CLSBE - Dissertações de Mestrado / Master Dissertations
URI permanente para esta coleção:
Navegar
Entradas recentes
- AI maturity and workforce displacement : a qualitative comparative study of governance across maturity stagesPublication . Schulz, Yannick Joël; Bohnsack, RenéArtificial intelligence (AI) is increasingly embedded in organisational activities, reshaping how tasks are performed, coordinated, and governed. While research on AI-related labour-market effects has focused mainly on occupational or sectoral exposure, AI maturity research has examined capability development, and AI governance research has rarely connected governance practices to specific displacement pressures. Less is known about how organisations at different AI maturity stages experience and govern workforce displacement. This study addresses this gap by examining AI-enabled workforce displacement at the firm level.The study draws on thirteen semi-structured expert interviews with firm-side and cross organisational informants. Transcripts were analysed through a Gioia-inspired abductive coding procedure.The findings show that AI implementation differed not only in tool availability, but in how AI was coordinated and embedded. AI capability remained unevenly distributed across functions, sites, and employee groups. Workforce displacement appeared most visibly as task reallocation within existing roles; role elimination and headcount reduction remained limited, indirect, or anticipated. The conversion of technical capability into realised workforce effects was shaped by organisational, legal, technical, and institutional filters. Governance developed unevenly: mitigation, the organisation of adaptation, became more formalised, while legitimation, the explanation of workforce consequences, remained indirect.The study contributes a firm-level account of AI-enabled workforce displacement as a stage conditioned conversion process rather than a direct result of occupational exposure. It further shows that AI maturity should be read as a dominant organisational configuration, and that responsible AI governance should be assessed by whether organisations communicate workforce consequences in explainable, contestable, and justifiable ways.
- How can a location-based digital platform support the coordination of shared sports activities while enabling a sustainable platform-based business model in urban environments?Publication . Kuckherm, Leandro; Xavier, RuteThis thesis examines how a location-based digital platform can support the coordination of shared sports activities while enabling a sustainable platform-based business model in urban environments. Drawing on digital platform theory, technology acceptance models, sports participation research, and platform business model frameworks, a sequential mixed methods approach was applied combining a secondary data-based market and competitor analysis, ten semi-structured interviews with urban sports practitioners, and a quantitative survey of 255 respondents. The market analysis identified a structural gap in the existing landscape of sports coordination platforms, with no current solution simultaneouslyaddressing the social and coordinative dimensions of shared outdoor sports participation. Qualitative findings revealed that coordination overhead, pace and schedule mismatch, social barriers, and safety concerns represent the primary obstacles to shared sports engagement, while user needs center on social connection, reliability, personal control, and coordination simplicity. The quantitative analysis confirmed that Perceived Usefulness is the dominant predictor of behavioral intention to use the platform (β=0.637, p<0.001), followed by Trust as a meaningful secondary driver (β=0.116, p=0.030). Feature preferences converged on the Live Map and Swipe feature as the most valued core functionalities, supported by a rating system and route privacy controls as trust-enabling components. A freemium business model with a geographically focused launch strategy is recommended, prioritizing local user density before scaling. The findings contribute to research on digital platform adoption in socially embedded consumer contexts and provide actionable insights for the design and commercialization of location-based sports coordination platforms.
- Stress-testing the 1/N rule : portfolio strategy performance across financial crisis phasesPublication . Birkenstock, Julian; Meira, MárioThis thesis investigates whether optimised portfolio strategies can consistently outperform naïve equal-weight diversification across distinct phases of financial crises. The analysis extends an unconditional portfolio evaluation framework by introducing explicit crisis-regime conditioning. Crises are identified using a drawdown-based classification methodology that partitions the market timeline into four sequential phases: pre-crisis, crisis, drawdown, and recovery. The empirical analysis spans three geographic regions, the United States, Europe, and Japan, and four asset universes derived from common equity factor models.Two layers of strategies are evaluated. The first comprises classical mean-variance and minimum-variance optimisation strategies. The second comprises phase-conditioned long-short strategies that construct portfolio signals from the return histories of prior crisis episodes. Results reveal a striking crisis and draw down-recovery asymmetry. During the sub-phases crisis and draw down, constrained minimum-variance strategies produce statistically significant out performance of the equal-weight benchmark across all regions and universes. During the recovery phase, all classical strategies significantly under perform the benchmark, reflecting the cost of backward-looking covariance estimates calibrated on high-stress market environments.Phase-conditioned long-short strategies generate full-sample Sharpe ratios and certainty equivalent returns above the equal-weight benchmark, an achievement absent from every phase of every classical strategy, with out performance concentrated in the sub-phases crisis and draw down. These findings demonstrate that the aggregate conclusion of the prior literature conceals structurally distinct phase-level behaviour. Regime-aware signal construction, rather than optimisation refinement alone, is the primary source of exploitable out performance relative to naïve diversification.
- Reconfiguration or replication? : strategic response to geopolitical disruption: a Finnair case studyPublication . Isringhausen, Luca Tom; Reis, RicardoOn the 24th of February 2022, Russia's invasion of Ukraine triggered an airspace closure that eliminated the competitive foundation of Finnair's business model overnight. For a decade, the Finnish carrier had built its fleet, network and organizational identity around a single geographic logic: Helsinki-Vantaa airport's position along the great-circle route between Europe and Northeast Asia, made viable by unimpeded Siberian corridor access.This thesis examines whether Finnair's response constitutes genuine strategic reconfiguration or operational adjustment strategically framed. The analysis adopts a longitudinal teaching case study format, drawing on secondary sources. Four frameworks from strategic management literature structure the analysis: Core Capabilities and Core Rigidities, Dynamic Capabilities, Strategic Crisis Response Typology and PESTEL Analysis.The findings support a position of genuine but incomplete reconfiguration. The commitment to restoring profitability regardless of whether Russian airspace reopens, the network rebalancing toward European and North Atlantic routes and the Embraer narrow-body renewal program represent real change in the firm's capability base. The Values dimension tells a different story: Helsinki continues to function as an intermediary hub dependent on sixth-freedom connecting traffic. The 2026–2029 strategy remains calibrated for one geopolitical assumption rather than designed to hold across several. The structural vulnerability has not been eliminated; it has been reoriented.
- Algorithmic brand discovery on Instagram and Its effect on brand trust and purchase intentionPublication . Schumacher, Sabina; Queirós, RitaAlgorithmic recommendations have transformed brand discovery on social media, creating new possibilities to personalize consumer experiences. This study analyses the impact of algorithmically driven recommendations on brand trust and purchase intention.The research employs a quantitative survey with an experimental between-subjects design. Independent-samples t-tests are used to examine the effectiveness of algorithmic cues, while linear regression and Hayes’ PROCESS macro are applied to investigate the mediating role of brand trust and moderating factors like involvement.The findings indicate that explicit algorithmic attribution does not significantly influence brand trust, suggesting a normalization effect and high baseline awareness among digital natives. Instead, perceived serendipity and objectivity are the main factors enhancing brand trust, which subsequently mediates purchase intention. Furthermore, the findings reveal a backfire effect for users with high privacy concerns and tech savviness. For those consumers, the perception of algorithmic influence triggers reactance rather than trust. Additionally, the results reveal that product involvement may not necessarily moderate the response toward algorithmic cues in fast-moving social media environments.The research concludes that effective algorithmic brand discovery requires a shift from mere technical precision towards responsible, serendipity- and objectivity-driven curation. These insights carry implications for brand managers, platform developers and policymakers, particularly regarding the need for clearer governance to protect consumers in increasingly automated environments.
- Grid-scale battery energy storage systems in Germany : responsible business opportunities and challenges in aligning deployment with societal goalsPublication . Sieler, Leander Noah; Cruz, Nuno Moreira daGrid-scale Battery Energy Storage Systems (BESS) are considered a critical flexibility technology for the energy transition. However, it is unclear whether the German market framework enables their deployment and operation in ways that align with societal objectives and electricity system needs. Based on the concepts of Creating Shared Value, Triple Bottom Line, and the Sustainable Development Goals, this dissertation examines the responsible business opportunities of grid-scale BESS, as well as the market barriers that may prevent this alignment. A qualitative research approach was used, involving nine semi-structured expert interviews with BESS market actors, grid operators and consultants. The findings show that grid-scale BESS represent an attractive business opportunity, with revenues primarily generated through electricity arbitrage and ancillary services. Meanwhile, BESS can create significant societal value by supporting the integration of renewables, improving grid stability and reducing overall electricity system costs. Yet, this alignment between economic and societal value does not emerge automatically. Current market structures fail to incentivize deployment and operation of facilities that would best serve the system. Key barriers include missing locational price signals, grid connection bottlenecks and regulatory uncertainty. The findings suggest that market actors behave rationally in response to existing incentives. Achieving system-optimal market outcomes, in line with societal objectives, therefore depends mainly on policymakers to create a coherent, predictable and incentive-aligned framework.
- Extreme weather disturbance of electrical substations : a geospatial prediction and impact assessment model for PortugalPublication . Bajramaj, Tim; Bertani, NicolòExtreme weather events are an increasing threat to electrical distribution infrastructure, yet data-driven approaches to substation-level disturbance prediction remain scarce, particularly at the most granular level of the high-to-medium voltage interface. This thesis develops and deploys a modular, open-data framework for predicting and contextualising weather-driven disturbances of substations on the Portuguese mainland. A three-type sampling strategy is used to train a calibrated Gradient Boosting classifier, whichachieves 81.3% accuracy, 70.9% recall, and a ROC-AUC of 0.866 on held-out test set of unseen installations. Daily and cumulative precipitation, temperature extremes, and geographic latitude emerge as the dominant disturbance predictors, with isolated 15 kV substations exhibiting disproportionate vulnerability. A self-developed line-length-proportional allocation mechanismassigns parish-level population, enterprise activity, and critical infrastructure exposure to individual substations across all 2,882 parishes, producing a bounded Criticality Index that ranks installations by their socio-economic consequence in case of disturbance. Both models are operationalized in Grid Sentinel PT, an open-source dashboard delivering automated, daily-updated,three-day-ahead disturbance risk forecasts for every substation on the Portuguese mainland. The framework is designed as a transferable blueprint for other European distribution operators.
- Measuring and explaining financial efficiency in brazilian listed firms : evidence from the social pillar and ESG disclosurePublication . Dietinger, Natalie; Wanke, Peter FernandesObjective: This study examines how efficiently publicly listed Brazilian firms transform financial inputs into financial outputs and to what extent the Social Pillar contributes to explaining differences in financial efficiency. Methodology: Financial efficiency scores are estimated using a hybrid Data Envelopment Analysis–Non-negative Matrix Factorization (DEA–NMF) approach with continuously differentiable hyperparameters (ρ, α, τ), selected through automated calibration. Followed by the second stage regression analysis where the efficiency scores are regressed against ESG and sector variables. Findings: The Social Pillar does not show a statistically significant relationship with financial efficiency for the dataset explored. However, Stakeholder Engagement exhibits a positive and significant association with financial efficiency, like the governance Pillar. The Environmental Expenditure variable shows a significantly negative correlation with financial efficiency. Conclusion: While the Social Pillar does not show an effect, Firms should focus their social incentives on their core stakeholders to boost the social pillar and create efficiency within the firm.
- Assessing the impact of next-generation products on the business model resilience of the austrian tobacco monopolyPublication . Prankl, Simon; Rajsingh, PeterThis dissertation assesses the impact of next-generation products (NGPs) on the business model resilience of the Austrian tobacco monopoly, administered by the Monopolverwaltung GmbH through a network of 4,308 licensed retailers (Trafiken). The market has undergone a structural transition from combustible to non-combustible revenue since 2020, accelerated by the Tax Amendment Act 2025. The study employs a mixed-methods triangulation design, combining quantitative time-series analysis (2018–2025) and revenue projections through 2030 with data from twelve semi-structured expert interviews conducted between April and May 2026. The findings confirm a measurable substitution dynamic: the NGP segment grew from 7.7% of total monopoly revenue in 2022 to 16.2% in 2025 and is projected to reach 26.6% by 2030. The margin differential between combustible cigarettes (12.9%) and newly regulated NGP categories (25.0% to 26.7% for specialty shops) embeds an economic incentive for portfolio reconfiguration within the legislative framework. Evaluated against the dynamic capabilities framework, the monopoly demonstrated strong institutional sensing capacity and an effective but operationally constrained seizing response, alongside an incomplete retail-level reconfiguration that risks bifurcating the network. The institution ultimately exhibited adaptive resilience, maintaining its core exclusive distribution function. The study's principal theoretical contribution is the inversion of the regulatory entrepreneurship mechanism, demonstrating that the regulated incumbent, rather than a disruptive entrant, deployed regulatory entrepreneurship tactics to achieve legislative consolidation.
- Two adoption logics, one tool : how and why talent agencies and content creators differ in evaluating AI operating systems in the DACH creator economyPublication . Geesink, Marleen Sophie; Almeida, SílviaThis dissertation examines how and why talent management agencies and individual content creators in the German-speaking Instagram creator economy differ in their evaluation of integrated AI operating systems (AI OS). The question is not whether established adoption frameworks predict who adopts, but where they break down when two structurally distinct actor types confront the same nascent technology. The study draws on the Technology Acceptance Model, the Unified Theory of Acceptance and Use of Technology, Diffusion of Innovations, and the Technology-Organisation-Environment framework as sensitising lenses. A comparative qualitative design analysed eighteen semi-structured interviews with ten individual creators and eight talent management agencies through a two-wave abductive thematic strategy. Participants evaluated a concrete prototype (marlu.ai) as empirical stimulus. Two structurally distinct adoption logics emerge. Adoption intention in this nascent category does not follow directly from perceived usefulness. Agencies and creators are not simply more or less willing to adopt the same tool, but follow different adoption logics. Agencies are often blocked before evaluation begins, because the category itself is not yet recognised in the field. Creators reach evaluation more readily, but they filter usefulness through autonomy, identity, trust, and delegation boundaries. Managed creators occupy a structurally intermediate position between the two logics, and a preliminary vicarious-adoption configuration decouples adopter from recommender. Established frameworks describe what creators evaluate but not what agencies do not. Such nascent markets require constructs operating upstream of the perceived usefulness pathway.