Perspectives in Business Informatics Research

While artificial intelligence promises a wide range of potential for businesses, its adoption poses major problems for some organizations. This paper presents a modeling framework that aims to specify AI use cases. It models three views: Business, data and analytics, that are adopted for the requirements of AI. The framework was applied in a real-world case study leading to several AI use cases and two proof of concepts. While the business view is a useful tool to derive ideas for AI use cases in general, the data and analytics views are very specific to each use case. The framework serves as a means to an end to communicate the project goals, deliver practical guidance and to capture the main results. As its application is time consuming and challenging, this paper closes with guidelines for its efficient use in practice.

Advocating a convergence between generative AI and BPMN-based process analysis, this study reports on experiments with multi-modal business process representations. By leveraging the capabilities of the Bee-Up modeling tool for RDF serialization and the standard XML export of SAP Signavio, the report probes into the generative AI ability of BPMN interpretation according to these different serializations. In addition, the deployment of multi-modal AI – that directly processes image inputs – transcends traditional constraints of machine readability of BPMN diagrams. For prompt engineering, we employ a combined strategy utilizing semantic processing offered by Ontotext GraphDB integrated with LLM services from OpenAI, which, applied on RDF representations of BPMN, can push the boundaries of natural language interactions with visual process models. The investigation experiments with the interpretation of BPMN process models through such AI-based user interactions, highlighting possibilities of integrating conversational AI with the Business Process Management lifecycle. Assessments of outcomes are based on the RAGAs framework.

Large language models (LLM) have been successfully applied in enterprise modelling (EM) for various tasks, such as supporting modellers and domain experts in modelling the current situation in enterprises. An important factor for the successful application of LLM is the quality of the LLM output. This paper’s research investigates whether LLM can be used as a tool for the quality control of LLM output. Starting from an analysis of LLM evaluation approaches, the paper focuses on investigating scenarios for LLM use and relevant quality criteria. The main contributions of this paper are (1) an approach for using LLM as support for quality control of LLM output in enterprise modelling consisting of quality criteria for defined application scenarios and their operationalization and (2) quasi-experiments showing the applicability of the approach.

Amidst the digital transformation, traditional linear TV faces major challenges, including fragmented viewership, fixed schedule, and inaccurate targeting. Therefore, this paper proposes a novel Machine Learning framework to understand the audience’s demographics from their viewing behaviour. By employing state-of-the-art classification models on an extensive TV first-party dataset, we achieved an average 88.6% accuracy in correctly identifying each household demographics. Our result offers promising outcomes for refining strategies within linear TV to improve viewer engagement, content programming, and market insights.

Automatic text classification is an important area of study in natural language processing (NLP) and machine learning. Text classification has become essential for businesses and organizations to handle incoming documents effectively and efficiently. The main objective of this study is to introduce and evaluate a selection of Free Open Source Software approaches for document classification and compare them against each other regarding their prediction performance and efficiency to identify the best candidate for a specific use case. In addition, the study compares the selected approaches prediction performance, efficiency, and cost-effectiveness with commercial providers’ proprietary software. This comparison provides insights into different approaches’ relative strengths and weaknesses to help businesses decide on the best strategy for their needs.

In the continuously evolving landscape of healthcare technology, the evolution of robotics has revolutionized a wide spectrum of services. This situation created a variety of opportunities and challenges. One of the main challenges concerns the interaction of robots with human users and other robots, along with the safety aspect of such interactions. The current paper introduces a project aiming to tackle this challenge, initially via a literature review aiming to explore and structure the domain of robotic interactions in healthcare. Subsequently, the results are conceptualized in a domain meta-model, aiming to establish the foundation for an information system, based on a multi-view modeling approach, with the capabilities to support by documenting and facilitating safer interactions between healthcare robots and external agents. The current study focuses on the safety viewpoint of robotic interaction. The developed model is suitable for initiating the development of the abovementioned system that will use multi-view modeling principles. It is demonstrated in a use case derived from the associated ENDORSE EU project.

Organisational rules, both created internally and externally mandated, are vital to an enterprise. Yet, understanding and managing these rules is problematic, as they are a part of a complex system. Thus, there is a need to view them in a complex setting of organisational actors and interactions. It has been suggested that enterprises, particularly in situations like collaboration in healthcare, should be analysed as complex adaptive systems (CAS). However, only some enterprise modelling contributions can represent perspectives of CAS theory. In this paper, we set out to examine how organisational rules in complex adaptive systems has been modelled. A systematic mapping study was conducted on modelling languages of organisational rules in collaborations, resulting in 22 identified languages. The constructs and modelling patterns of the identified languages were mapped against an analytical framework that included 15 concepts from CAS theory. Overall, even though most CAS concepts had yet to be addressed by the identified languages, potentially useful approaches were found, related to: abstraction of large organisational rule systems through power relations; interpretation and implementation of rules; feedback loops to rule-makers, including delays.

Studies on the effects of digital transformation on enterprise architectures (EA) indicate that variability in the EA increases on different levels, such as the business and the data architecture. Dealing with variability becomes a common challenge in the daily operations of many enterprises and requires methodical and technological support. Methodical support for controlling variability in EA can help enterprises manage variability more efficiently. In this context, the conjecture motivating this paper is that building blocks integrating business and data architecture or allowing for data-aware business process building blocks can help to ensure a high level of flexibility and, at the same time, control complexity in variability management. The paper aims to contribute to a better understanding of the requirements, necessary activities and frame conditions of method support for variability management of EA. The paper’s main contribution is an initial method for identifying building blocks in enterprise architecture models that integrate several architecture layers and a way to capture such building blocks as ArchiMate models.

Business Intelligence systems have become one of the most widely used information systems yet their efficient and successful usage requires advanced digital skills of business users, implementation team and digital transformation of organization in general. These requirements are particularly profound in implementation of self-service business intelligence solutions. In order to evaluate factors influencing success or failure of implementation of self-service business intelligence solutions, this paper adopts a modified DeLone and McLean IS success model. The modified model emphasizes importance of organizational factors and correlates user intent and satisfaction with leading and lagging indicators characterizing the implementation. The model is applied to analyze more than a hundred business intelligence implementation projects in small and medium size companies. An in-depth analysis of one successful project and one failed project is performed. The model offers a tool for business intelligence project management teams to increase success rates, to get insights from ongoing projects, to understand past failures and to plan better work for future projects.

Ethical considerations in software requirements engineering are a critical but often overlooked aspect of the software development process. However, requests for transparency and autonomy in the way IT artefacts are designed, described, used, applied and integrated in the everyday life are getting more pressing within the society. In this research, the process of software engineering is taken as an illustrative model for the proactive incorporation of ethical principles in the system design within the design of a software artefact. Specifically, the phase of requirements elicitation and analysis are expanded with ethical aspects, since that is where the first steps of the software construction as an artefact are initiated and the first common ground of understanding is achieved. Being rooted in the security engineering, the SQUARE process is expanded to provide a basic structure for incorporating of ethical aspects into software design. By doing so, the social and moral values become central to the design and development of new technologies.

Accurate identification and documentation of requirements for software development is fundamental to successful project implementation. Business process models help to clearly define and document the steps involved in a process, providing an opportunity to use the models for deriving system requirements. Thus, selecting appropriate business process modeling methods or languages for requirements identification can significantly impact the quality of the final software product. The purpose of this paper is to examine several business process modeling languages with the purpose of showing and comparing their potential to help derive functional and non-functional software requirements out of business process models.

Software architecture is concerned with identifying the essential components of a software system and the process of creating an abstract representation of the system. A representation encompasses software elements, their properties and relationships with other elements. Similar to the architecture of a physical building, it serves as a blueprint for both the system itself and the development process, delineating the tasks to be carried out by the development team. In e-business, a wide variety of such architectures can be found. These have evolved over time, and the emergence and subsequent developments of the Internet have been decisive factors in shaping applications and their associated software architectures. In the 1980s, the purpose of an application was to run on a single computer, but today, applications are increasingly interconnected and accessible from anywhere. Practitioners have realized that a sound architecture is critical for success in both design and development. In order to incorporate domain-specific semantics, knowledge graphs must become a key ingredient to architectural thinking and the codification of principles, methods, and practices has led to repeatable architectural design processes. However, despite these advancements, the field of software architecture remains relatively immature in its relation to knowledge graphs and how to leverage them for e-business – towards this goal, traditional architectural patterns such as MVCs will be revisited in this paper.

Due to data and its use being an upcoming source of value for all industries, the use of IT systems becomes increasingly important to the daily business of most companies. As digitalization efforts increase, some existing obstacles come into focus – such as technical debt (TD). TD is well-researched in the software industry, but not so much in other industries. This paper aims at answering the question of how clients of software vendors in other industries are confronted with TD by performing a case study in a manufacturing SME and using grounded theory to develop a theory model on how TD occurs on the client-side, considering the entire system landscape and its evolution.

The paper introduces rules to help identify and depict in a model enterprise activities that engage tacit knowledge. This is done using a specific enterprise modeling technique called Fractal Enterprise Model (FEM). However, the result can be of interest to researchers and practitioners using other modeling techniques. Though representing tacit knowledge is more or less mandatory for research and practice of Knowledge Management (KM), it is very seldom depicted explicitly in enterprise models of any kind. The type of rules presented in this paper follows our suggestion to consider the so-called discovery power of enterprise modeling languages alongside its expressive power.

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.

Product information flows internally between stakeholders and their practices, externally between partners and across supply chains. Upcoming EU regulations mandate that all physical goods placed on the EU market or put into service must be linked to a digital product passport (DPP). A DPP provides digital information about a product’s entire lifecycle, materials utilised, environmental footprint, disposals, warranty and other data. This poses significant challenges for all organisations regarding understanding and designing their information systems and flows, feeding accurate public and restricted product information into their DPPs. The Work-Oriented Approach to Information Products (WOA) has been designed to address challenges with multi-stakeholders and multi-practice information flows involving information products, including DPP. This paper presents an industrial case study that aims to demonstrate the application of the WOA method and constructs in a knowledge development project by a global manufacturer of premium metal-cutting tools that aims to gain insights into future production and flow of product (lifecycle) information to their DPP.