realize APM systems and elaborate on four key capabilities that agents in APM systems must support: framed autonomy, explainability, conversational actionability, and self-modification. These capabilities jointly ensure that agents’ goals are aligned with organizational goals and that agents behave in a framed yet proactive manner in pursuing those goals. We discuss the extent to which the capabilities can be realized and identify research challenges whose resolution requires further advances in BPM, AI, and multi-agent systems. The manifesto thus serves as a roadmap for bridging these communities and for guiding the development of APM systems in practice.

Clinical Practice Guidelines (CPGs) are continually updated, yet translating such narrative recommendations into formal Care Pathway (CPW) processes remains labor-intensive and prone to inconsistency. We introduce a modular, end-to-end extraction pipeline that leverages Large Language Model (LLM) frameworks to automatically generate and evaluate BPMN models of CPWs from CPG text. Our system standardizes different output formats, applies LLM-based label alignment to harmonize terminology, and evaluates models against reference models using node- and structural-similarity metrics. Experiments on four stroke-related CPGs, using two state-of-the-art LLM frameworks—ProMoAI and a multi-agent orchestration approach (MAO)—shows the feasibility of automated CPW process extraction. We further observe that the multi-agent framework (MAO) demonstrates markedly higher fidelity. A web-based UI supports experiment configuration and result inspection. We released all code, prompts, and datasets as open-source to promote reproducibility and future enhancements.

Requirements Engineering (RE) is a critical sub-field of Software Engineering (SE), involving activities to identify, specify, model, analyze, and validate system needs and constraints. RE plays a fundamental role in the SE process. With the current advances in AI, new opportunities for RE arise to ensure the development of trustworthy AI and to utilize AI to perform various RE-related activities more effectively and efficiently. The workshop on Multi-disciplinary, Open, and IntegRatEd RE (MO2RE) successfully addresses the underrepresentation of the RE within the SE community, raises awareness of RE's diverse aspects, and fosters collaboration. The second edition took place on April 27th, 2025, co-located with the 47th IEEE/ACM International Conference on Software Engineering (ICSE) in Ottawa, Canada.

Eliciting requirements for Business Intelligence (BI) systems remains a significant challenge, particularly in changing business environments. This paper introduces a novel AI-driven system, called AutoBIR, that leverages semantic search and Large Language Models (LLMs) to automate and accelerate the specification of BI requirements. The system facilitates intuitive interaction with stakeholders through a conversational interface, translating user inputs into prototype analytic code, descriptions, and data dependencies. Additionally, AutoBIR produces detailed test-case reports, optionally enhanced with visual aids, streamlining the requirement elicitation process. By incorporating user feedback, the system refines BI reporting and system design, demonstrating practical applications for expediting data-driven decision-making. This paper explores the broader potential of generative AI in transforming BI development, illustrating its role in enhancing data engineering practice for large-scale, evolving systems.

Despite substantial investments from the Canadian government in constructing new and renovating existing affordable and social housing, communities continue to face unmet housing needs, and the gap between supply and demand for this type of housing is widening. Effective decision-making around social and affordable housing investments and policy requires comprehensive, system-level data. For such data to be useful, it must be pulled together from various sources to provide a meaningful picture of housing needs at local, regional, and national levels. This paper outlines our vision for an open data solution that integrates existing databases on social housing needs, shelter system users, and available housing stock to improve how decisions are made around housing priorities.

As Industry 4.0 transforms the food industry, the role of software in achieving compliance with food-safety regulations is becoming increasingly critical. Food-safety regulations, like those in many legal domains, have largely been articulated in a technology-independent manner to ensure their longevity and broad applicability. However, this approach leaves a gap between the regulations and the modern systems and software increasingly used to implement them. In this article, we pursue two main goals. First, we conduct a Grounded Theory study of food-safety regulations and develop a conceptual characterization of food-safety concepts that closely relate to systems and software requirements. Second, we examine the effectiveness of two families of large language models (LLMs) -- BERT and GPT -- in automatically classifying legal provisions based on requirements-related food-safety concepts. Our results show that: (a) when fine-tuned, the accuracy differences between the best-performing models in the BERT and GPT families are relatively small. Nevertheless, the most powerful model in our experiments, GPT-4o, still achieves the highest accuracy, with an average Precision of 89% and an average Recall of 87%; (b) few-shot learning with GPT-4o increases Recall to 97% but decreases Precision to 65%, suggesting a trade-off between fine-tuning and few-shot learning; (c) despite our training examples being drawn exclusively from Canadian regulations, LLM-based classification performs consistently well on test provisions from the US, indicating a degree of generalizability across regulatory jurisdictions; and (d) for our classification task, LLMs significantly outperform simpler baselines constructed using long short-term memory (LSTM) networks and automatic keyword extraction.

This paper investigates the use of Contract Specification Languages (CSLs), smart contracts, and their associated software methods and tools for formally specifying, verifying, and monitoring contracts, as a means to improve compliance with data protection laws, by harnessing the back end of data supply chains. The privacy literature has focused predominantly on B2C relationships (front end data processing) to assess corporate compliance with data protection laws and has explored the use of technologies primarily as a means to empower data subjects vis-à-vis data processors. This article shifts the perspective. First, it shows that the effective implementation of data subjects’ rights hinges upon the capacity of data processors to monitor compliance in back-end processing, i.e., the data supply chain. Second, it spells out the way CSLs and smart contracts can contribute to bridge the compliance gap in the data supply chain, which currently hinders data controllers from effectively monitoring the flow and use of data documents and practices across the networks of subcontractors. Indeed, while data is processed in an iterative and dynamic manner through a wide variety of means by actors within and across companies in the supply chain (e.g., controller, processors, and sub-processors), the legal instruments used to ensure compliance are dispersed, static, convoluted and legalistic, ultimately lacking the capability to operationalize data protection across different management levels (legal, technical, business, etc.). This paper proposes methods for specifying and controlling obligations related to data processing and distributions by using CSLs and smart contracts, and the contribution these methods can make for monitoring compliance in the data supply chain. It briefly introduces the Symboleo CSL for the formal specification of contracts. It then offers a concrete case study testing the properties of the Symboleo in a dataset composed of contractual documents used in the data supply chain of a network of IT companies. The paper concludes with a discussion of the challenges laying ahead to narrow the compliance gap and highlights how the proposed methods can evolve to become effective compliance tools in data protection.

Over the past decade, different domain-specific languages (DSLs) were proposed to formally specify requirements stated in legal contracts, mainly for analysis but also for code generation. Symboleo is a promising language in that area. However, writing formal specifications from natural-language contracts is a complex task, especially for legal experts who do not have formal language expertise. This paper reports on an exploratory experiment targeting the automated generation of Symboleo specifications from business contracts in English using Large Language Models (LLMs). Combinations (38) of prompt components are investigated (with/without the grammar, semantics explanations, 0 to 3 examples, and emotional prompts), mainly on GPT-4o but also to a lesser extent on 4 other LLMs. The generated specifications are manually assessed against 16 error types grouped into 3 severity levels. Early results on all LLMs show promising outcomes (even for a little-known DSL) that will likely accelerate the specification of legal contracts. However, several observed issues, especially around grammar/syntax adherence and environment variable identification (49%), suggest many areas where potential improvements should be investigated.

As Large Language Models (LLMs) become increasingly integral to software engineering tasks, the need for extensive evaluation benchmarks grows. Yet, creating these benchmarks manually is costly and time-consuming, posing a significant barrier to the effective testing and deployment of LLM-based systems. In this position paper, we highlight the challenges associated with manual benchmark creation using LLMs and discuss the potential of using LLMs themselves to automate this process. We explore critical issues such as ensuring dataset quality and comparability to human-crafted benchmarks, assisting users in validating generated examples, reducing the need for manual feedback, lowering associated costs, and facilitating customization across diverse tasks. By examining these challenges and sharing insights from our preliminary tool development, we aim to spark discussion and collaboration within the community to develop scalable solutions for benchmark generation.

Monitoring the compliance of contract performance against legal obligations is important in order to detect violations, ideally, as soon as they occur. Such monitoring can nowadays be achieved through the use of smart contracts, which provide protection against tampering as well as some level of automation in handling violations. However, there exists a large gap between natural language contracts and smart contract implementations. This paper introduces a Web-based environment that partly fills that gap by supporting the user-assisted refinement of Symboleo specifications corresponding to legal contract templates, followed by the automated generation of monitoring smart contracts deployable on the Hyperledger Fabric platform. This environment, illustrated using a sample contract from the transactive energy domain, shows much potential in accelerating the development of smart contracts in a legal compliance context.

We present a case study of Process Mining (PM) for personnel security screening in the Canadian government. We consider customer (process time) and organizational (cost) perspectives. Furthermore, in contrast to most published case studies, we assess the full process improvement lifecycle: pre-intervention analyses pointed out initial bottlenecks, and post-intervention analyses identified the intervention impact and remaining areas for improvement. Using PM techniques, we identified frequent exceptional scenarios (e.g., applications requiring amendment), time-intensive loops (e.g., employees forgetting tasks), and resource allocation issues (e.g., involvement of non-security personnel). Subsequent process improvement interventions, implemented using a flexible low-code digital platform, reduced security briefing times from around 7 days to 46 h, and overall process time from around 31 days to 26 days, on average. From a cost perspective, the involvement of hiring managers and security screening officers was significantly reduced. These results demonstrate how PM can become part of a broader digital transformation framework to improve public service delivery.

Smart contracts act as immutable and trusted intermediaries, ensuring transactions comply with contractual terms. As translating natural-language legal contracts into executable smart contracts is time-consuming and error-prone, formal specification languages such as Symboleo were introduced as intermediate representations that enable verification and code generation. Yet, writing or generating such specifications correctly still requires time and specialized expertise. To reduce manual effort in this context, this paper investigates a new repair approach that combines Large Language Models with compiler feedback. We propose an automated agent (SymboleoFix) that iteratively repairs Symboleo specifications, correcting many syntactic, semantic, and type-related errors. Preliminary experimental results from a case study suggest that SymboleoFix can help most eliminate syntactic errors, reduce the need for manual intervention, and improve specification accuracy.