Table 1 Analytical framework: TOE... | Emerald Publishing

Table 1

Analytical framework: TOE factors retained, with literature-based justification

Dimension	Factor/sub-factor	Key literature and rationale for inclusion
Technological	Relative advantage	Rogers (2003), Tornatzky and Fleischer (1990) and Pumplun et al. (2019). Most robust predictor of adoption; in welfare AI, advantage must be weighed against constitutional trade-offs (augmentation over substitution)
Technological	Compatibility (legacy, workflows and data)	Baker (2012) and Neumann et al. (2024). Legacy infrastructure is decisive in bureaucracies with long-standing IT stacks; compatibility with on-premise data regimes is particularly binding under GDPR
Technological	Complexity	Rogers (2003) and Pumplun et al. (2019). AI-specific opacity magnifies complexity; mitigated by in-house expertise and IT–policy integration
Organisational	Top-management support	Mikalef et al. (2022) and Chen et al. (2024). Leadership bridges environmental pressures and allocative decisions
Organisational	Culture (innovation climate and change readiness)	Neumann et al. (2024) and Damanpour (1991). AI adoption collides with routine-driven bureaucratic cultures, culture conditions the pace of change
Organisational	Resources (financial, human capital and data)	Mikalef et al. (2022) and Pumplun et al. (2019). Data resources are a distinctive AI determinant; human-capital scarcity can paradoxically accelerate adoption
Organisational	Structure	Baker (2012) and Fountain (2001). Formalisation and centralisation condition the form of adoption
Organisational	Absorptive capacity	Cohen and Levinthal (1990). Cumulative knowledge shapes ability to integrate new technologies; extended in Section 5.2 into organisational metabolism
Environmental	Regulatory framework (GDPR, AI Act and constitutional mandates)	Neumann et al. (2024) and Grimmelikhuijsen and Meijer (2022), extended to constitutional norms
Environmental	Stakeholder pressures (citizens, employees and unions)	Wang et al. (2024) and Alon-Barkat and Busuioc (2023). Stakeholder pressures shape legitimacy judgements, particularly for citizen-facing systems
Environmental	External support (policy, funding and networks)	Mikalef et al. (2022). National policies and EU funding operate as enablers, particularly under resource scarcity
Environmental	Demographic/fiscal pressures	Berryhill et al. (2019), this study. In welfare institutions, demographic trajectories function as non-discretionary adoption triggers

Dimension	Factor/sub-factor	Key literature and rationale for inclusion
Technological	Relative advantage	Rogers (2003), Tornatzky and Fleischer (1990) and Pumplun et al. (2019). Most robust predictor of adoption; in welfare AI, advantage must be weighed against constitutional trade-offs (augmentation over substitution)
Technological	Compatibility (legacy, workflows and data)	Baker (2012) and Neumann et al. (2024). Legacy infrastructure is decisive in bureaucracies with long-standing IT stacks; compatibility with on-premise data regimes is particularly binding under GDPR
Technological	Complexity	Rogers (2003) and Pumplun et al. (2019). AI-specific opacity magnifies complexity; mitigated by in-house expertise and IT–policy integration
Organisational	Top-management support	Mikalef et al. (2022) and Chen et al. (2024). Leadership bridges environmental pressures and allocative decisions
Organisational	Culture (innovation climate and change readiness)	Neumann et al. (2024) and Damanpour (1991). AI adoption collides with routine-driven bureaucratic cultures, culture conditions the pace of change
Organisational	Resources (financial, human capital and data)	Mikalef et al. (2022) and Pumplun et al. (2019). Data resources are a distinctive AI determinant; human-capital scarcity can paradoxically accelerate adoption
Organisational	Structure	Baker (2012) and Fountain (2001). Formalisation and centralisation condition the form of adoption
Organisational	Absorptive capacity	Cohen and Levinthal (1990). Cumulative knowledge shapes ability to integrate new technologies; extended in Section 5.2 into organisational metabolism
Environmental	Regulatory framework (GDPR, AI Act and constitutional mandates)	Neumann et al. (2024) and Grimmelikhuijsen and Meijer (2022), extended to constitutional norms
Environmental	Stakeholder pressures (citizens, employees and unions)	Wang et al. (2024) and Alon-Barkat and Busuioc (2023). Stakeholder pressures shape legitimacy judgements, particularly for citizen-facing systems
Environmental	External support (policy, funding and networks)	Mikalef et al. (2022). National policies and EU funding operate as enablers, particularly under resource scarcity
Environmental	Demographic/fiscal pressures	Berryhill et al. (2019), this study. In welfare institutions, demographic trajectories function as non-discretionary adoption triggers