Effective risk management is crucial yet challenging in modern projects because early warning signals often surface in informal, unstructured data sources and evolve as project contexts change. Since prevailing project risk management (PRM) practices largely depend on structured inputs and episodic control routines, they can miss these subtle signals, creating a temporal detection gap between risk emergence and formal identification. This paper introduces an LLM-based architecture for information extraction in PRM. Following a Design Science Research approach, we design, instantiate, and evaluate it, governed by two complementary flows: an Information Flow that continuously transforms heterogeneous project communication into structured, evidence-linked risk indicators through domain-specialized extraction and cross-source reconciliation, and an Extraction Control Flow that versions extraction behavior to remain adaptable, transparent, and auditable. We abstract four nascent design principles capturing transferable design knowledge for LLM-based information extraction in dynamic, context-rich domains.
Climate change has disrupted the global water cycle, exposing hundreds of millions of people to unprecedented hydrological conditions. As water scarcity grows, urban green spaces increasingly struggle to regulate micro-climates, a key factor in maintaining liveable cities. Thus, smart city initiatives increasingly turn to digital solutions to conserve scarce resources, and smart, sustainable irrigation systems offer a promising approach. This study employs an Action Design Research approach to design and evaluate such a decision support system for urban green spaces, in close collaboration with the city of Frankfurt am Main, its urban green spaces department, and the botanical garden, which are responsible for irrigating 8000 young trees and 3800 shrubs. Combining theoretical foundations with practitioner insights from city administrators, smart city specialists, and botanists, the study shows how IoT-based soil moisture sensing and data analytics enable demand-oriented, resource-efficient irrigation. The work generates design knowledge contributions in the form of problem statements, design requirements, design features, and design principles, as well as a software prototype.
The growing integration of generative AI (GenAI) in workplaces raises concerns about the concealment of GenAI usage. This study examines how Dark Triad traits (narcissism, Machiavellianism, psychopathy) are associated with GenAI concealment behavior via moral disengagement and AI identity and how competitive climate moderates these relationships. Survey data from working professionals were analyzed using covariance-based structural equation modeling and moderation analysis, complemented by propensity score matching as a robustness check. Results show that all Dark Triad traits are positively associated with moral disengagement, while narcissism is additionally associated with AI identity. Contrary to expectations, moral disengagement is not associated with concealment. Instead, AI identity emerges as the construct most strongly associated with GenAI concealment, particularly in competitive work environments, with narcissism showing an indirect association via AI identity. These findings suggest that GenAI concealment reflects identity-based visibility regulation rather than purely moral disengagement. Organizations are therefore encouraged to build transparent environments that reduce competitive pressures and embed ethical considerations into AI practices, helping to prevent strategic concealment dynamics and promote responsible GenAI use.
As social media increasingly function as platforms for information communicating, they have become key venues for health information seeking while also enabling the rapid diffusion of misinformation. This study examines how linguistic characteristics of misinformation and information consistency between original posts and User-Generated Content (UGC) shape misinformation diffusion, with influencer status considered as a moderating factor. Using a hybrid analytical approach that combines text mining and negative binomial regression, we analyze 10,651 misinformation posts and 615,562 related UGC items concerning winter influenza and vaccines on Rednote. Consequently, we reveal that syntactic linguistic features, including sentence length, word composition, and emoji usage, are significantly associated with all three diffusion-related engagement behaviors, i.e., likes, comments, and reposts. Emotional cues exhibit heterogeneous effects across engagement types, where anxiety and anger suppress diffusion-related behaviors while sadness significantly increases reposting. Interestingly, users are more likely to comment on misinformation when UGC is topically similar to the original post, rather than simply liking or reposting it, suggesting evaluative engagement rather than passive endorsement. Further, at the pragmatic level, metaphorical expression and information consistency across syntactic, semantic, and pragmatic dimensions between source content and UGC significantly amplify misinformation diffusion. Importantly, influencer status positively moderates these relationships, strengthening the effects of linguistic features and information consistency, particularly for reposting. Overall, this study advances misinformation diffusion research by integrating linguistic, cross-text, and social influence mechanisms, and provides actionable insights for platform governance aimed at balancing user engagement with information integrity in digital information platforms.
Nowadays, retailers often use discounts to promote and monetize their digital goods. Nonetheless, little research attention has been paid to the way discount strategies are designed in a competitive market. To fill this gap, we investigate well-known competitor’s price effect on retailers’ optimal discount decisions (in both discount size and duration). Inspired by our observation, we developed an analytical model including a less-known retailer and a well-known retailer with the less-known retailer offering a discount. Our analytical results show that retailer’s discount size increases with the intensity of price competition. Furthermore, different from prior findings, the less-known retailer prefers pairing a larger (smaller) discount size with a longer (shorter) discount duration under competition. In addition, our findings reflect that offering a longer promotion helps offset dominant competitor’s advantage. Using a unique dataset of audiobooks, we empirically test the derived hypotheses. Most of our predictions are consistent with empirical results. Our study provides retailers with useful managerial insights of discount promotion.
Generative search is reshaping e-commerce discovery by replacing link-centered browsing with answer-first interfaces in which external sources appear as citations. This study argues that citations function not merely as transparency cues, but as endorsement tokens through which platforms allocate visibility, credibility, and verification opportunities across commerce-relevant actors. Using a paired field audit design, it compares Google AI Overviews with the search engine results page (SERP) top-10 baseline across shopping intents. The findings show that generative citations do not simply mirror conventional search rankings, but systematically reallocate marketplace exposure across official, evaluation, and transaction channels. By distinguishing citation breadth from citation intensity, the study develops a dual-lens account of exposure governance and shows that apparent citation diversity may coexist with concentrated endorsement intensity. The study contributes to research on electronic markets by explaining how generative interfaces govern marketplace visibility and credibility in online product discovery.
Online review systems are central to trust formation and governance on e-commerce platforms, yet fake reviews remain persistent despite ongoing regulatory efforts. Existing studies offer limited insight into how governance outcomes evolve over time. This study examines fake review governance from an evolutionary perspective by developing a tripartite evolutionary game model involving platforms, merchants, and consumers, analyzing how institutional parameters and initial conditions jointly shape long-term strategic trajectories. The results show that governance effectiveness does not improve linearly with regulatory intensity. Instead, the system may converge to multiple stable equilibria and exhibit strong path dependence. Early tolerance of review manipulation can lock the market into low-quality governance states, while later regulatory strengthening may fail to reverse these trajectories. Platform regulation affects merchant behavior in a nonlinear manner once key parameters cross-critical thresholds. Consumer behavior does not directly determine equilibrium selection, but instead influences the speed of strategic adjustment through information feedback mechanisms. These findings highlight the importance of early institutional design and sustained incentive alignment in platform governance.
GovTech is emerging as a policy-relevant frame at the intersection of digital government, public procurement, and startup innovation. Yet research remains fragmented and often disconnected from the institutional conditions that shape implementable solutions. This paper develops a stakeholder-validated, dimensional publicness theory informed research agenda for GovTech in the European context using a multiround agenda-setting Delphi study with experts from public administrations, industry, and academia. Synthesizing qualitative inputs and quantitative prioritization, we identify thematic clusters and topic areas, formulate exemplary research questions, and provide rationales that connect each area to concrete governance and implementation challenges. To assess dynamism, we additionally examine perceived relevance trends across two measurement points (2023 and 2025), highlighting where urgency is increasing and where research sequencing is needed. The resulting agenda informs scholarship by clarifying GovTech as a hybrid, regulation-shaped domain and offers entry points for empirical and design-oriented research on ecosystem formation, procurement, regulation, and strategic scaling.
Generative artificial intelligence (GenAI) is widely used in the gig economy and electronic markets. However, its progress incurs substantial environmental costs due to its high and constantly increasing energy consumption. As inference—the process by which GenAI generates output—accounts for a large part of this demand, individual usage patterns are becoming a key factor in understanding the overall energy footprint and a crucial prerequisite for developing effective educational strategies and policies. Drawing from the construal-level theory of psychological distance, this study explores individuals’ perceptions related to GenAI energy consumption. Based on a nationally representative sample of 1080 participants from the USA, we provide evidence that individuals are aware of GenAI-related energy consumption. Moreover, we show that, compared to household appliances, there is a diffusion of responsibility for GenAI’s energy consumption, with more emphasis placed on policy-makers and developers than on users. We also explore how individuals intend to implement pro-environmental strategies and shed light on different avenues suggested by individuals to lower energy consumption when using GenAI. We conclude with implications for theory, practice, and policymaking.
Generative AI is embedded in everyday knowledge work, but we know little about how workers preserve human agency when large language models (LLM) output enters real deliverables. Drawing on 15 interviews in two early-adopting German tech firms and analysis of handbooks and workflow artifacts, we examine how professionals keep control and responsibility when LLM generate parts of their work. We focus on three phases where people step in: drafting with the AI, refining its output, and reviewing the final product with others. We derive a nine-step checklist that specifies what practices to apply, and when, to keep AI-assisted text reliable, traceable, and accountable. We contribute by (1) explaining how ownership and discretion jointly ground human agency in GenAI workflows, (2) mapping when and how judgment shifts between humans and LLMs across drafting, refining, and reviewing, and (3) deriving a nine-step Oversight Protocol (OP-9) that embeds simple checks at key points in the workflow.
Machine learning (ML) advice complements human expertise by offering distinct strengths in judgment and prediction, thereby improving decision outcomes. While previous research has examined individual responses to human or ML advisors, the dynamics of human–ML ensemble advice remain unexplored. This study investigates how decision-makers integrate knowledge claims when receiving advice from humans and/or ML advisors, and how emerging interactions affect decision quality at individual and group levels within an organizational context. Based on a field experiment in Porsche AG’s after-sales department, we identified distinct activities of engaged and unengaged advice-taking that facilitate knowledge transfer and shape decision quality. Our findings reveal that while individuals integrate their own knowledge more often in single-advisor settings (with only human or ML advice), human–ML ensemble advisory settings yield better decision outcomes when decision-makers actively reconcile diverse inputs. Conversely, in single-advisor settings, aligning with the advisor improves decision quality. Notably, engagement dropped from individual to group settings. Participants who integrate all decision-makers and advice input also tend to perform better. These results underscore the importance of the advisory context and offer design implications for collaboration in complex organizational decision-making, highlighting the role of ML-supported platforms in shaping the future of digital knowledge work.
As artificial intelligence transforms customer interactions, enterprise conversational agents (ECAs) have become essential brand touchpoints. However, organizations often struggle to ensure that these agents consistently express their brand identity. While prior research highlights the importance of conversational design for brand identity, existing studies offer a fragmented understanding of how brand identity manifests in ECAs. This study presents a systematic review (n = 69) to consolidate current insights into the relationship between ECA design and brand identity. The review identifies key design cues that support alignment between conversational style and brand expression and explains how these cues, shaped by user and contextual factors, influence perceptions of both the ECA and the brand. The resulting framework links design elements to brand-related outcomes and provides a structured foundation for practice and research, enabling the development of ECAs that communicate brand identity more effectively within AI-mediated customer interactions.
Generative AI enables autonomous software agents that can search, compare, and transact across digital marketplaces, promising large reductions in consumer search costs and improved matching between buyers and sellers. This paper argues that such gains are not automatic. Drawing on economic search theory, we first discuss the impact of reduced search costs on markets. Then, we show how the behavior of current AI agents introduces frictions that limit competitive outcomes. Empirical studies reveal persistent deviations of AI agents from optimal search behavior that function as behavioral search costs even when technical search costs approach zero. At the same time, AI-generated content contributes to signal dilution, reducing the informativeness of offers by AI agents and weakening effective product differentiation. These forces interact with low entry costs, which encourage excessive and often low-value entry. Together, they can trap agentic markets in inefficient equilibria. We outline key implications for electronic marketplaces and highlight promising directions for future research on agentic markets.
Past studies on psychological ownership and financial technology (Fintech) innovation have typically examined these domains separately, providing limited insight into how psychological ownership of Fintech innovation influences financial inclusion. This study integrates the Customer Value Theory (CVT), Self-determination Theory (SDT), and the Theory of Planned Behaviour (TPB) to explore how psychological ownership of Fintech innovations affects financial inclusion through socio-economic engagement. Data were gathered from 630 mobile money (MoMo) users in Ghana through a survey and 23 interviews. The findings show that continued use of MoMo promotes psychological ownership and positively influences social, economic, and financial engagement. However, these benefits are undermined by the perceived government policy (electronic levy), which negatively affects socio-economic engagement. The findings call for Fintech stakeholders to enhance users’ psychological ownership and encourage the government to adopt pro-poor policies to improve financial inclusion. This study uniquely combines the CVT, SDT, and TPB to investigate the role of psychological ownership in Fintech diffusion and its socio-economic impact.
Products developed through human-artificial intelligence collaboration (HAIC) are emerging in high-stakes domains, such as pharmaceuticals, where failure can lead to harm. These products contain information asymmetries and tensions in agency configuration, whether the process is AI- or human-dominant. The information systems (IS) literature has few studies examining consumer acceptance of these products through the lens of transparency, social acceptance, and agency configuration. In response, this study explores two research questions: (1) How does the transparency in the HAIC process and the social value of the HAIC-developed product shape consumers’ trust and acceptance? and (2) How does the agency configuration affect the relationship between social value and consumers’ trust and acceptance? These questions are grounded in a conceptual model drawing on signaling theory and the theory of consumption values to examine consumer acceptance of HAIC-developed products in high-stakes domains, such as pharmaceuticals. The study employs an experimental design with simulated packaging disclosures of agency configuration (AI-dominant vs. human-dominant) for an over-the-counter medicine. It uses partial least squares structural equation modeling (PLS-SEM) to analyze the data (n1 = 262 and n2 = 238 consumers in the AI-dominant and human-dominant conditions, respectively). The findings reveal that transparency in the HAIC process increases trust in both conditions but enhances social value only in the human-dominant condition. Social value increases trust and acceptance, with these effects being stronger in the AI-dominant configuration. The study advances IS research by offering guidance on building consumer trust and acceptance through the disclosure of internal decisions about HAIC processes.
The prevalence of spam reviews on e-commerce platforms undermines consumer trust, distorts product evaluations, and often leads to suboptimal purchasing decisions. This study proposes a spam-aware recommendation framework that first detects spam reviews using machine learning and deep learning models based on semantic, sentiment, and metadata features, and then integrates these filtered credible reviews into personalized product recommendations. Rather than treating spam detection and recommender design as isolated tasks, this work bridges the gap between review credibility and user preference modelling to improve recommendation accuracy. By ensuring that only trustworthy reviews contribute to recommendation generation, the proposed approach enables users to receive more reliable and preference-aligned suggestions, thereby supporting more informed and confident purchasing decisions.
Digital platforms rely on third-party complementors to create and sustain value over time. While early research has emphasized platform strategies to attract and onboard complementors, less is known about how complementors sustain their engagement with a platform after the initial product launch. We conceptualize complementor engagement as sustained, behaviorally enacted involvement in a platform and examine how ongoing product enhancement activities—specifically, routine maintenance and adaptation to platform-wide changes—act as observable manifestations of this engagement. Focusing on product complexity arising from the addition of features, which is facilitated by the unique characteristics of digital platforms, we investigate how complexity influences the sustained engagement behaviors of complementors. Using a multimethod approach that combines longitudinal panel data from the Apple iOS App Store with a randomized experiment, we find that more complex products are associated with more frequent product maintenance, indicating stronger ongoing engagement at the product level. At the same time, while the experimental participants perceive higher product complexity as substantially constraining their ability to adapt quickly to platform-wide changes, archival evidence reveals that complexity does not slow complementors’ actual adaptation in practice. This difference between perceived difficulty and enacted behavior provides an important insight into how complementors navigate real market incentives, development capabilities, and platform governance when responding to platform evolution. Collectively, our findings clarify how product complexity shapes sustained complementor engagement and contribute to research on digital product strategy, platform governance, and the behavioral foundations of engagement in digital platform ecosystems.
Despite the proliferation of agentic technologies in service systems, empirical research on bridging AI agents into physical settings remains scarce. This study examines the integration of conversational AI agents as dynamic touchpoints within physical service environments, exploring how one or more human actors form and renegotiate agency configurations with these non-human agents. The study conducts an in-depth single-case study within a garden center retail setting, comprising observations and condensed semi-structured interviews with 83 customers across 42 interactions. Our qualitative analysis develops a framework that explains how agency configurations are shaped by 20 attributes across the stages of agency formation and iterative agency re-configuration. Agency formation emerges as an evaluation of attributes such as AI agent role and value, shaped by social dynamics and contextual factors. Agency re-configuration is triggered by information exchange and situational contingencies and facilitated by the availability of instructive mechanisms and loop learning. The findings advance understanding of value co-creation in physical service environments involving AI agents, highlighting levers for orchestrating effective agency configurations and addressing tensions between human agency and AI agency in polyadic human-AI hybrids.
This study examines how social media influencers affect pricing dynamics in electronic marketplaces. Drawing on platform ecosystem theory and differentiated demand models as a theoretical lens, we conceptualise influencers as attention brokers who reallocate consumer attention within algorithmic marketplace environments, thereby reducing effective price sensitivity among exposed consumers. Using a panel dataset combining YouTube influencer campaign data and Amazon product-level data (price, sales rank, reviews, and ratings), we employ a two-stage least squares (2SLS) framework to assess how influencer exposure alters the observed price-demand relationship. Our findings indicate that products featured in influencer campaigns exhibit meaningfully lower observed price responsiveness in sales rank-based demand proxies compared to non-featured products. These results are consistent with the theoretical prediction that influencer-driven attention shifts reduce the slope of the effective demand curve, enabling sellers to maintain higher price positions with limited sales rank deterioration. We interpret these results as reduced-form evidence of altered competitive price pressure rather than structurally identified causal elasticity estimates, and we acknowledge the limitations of the identification strategy. The study contributes to platform pricing theory and influencer marketing research by providing empirical evidence that social influence processes can reshape pricing power in highly transparent electronic marketplaces.
