About: This study investigates how large organizations evaluate and manage trust when procuring third-party software. Drawing from literature reviews and nine semi-structured interviews with professionals, the research identifies key trust factors and maps how they are applied throughout the procurement lifecycle. The goal is to provide practical insights for improving procurement strategies and for software vendors seeking to build trust.
Problem: As businesses increasingly depend on external software for core operations, the risk of supply chain attacks and data breaches has grown significantly. While trust in a software vendor is critical, there is limited understanding of how large organizations practically assess and maintain this trust beyond formal checklists. This study addresses the gap between prescribed procurement models and the real-world, dynamic processes companies use to manage vendor risk.
Outcome: - Organizations employ a layered trust strategy, combining formal mechanisms like ISO/SOC certifications and legal contracts with relational ones such as vendor transparency and past performance. - Trust is not a one-time check but an ongoing process that is monitored and re-evaluated throughout the vendor relationship, especially at contract renewal or after security incidents. - Procurement processes are often flexible, adapting based on the software's risk level, cost, and the data it handles; high-risk tools undergo stricter scrutiny. - Legal and geopolitical factors, particularly concerning data residency and regulations like GDPR and the U.S. CLOUD Act, are major considerations in vendor selection. - The study found that organizations balance formal requirements with practical needs, sometimes compromising on certain criteria when no ideal alternative vendor exists.

About: This study investigates how generative AI can enhance software security practices through a multi-case study involving five different software organizations. Using semi-structured interviews with developers and managers, the research identifies key areas where AI can provide support, while also uncovering the sociotechnical risks that may hinder its adoption.
Problem: Software organizations face a rising number of security threats, and manual security processes are often time-consuming, error-prone, and expensive. While generative AI offers a promising solution to enhance security, its practical impact and the associated risks for software organizations remain largely underexplored.
Outcome: - Generative AI can enhance four key security practices: threat assessment, security testing, operational management, and education & guidance. - Significant sociotechnical risks impede AI adoption, including concerns about inadequate data management, inaccurate AI outputs, and a lack of trust and transparency. - Key identified risks include insufficient data management, data poisoning, lack of control, and limited data novelty. - The study proposes a theoretical model that highlights the balance between the perceived benefits and risks of adopting generative AI in security workflows. - Practitioners must balance AI-driven efficiency with effective risk management, as human validation of AI outputs remains necessary.

About: This study investigates the digital and workflow-related challenges that small and medium-sized enterprises (SMEs) face when collecting and reporting emissions data for Corporate Sustainability Reporting Directive (CSRD) compliance. Using a multi-method approach combining document analysis of regulatory standards and semi-structured interviews with SME professionals, the paper identifies key barriers and proposes features for digital reporting systems to address these issues.
Problem: New European Union regulations (CSRD) mandate extensive sustainability and emissions reporting, creating significant challenges for SMEs, which often lack the technical infrastructure, expertise, and resources to comply. There is a research gap concerning how these smaller companies can adopt or struggle with digital solutions to meet these demanding new requirements for collecting, measuring, and reporting emissions data accurately.
Outcome: - SMEs predominantly rely on manual, fragmented, and error-prone data collection methods, such as Excel spreadsheets and gathering information from emails and invoices. - A major hurdle is the complexity of calculating Scope 3 emissions (from the value chain), forcing SMEs to use imprecise estimation methods due to a lack of supplier data and specialized tools. - SMEs experience 'compliance overload' and declining motivation, as the reporting burden often shifts to finance departments, taking time away from implementing actual sustainability projects. - The study recommends that digital reporting systems should feature automated data integration from various sources, real-time data validation to ensure quality, and flexible tools for estimating Scope 3 emissions. - To ease the reporting burden and ensure compliance, recommended systems should also provide structured, machine-readable report templates and role-based access controls to facilitate collaboration across departments.

About: This paper presents the design, implementation, and evaluation of a serious game scenario titled 'ERP-Systems in Need' to raise cybersecurity awareness specifically for Enterprise Resource Planning (ERP) systems. The study applies a Design Science Research approach to build upon an existing game, simulating realistic security incidents to foster experiential learning. The game's effectiveness was evaluated in academic settings to measure its impact on increasing knowledge and preparedness for ERP-related cyber threats.
Problem: Enterprise Resource Planning (ERP) systems are fundamental to business operations but are increasingly vulnerable to sophisticated cyberattacks that can paralyze supply chains and cause significant financial damage. Traditional cybersecurity training methods like lectures are often ineffective at engaging users, and there is a lack of gamified training tools specifically addressing the unique security risks of ERP systems.
Outcome: - The serious game 'ERP-Systems in Need' is an effective training tool for increasing knowledge, participation, and preparedness for ERP-related cyber threats. - A dual-role setup (employee and management) proved valuable for promoting perspective-taking and fostering discussion across organizational levels. - Both technical and non-technical participants showed increased confidence in identifying and responding to incidents after the game sessions. - Evaluation results indicate high participant satisfaction and a significant increase in perceived learning content compared to games focused on general IT security.

About: This qualitative study investigates how organizations decide to adopt Artificial Intelligence (AI) technologies. Through 15 semi-structured interviews across various industries, the research applies Gioia's method to identify the key motivations, triggers, and decision-making factors influencing AI adoption. The study proposes a process model to explain this complex decision-making journey.
Problem: Many organizations struggle with the critical decision of when and how to adopt AI, often rushing into implementation without strategic clarity, which leads to disappointing results. Existing research models often overlook the complex, contextual processes through which organizations deliberate and evaluate AI, creating a gap in understanding the practical decision-making journey.
Outcome: - AI adoption is not a singular, technical decision but a complex, multi-stage reasoning process involving feedback loops. - The decision is shaped by three interdependent dimensions: strategic motivations (the 'why'), preconditions and triggers (the 'when'), and evaluation criteria (the 'how'). - Systemic challenges, such as poor data quality, act not only as barriers but also as catalysts that prompt organizations to explore AI solutions. - Organizations evaluate AI adoption based on a layered process that includes strategic alignment and organizational readiness, not just financial ROI or technical feasibility. - The study produced a process model that illustrates how organizations move through various phases of AI adoption decision-making, including stages for re-evaluation and reconsideration.

About: This study investigates the challenges and opportunities of integrating Artificial Intelligence (AI) and Machine Learning (ML) in regulated manufacturing sectors like pharmaceuticals and medical devices. Through 20 qualitative interviews with data architects, AI specialists, and compliance officers, the research identifies key barriers and proposes a conceptual framework for designing compliant, AI-driven data architectures.
Problem: Life science manufacturers are increasingly adopting AI/ML to improve production and compliance, but progress is slow. They face significant hurdles from fragmented data systems, legacy infrastructures, and data silos, which are compounded by complex and evolving regulatory requirements that create uncertainty around system validation and traceability.
Outcome: - Current data infrastructures in regulated manufacturing are highly fragmented, with data silos and legacy systems acting as the primary technical barriers to AI adoption. - Evolving regulatory frameworks and uncertainty in how to validate AI systems create significant compliance challenges that slow down innovation. - Experts agree that overcoming these barriers requires unified data architectures, embedded governance, enhanced security, and proactive regulatory operations (RegOps). - The study proposes a conceptual framework for a validation-aware, layered data architecture that can bridge fragmented systems and support the entire AI/ML lifecycle in a compliant manner.

About: This paper presents the initial findings from a qualitative interview study conducted with two Finnish business-to-business (B2B) software companies. The research goal was to understand how digital identities are perceived and managed from the perspective of identity providers and software product companies. The study identifies key stakeholder priorities, as well as the factors that disrupt or accelerate the process of meeting those priorities.
Problem: Software companies that provide software-as-a-service products must manage the growing complexity of digital identities, balancing numerous stakeholder needs and evolving regulations. There is a research gap in understanding how these companies practically define and handle digital identities, navigate competing priorities like security and user experience, and address challenges in a real-world business context.
Outcome: - Software companies tend to avoid the abstract term 'digital identity', preferring more concrete, practical terms like 'authentication', 'login', and 'user roles'. - Different stakeholders have distinct priorities: customers value control and user experience; identity providers focus on interoperability and security; software companies aim to balance security with competitive advantage; and regulators emphasize compliance. - Key factors that slow down (disruptors) the alignment of these priorities include security concerns, interoperability issues, unknown regulatory requirements, and a lack of standardization. - Factors that speed up (accelerators) this alignment include a strong emphasis on user experience, clear company strategies for security and identity management, and collaboration among stakeholders.

About: This study introduces and conceptualizes the 'Cost of Not Investing' (CONI) in intelligent process automation (IPA). The authors develop a framework identifying three dimensions of CONI—operational inefficiencies, strategic disadvantages, and human capital impacts—and test initial perceptions through a cross-sector survey of 108 professionals in the US and EU.
Problem: Traditional methods for evaluating technology investments fail to capture the significant strategic risks associated with delaying the adoption of AI-driven automation. This oversight can lead businesses to make suboptimal decisions by underestimating the cascading negative consequences of inaction, such as losing competitive advantages and key talent.
Outcome: - Delaying automation has severe consequences, with professionals identifying talent migration (81% agreement), lower productivity (77%), and loss of early-mover advantages (74%) as major risks. - A critical gap exists between awareness and action, as only 16.7% of organizations systematically incorporate these strategic costs into their investment decisions. - The cost of delay is a cascading mechanism, where initial operational inefficiencies lead to lost strategic flexibility, which in turn erodes capabilities and weakens long-term competitive position. - The most significant perceived impact is on human capital, with 81% of respondents believing that valuable employees will leave for competitors with more advanced automation.

About: This study investigates how environmental sustainability is implemented in software-focused companies and identifies the factors that motivate or obstruct these green practices. The findings are based on a thematic analysis of 22 semi-structured interviews with 38 professionals from various Finnish organizations involved in Green ICT projects.
Problem: The software industry has a dual role in sustainability; it can enable efficiency and resource optimization, but it also contributes to a significant environmental footprint through energy consumption and hardware dependencies. Despite increased awareness, a persistent gap exists between research on sustainable practices and their actual implementation in business operations.
Outcome: - The primary barriers to implementing sustainable practices are low prioritization compared to cost or quality, unclear roles and responsibilities, limited client influence, and a lack of specialized expertise. - Key benefits are primarily business-driven, including improved competitiveness, enhanced company reputation, cost savings, and eligibility for public contracts requiring sustainability criteria. - Sustainable practices appear in four main areas: the organization (strategy, carbon footprint), the software process (service design), the software product (efficient code, simple architecture), and external factors (cloud infrastructure). - The study highlights a crucial need for a shared understanding of sustainability within organizations, integrated competencies, and measurable metrics to drive meaningful change.

About: This study investigates how software development organizations in Finland and globally are practically adopting green coding practices. It uses a qualitative research method, combining a systematic literature review with semi-structured interviews with 15 software professionals to compare academic recommendations with real-world industry applications.
Problem: The Information and Communication Technology (ICT) sector has a significant and growing environmental footprint, contributing to global greenhouse gas emissions. While green coding offers a way to create more energy-efficient software, there is a notable gap between academic proposals and their practical adoption in the industry, leaving developers without standardized guidelines or integrated tools.
Outcome: - Organizations widely adopt practices that improve both performance and energy efficiency, such as caching, lazy loading, and database query optimization. - The explicit focus on 'green coding' is more evident in European companies, influenced by regulations, while in non-EU regions it is often a secondary benefit of writing high-quality code. - Green coding initiatives are typically driven from the bottom up by software engineers rather than through formal, top-down management mandates. - There is a significant disconnect between innovative academic tools (e.g., LLM-based assistants) and their lack of adoption in the industry. - The study highlights a need for standardized guidelines, better industry-academia collaboration to create practical tools, and integrating sustainability into computer science education.

About: This study investigates how playful interaction design can mitigate creative burnout in software development teams' Sprint Retrospectives. Using a Design Science Research approach, the researchers developed and evaluated a design artifact called 'Scrumstinct' in workshops with eight professional teams to understand the impact of playfulness on team well-being and creativity.
Problem: Creative burnout is a significant issue in software development, leading to stagnation, exhaustion, and rigid routines that harm team creativity and well-being. While acknowledged in the industry, this phenomenon is academically underexplored, particularly at the team level, creating a need for research-backed strategies to foster sustainable creativity.
Outcome: - Playful retrospectives can successfully reframe problems, foster openness, and reduce emotional fatigue among software development teams. - The study proposes six design principles for integrating playfulness to address team-level creative burnout by countering stagnation, exhaustion, and rigidity. - Introducing fun, excitement, and desire through a designed artifact ('Scrumstinct') increased team engagement and made retrospectives more desirable. - Using humorous and metaphorical artifacts helps teams discuss serious concerns in a constructive, less confrontational manner. - Playful activities that deviate from routine were found to challenge rigid thinking and open up teams to new perspectives and solutions.

About: This study aims to create clear guidelines for businesses looking to implement IT-supported circular economy (CE) projects. Using a mixed-methods approach that included a literature review, text analysis, and a workshop with industry practitioners, the researchers developed a set of core design principles.
Problem: While the circular economy is a popular concept, its broad and often vague definitions make it difficult for companies to implement practical, IT-driven initiatives. This lack of conceptual clarity hinders organizations from identifying effective strategies, preventing them from fully realizing the potential of IT to advance sustainability goals.
Outcome: - DP1: Regulatory and Goal Alignment for IT-Driven CE Initiatives: This principle ensures that projects comply with regulations and balance economic goals with clear ecological and social objectives. - DP2: Digital Foundations for a Holistic CE Management: This principle focuses on leveraging existing technology, systematically evaluating circular business models, and mapping stakeholders and data to close information gaps across the value chain. - DP3: Implementation of CE through IT-Enabled Processes: This principle guides companies to adopt comprehensive strategies covering the entire product lifecycle and to develop concrete implementation plans that engage all stakeholders.

About: This study aims to clarify the defining features of data spaces within the Industry 4.0 context by developing a comprehensive classification system, or taxonomy. The researchers analyzed 19 existing data spaces to identify their common dimensions and characteristics, providing a structured way to understand and compare them.
Problem: As companies increasingly rely on data sharing for innovation in Industry 4.0, 'data spaces' have emerged as a key solution for secure collaboration. However, the concept is poorly defined, and the existing research is fragmented, making it difficult for businesses, especially small and medium-sized enterprises, to understand and engage with these new data ecosystems.
Outcome: - The study developed a taxonomy to classify data spaces, identifying nine key dimensions (like funding, key purpose, and underlying technology) and 40 distinct characteristics. - Most Industry 4.0 data spaces are publicly funded, primarily aim to deliver economic benefits, and are based on the IDS Reference Architecture Model. - Access to these data spaces is typically granted through project membership or participation, rather than direct fees; no pricing models were found. - The documentation for many data spaces often lacks clear details on whether they are fully operational, what specific types of data are shared, and what technical enablers (like AI or blockchain) are used.

About: This study introduces Reqwire, an AI-driven, multi-agent system designed to integrate sustainability into agile software development. The system automatically generates user stories from requirement documents, enriches them with sustainability attributes, and integrates with project management tools like Jira. The research follows a Design Science Research (DSR) approach, using iterative cycles of design and evaluation with feedback from industry and academic partners.
Problem: During the requirements engineering phase of software projects, sustainability is often not a primary consideration. As a result, software engineers typically lack the tools to assess the environmental, social, and economic impacts of their products, leading to a gap between development practices and sustainable outcomes. This study addresses the lack of systematic tools for incorporating sustainability directly into early-stage agile workflows.
Outcome: - Reqwire reduces manual effort by automatically generating structured user stories, estimating story points, and assigning sustainability tags. - The system successfully categorizes sustainability impacts across five dimensions: environmental, economic, social, individual, and technical. - Its multi-agent framework enables integration with third-party tools like Jira, supporting consistent and systematic project tracking. - Initial tests indicate that Reqwire shows promise for enhancing agile workflows and promoting sustainable software development practices.

About: This study evaluates how effectively five different Generative AI (GenAI) tools can complete exercises and quizzes in five university programming courses. The research objective was to measure and compare the performance of these AIs across different courses to understand their capabilities in an academic setting. The methodology involved using the GenAI tools to solve all weekly exercises, mimicking a student who relies solely on AI to pass.
Problem: The rise of powerful Generative AI presents a significant challenge to university education, particularly in programming, as students can potentially use these tools to complete assignments without genuine learning. This creates a gap in verifying student knowledge and skills, forcing educators to reconsider how courses are designed and assessed. The study addresses the urgent need to understand the exact capabilities of these AIs to adapt educational practices and maintain academic integrity.
Outcome: - Generative AI tools were able to solve 43-90% of programming assignments and 80-100% of quizzes across the tested courses. - The results indicate that a student could pass the exercise portions of most of these courses by solely relying on readily available GenAI tools. - GitHub Copilot consistently performed the best, while the free version of ChatGPT (GPT-3.5) and Codeium were often the worst performers. - AIs struggled most with complex tasks, exercises that built upon previous solutions, and tasks requiring interaction with external files or APIs. - The study concludes that programming courses must incorporate assessments in controlled environments (e.g., monitored exams) to accurately verify student learning and skills.

About: This study investigates how companies in the software-intensive systems industry manage and utilize the vast amounts of data collected from their products. Through a multi-case study of eight companies, the research identifies beneficial practices, or 'patterns', that lead to successful data use, and detrimental practices, or 'anti-patterns', that companies should avoid.
Problem: Companies are collecting ever-increasing volumes of data from their products but often struggle to effectively transform this raw data into customer and business value. There is a lack of practical guidance on what constitutes best practices, leading to significant challenges in data management, quality assurance, access control, and cost management.
Outcome: - Successful companies follow beneficial patterns such as: exploring new data-driven services at low cost on internal infrastructure, operating monetized services with high-quality data on reliable public clouds, and collaborating with partners to generate new insights. - Detrimental 'anti-patterns' that hinder success include: using raw, low-quality data for commercial services, allowing ungoverned data access across teams which leads to system failures, and incurring excessive costs from using public clouds for exploratory data analysis. - Key operational challenges across companies are poor data quality, which consumes significant time for cleaning (up to 80% of effort in some cases), and the high costs associated with data storage and computation, especially in the cloud.

About: This study explores how ethical considerations can be effectively integrated into the digital strategy development process within organizations. Through a qualitative case study at a German public university, researchers interviewed sixteen employees across various hierarchical levels to understand how ethics are perceived and can contribute to a more effective strategy by balancing power and participation.
Problem: As organizations undergo digital transformation, ethical concerns are often treated as secondary or are not effectively embedded in the strategy development process itself. This creates a gap where decisions about technology and resource allocation are viewed as purely technical, overlooking their inherent ethical dimensions and the central tension between centralized power and broad employee participation.
Outcome: - Ethics are often perceived as an individual's personal concern rather than a formal, integral part of the digital strategy process. - The core challenge in digital strategy development is the ethical tension of 'balancing power and participation,' particularly concerning resource allocation and decision-making authority. - Seemingly technical or economic decisions, such as IT project prioritization, are inherently ethical because they involve the distribution of scarce resources and impact stakeholders differently. - In the absence of a clear central digital strategy, employees and departments tend to act independently, pursuing their own priorities, which can lead to fragmented and inefficient outcomes. - The study proposes a procedural, discourse-ethical approach using participatory and agile methods to negotiate interests and fairly allocate resources, making ethics a structural part of strategy development.

About: This paper explores how 'Living Labs' can serve as effective frameworks for developing and testing artificial intelligence (AI) innovations in the mobility sector. Using the AIAMO (Artificial Intelligence And Mobility) research project as a case study, it demonstrates how these collaborative, real-world environments facilitate the creation of practical, user-centered AI applications for traffic management.
Problem: Developing and implementing complex AI technologies for real-world scenarios, like smart mobility, faces significant hurdles. There is a need for open, practical testing environments that involve diverse stakeholders to bridge the gap between lab-based research and market-ready solutions, ensuring innovations are both effective and accepted by users.
Outcome: - Infrastructure Requirements: Integrating AI systems effectively requires a dedicated, accessible, and interoperable infrastructure, as demonstrated by the AIAMOnexus platform developed for the project. - Data Protection and Governance: Handling sensitive mobility data necessitates a comprehensive, multi-perspective approach to data protection and ethical compliance to build user trust. - User Involvement: Iterative co-development with stakeholders and end-users is crucial for aligning AI systems with real needs and regulatory requirements, leading to more successful and accepted innovations. - Transferability and Scalability: The modular architecture of the system supports expansion, but actual scalability depends heavily on factors like data availability, stakeholder engagement, and institutional support. - Practical Success: The study concludes that cooperation within a Living Lab leads to practical AI solutions with a high level of user acceptance, increasing the likelihood of commercial success.

About: This paper presents a conceptual framework for integrating Generative AI (GenAI) into higher education and corporate communication. Drawing on practice-based teaching experiences, it argues that the transformation driven by GenAI requires a threefold competence profile encompassing technical proficiency, narrative design, and ethical reflexivity. The study uses a conceptual approach based on media ethics, AI governance, and narrative theory to bridge the gap between academic theory and real-world business applications.
Problem: The rapid adoption of Generative AI is reshaping fundamental concepts like authorship, narration, and authenticity in media and academia. This presents a significant challenge for educational institutions and businesses, which must move beyond simple technical adoption to address the profound ethical, intellectual, and practical implications. The paper addresses the lack of a structured framework for developing the necessary skills and ethical awareness to navigate this new landscape effectively.
Outcome: - The transformation by GenAI demands a threefold competence profile: technical proficiency, narrative design capability, and ethical reflexivity. - In business, the focus is shifting from direct content production to strategic 'narrative orchestration,' where professionals design and manage coherent narrative worlds using AI tools. - Three core competencies are essential for the AI-augmented era: 'Worldbuilding' (creating immersive brand stories), 'Prompt Literacy' (skillfully guiding AI), and 'Ethical Framing' (ensuring responsible AI use through governance and stakeholder analysis). - University curricula should be redesigned to include four developmental dimensions: understanding how GenAI works, practical application, critical reflection on AI outputs, and co-creation with AI tools. - Key teaching topics must include the ethical and legal aspects of AI, such as sustainability, data privacy, copyright, AI content labelling, and liability.

About: This study investigates the sustainability perceptions and behaviors of consumers regarding digital technology products and services. Through a global survey of 490 technology end-users, the research identifies their key concerns, unmet expectations, and views on the current sustainability efforts of technology companies.
Problem: While there is extensive research on technology and sustainability, most of it focuses on the digital technologies themselves or the companies behind them, leaving a gap in understanding the end-user's perspective. This study addresses the lack of insight into consumer sustainability perceptions, behaviors, and expectations, which is crucial for promoting sustainable consumption in the digital economy.
Outcome: - Consumers' most pressing unmet sustainability needs are product durability, transparency, corporate legitimacy, ethical practices, the right to repair, and reducing water and energy consumption. - A significant communication gap exists between technology companies and their customers regarding sustainability efforts, leading to consumer skepticism and accusations of 'greenwashing'. - While over 80% of users rate sustainability as very or extremely important, they perceive tech companies' performance as only moderate, with most ratings clustering between 6 and 8 on a 10-point scale. - The primary barrier preventing consumers from adopting sustainable technology is the high cost of quality products, cited by 73.3% of respondents. - Consumers prioritize product longevity and repairability over downstream activities like recycling when it comes to sustainable digital consumption.

About: This study explores cooperative software businesses as an alternative model to address social-environmental crises and corporate short-termism. Researchers analyzed a dataset of 76 cooperative software firms and conducted interviews with members from 23 of them. The goal was to understand member motivations, democratic management practices, and the potential impact of this business model on the software sector.
Problem: Traditional corporate structures often focus on short-term financial gains, which can lead to negative social and environmental consequences. While cooperatives present a democratically managed alternative, software worker cooperatives have remained an under-explored area of research. This study addresses this gap by investigating their characteristics and potential for positive transformation in the tech industry.
Outcome: - Members of cooperative software businesses are motivated by both practical benefits, like greater involvement and work-life balance, and ideological alignment with Social Economy values. - These businesses actively promote workplace democracy, stakeholder engagement, and long-term value creation. - Three distinct profiles of members were identified: ideologically-driven, autonomy-seeking, and newly-aware members who join without prior knowledge of the cooperative model. - While these firms care about their societal impact, the formal measurement of this impact remains a limited and underdeveloped practice. - With broader support, cooperatives have the potential to play a transformative role in making the software sector more inclusive, participatory, and socially responsible.

About: This study investigates the potential of an AI-based decision support system for prioritizing product features within software startups. Using a design science research methodology, the authors developed a proof of concept (PoC) tool that uses AI to score features based on criteria like ROI and time-to-value. The PoC was then demonstrated to and evaluated by seven software entrepreneurs through semi-structured interviews to gauge its perceived usefulness and potential for adoption.
Problem: Software startups operate in fast-paced, uncertain environments with limited resources, making the prioritization of new product features a critical and constant challenge. Decisions are often made based on intuition or incomplete information, creating a need for more structured, data-driven approaches. This research addresses the gap by exploring how AI can enhance the feature prioritization process, which is currently often ad-hoc and subjective in startups.
Outcome: - Startups perceive AI as a useful aid for feature prioritization but not as a replacement for human decision-making and judgment. - The adoption of AI prioritization tools is contingent on improving their transparency, explainability, interoperability with existing workflows, and overall usability. - The conversational nature of the AI tool was found to be effective in forcing users to deliberately reflect on the value and metrics of each feature. - Trust in the AI's output is directly linked to understanding its reasoning; users need to be able to justify the AI's suggestions to other stakeholders. - Visual representation of the prioritization scores was a key strength, helping teams make faster, more objective decisions.

About: This study examines how two types of AI, Analytical and Generative, can support decision-making in the software release process. Based on qualitative interviews with seven industry experts, the paper identifies the key benefits, hurdles, and practical applications of integrating AI into modern, fast-paced development environments.
Problem: In modern software development with rapid CI/CD cycles, release management has become increasingly complex, compressing decision windows and elevating risk. While Artificial Intelligence presents new tools to manage this complexity, significant practical challenges such as integration complexity, model opacity, and organizational resistance hinder its effective adoption.
Outcome: - AI integration offers key benefits including increased automation of repetitive tasks, more efficient resource allocation, enhanced risk forecasting, and improved overall decision-making. - A clear division of labor was identified: Analytical AI is best for quantitative tasks like forecasting incidents and detecting anomalies, while Generative AI excels at language and coordination tasks such as summarizing changes and drafting release notes. - The primary challenges to AI adoption are high integration complexity with existing systems, lack of transparency or trust in 'black-box' AI models, and human/organizational factors like skill gaps and resistance to change. - For successful adoption, practitioners should prioritize reliable data pipelines, pair automated decisions with clear explanations, maintain human oversight for critical steps, and pilot AI solutions before scaling.

About: This paper presents a case study on implementing Enterprise Architecture Management (EAM) within a company shifting from hardware to software and services. The researchers developed a practical evaluation tool, a "fitting matrix," by conducting expert workshops and interviews to help the company choose the most suitable EAM strategy. This matrix systematically compares various EAM implementation models based on key success criteria.
Problem: Companies today face rapid changes and growing complexity, making traditional management methods ineffective. While Enterprise Architecture Management (EAM) is a promising solution for creating a holistic organizational view, there's a lack of clear, practical guidance on how to select and implement the right EAM approach for a company's specific context.
Outcome: - Developed a practical assessment tool called a "fitting matrix" to help organizations evaluate and select the best EAM implementation strategy. - The study identified that a federated "Governance EAM" model, led by the Chief Information Officer (CIO), was the most suitable option for the specific company studied. - The research confirms that there is no one-size-fits-all approach to EAM, highlighting the need for context-specific evaluation. - Centralized EAM approaches were seen as potentially creating an "ivory tower syndrome," detached from business needs, whereas a federated model promotes better stakeholder integration.

About: This study critically examines how online gambling platforms use specific User Experience (UX) design strategies to foster compulsive engagement and maximize user retention. Through a heuristic evaluation of five popular platforms, the research identifies and analyzes common persuasive and manipulative design patterns that exploit users' cognitive vulnerabilities.
Problem: The rise of online gambling has created a high-risk digital environment where sophisticated design techniques are used to encourage continuous play, often leading to compulsive behavior and addiction. There is a significant lack of transparency and ethical consideration in how these platforms are designed, exploiting user vulnerabilities for profit without implementing necessary protective measures.
Outcome: - A consistent and standardized set of UX design strategies focused on user retention was found across all analyzed gambling platforms. - All ten addictive design heuristics, including immediate sensory reinforcement, variable rewards, and artificial scarcity, were present in 100% of the platforms. - Platforms were found to strategically manipulate social proof by constantly displaying messages of other users' alleged winnings to distort the perception of luck. - A complete absence of user protection mechanisms, such as warnings about risky behavior, loss limits, or mandatory breaks, was a key finding. - The study concludes that design principles intended for usability are being transformed into mechanisms of manipulation that prioritize engagement over user well-being, highlighting a significant ethical risk.

About: This study introduces the Ethical Requirements Stack (ERS), a structured framework designed to translate high-level ethical principles into concrete development tasks for AI systems. Using a design science research methodology, the paper demonstrates a collaborative workflow where humans and GPT-based Large Language Models (LLMs) work together to elicit, decompose, and manage these ethical requirements in an Agile-inspired environment.
Problem: Software development teams often struggle to convert abstract ethical principles like fairness and transparency into specific, actionable requirements. This leads to an accumulation of "ethical debt," where unresolved ethical issues create risks of reputational harm, regulatory penalties, and loss of stakeholder trust.
Outcome: - The study introduces the Ethical Requirements Stack (ERS), a practical framework for breaking down broad ethical themes into actionable Epics, Features, and Stories. - Human-AI collaboration was shown to be highly effective; AI generated a high volume of potential ethical requirements, while human experts provided crucial contextual nuance, critical judgment, and strategic direction. - AI tended to generate more granular, story-level requirements, whereas humans focused on higher-level strategic themes, demonstrating complementary strengths. - The iterative process of refining AI-generated outputs with human oversight significantly improved the clarity, relevance, and actionability of the ethical requirements. - The ERS framework provides a dynamic method for managing and reducing ethical debt by integrating ethics into the software development lifecycle from the start.

About: This study evaluates the effectiveness of Active Personas (APs), which are user archetypes powered by Generative AI, in producing realistic user feedback for new product development. Using a Design Science Research approach, the researchers created APs based on different personas and Large Language Models (LLMs) to provide feedback on a mobile transport app. The AI-generated feedback was then compared against feedback from human users and Google Play reviews to assess its alignment and validity.
Problem: Securing consistent and diverse user feedback is a critical part of product development, but it is often a resource-intensive and time-consuming process. Development teams struggle to get rapid feedback from a wide range of user types, which can slow down innovation. This study explores whether AI-powered personas can serve as a low-cost, on-demand alternative to generate valuable user feedback, bridging the gap in internal product experimentation.
Outcome: - A strong alignment was found between the feedback generated by AI Active Personas (APs) and that from human users, with APs successfully identifying similar usability and accessibility issues. - The specific characteristics defined in a persona (e.g., a user with a vision impairment) significantly dictated the criticality and focus of the AI-generated evaluation. - The choice of the underlying Large Language Model (LLM) also influenced the evaluative stance and tone of the feedback. - APs are a viable, rapid, and low-cost method to supplement early-stage usability evaluations, but they should augment, not replace, direct interaction with human users.

About: This study presents a software-supported approach designed to help businesses configure and manage complex, service-oriented value networks. The core of this approach is a detailed metamodel, implemented in a software tool, which integrates all relevant elements like partners, resources, and activities into a single data model. This allows companies to model, analyze, simulate, and optimize their collaborative service offerings.
Problem: As industries shift from product-centric to service-oriented models, businesses increasingly operate within complex networks of partners, suppliers, and customers. Traditional methods for business modeling are inadequate for managing the intricate, non-linear decisions required in these ecosystems, creating a need for a tool that can handle coordinated decision-making across the entire value network.
Outcome: - The developed software tool enables the systematic modeling, simulation, and comparison of different partner constellations in a service network. - It makes trade-offs between factors like cost, quality, and time explicit, linking operational activities directly to the value perceived by each stakeholder. - The approach proved effective in a hypothetical case and three real-world industry cases (mobility, construction, additive manufacturing) for designing feasible business models and identifying new stakeholders. - It provides a unified view of the value network, supporting the systematic analysis of dependencies and the development of collaborative strategies.

About: This paper investigates how software-as-a-service (SaaS) startups in Finland successfully scale their operations. Through a qualitative multiple-case study of six Finnish companies, the research develops an integrated model that examines the interplay between business, organizational, and engineering choices. The study identifies a coherent scaling strategy suited for companies operating in smaller, digitally advanced economies with limited domestic markets.
Problem: Scaling is a critical phase where many software startups fail, and existing research predominantly focuses on large, capital-rich ecosystems like Silicon Valley. This creates a knowledge gap regarding how startups navigate growth in smaller markets, such as Finland, which require early internationalization. The study addresses how these companies must integrate commercial, organizational, and technical strategies to achieve durable, global growth.
Outcome: - Finnish SaaS startups pursue early internationalization with lightweight local hubs due to the small domestic market. - Growth is driven by a capital-efficient, product-led growth (PLG) and inbound marketing model, with outbound sales added selectively for larger customers. - Companies maintain a disciplined single-product focus, using one codebase for all clients to ensure maintainability and scalability. - Technical debt is managed pragmatically, with teams prioritizing and fixing issues only when they become direct bottlenecks to business growth. - Operations rely on distributed, remote-first teams that use lean planning, persistent experimentation, and data-driven management. - External funding is typically used to accelerate proven growth strategies, not to initiate the scaling process.