Digital transformation initiatives frequently fail due to organizational silos that impede knowledge flow. This conceptual paper applies Granovetter’s theory of weak and strong ties to conceptualize projects as network catalysts for organizational development. Based on a systematic literature synthesis, we develop an exploratory framework explaining how projects may create weak ties between separated organizational units. We derive four preliminary propositions specifying the conditions under which project-induced network structures may promote both operational objectives and strategic change. The framework integrates three theoretical streams: network theory foundations, contingency perspectives on knowledge transfer, and dual-objective achievement mechanisms. The framework suggests that network-aware project design represents an underutilized lever for organizational transformation, particularly in digital transformation contexts where cross-functional collaboration is essential.

Agentic systems have emerged as promising approaches for automating complex software development processes, particularly within data analysis workflows. However, their application in industrial settings, where data is heterogeneous, schemas evolve frequently, and analytical errors carry significant operational consequences, remains largely unexplored. This study develops and deploys an agentic data analysis pipeline for a Fast Moving Consumer Goods (FMCG) production plant, processing production data to generate insights on Key Performance Indicators (KPIs). Over a ten-day evaluation period, deployed across 18 production lines, the system achieved 80.59% task completeness and 94.02% correctness, in addition to expert ratings of 4.53 mean understandability and 4.36 mean goal alignment on a five-point Likert scale. These results demonstrate that agentic systems can operate reliably in industrial environments, delivering accurate and actionable insights despite data complexity and high stakes decision making contexts. The framework configuration and implementation are publicly available on Github.

Software selection represents a far-reaching decision problem for organizations, particularly in large enterprises where platform decisions affect multiple business units and long-term data strategies. In the domain of data platforms in particular, inadequate software selections may result in global and systemic challenges for data-driven enterprises. Consequently, a structured, objective, and multicriteria decision-making process is critical for maintaining competitiveness. Decision Support Systems (DSS) provide suitable means to address this complexity. This paper investigates the value contribution of a DSS by applying a two-stage Analytic Hierarchy Process (AHP) within an Original Equipment Manufacturer (OEM), using the BMW Group as an illustrative case. The effectiveness of the DSS is evaluated through a structured user study, enabling a comparison with software selection approaches based on individual consultancy. Furthermore, the DSS is analyzed both with and without the support of an individualized generative AI (GenAI) chatbot in order to assess its impact on the decision-making process. The results demonstrate the potential of DSS to enhance transparency, objectivity, and decision quality in complex software selection scenarios.

The increasing complexity of global supply chains has exposed the limitations of traditional simulation paradigms, which often rely on centralised assumptions and fail to reflect the distributed nature of decision-making. This paper introduces the conceptual foundations of the Supply Chain Simulator as a Service (SC-SaaS), a modular and privacy-preserving framework designed to support collaborative simulation across autonomous actors. SC-SaaS allows each stakeholder to model its own segment independently, share only the data selected, and create system-level insights only when collaboration is explicitly enabled. The framework investigates the main challenges in supply chain modelling, such as the achievement of different objectives, information asymmetry, barriers to interoperability, and the spread of uncertainty.