Integrated continued care as a space for collective action: the strategic actor game in the era of artificial intelligence Open Access

Health systems worldwide are undergoing profound structural transformations, driven by the convergence of cultural, social, political and economic forces that are fundamentally reshaping the way care is delivered. These transformations are particularly evident in integrated care frameworks, which face mounting challenges because of demographic shifts marked by population ageing, changes in morbidity patterns that reflect evolving disease profiles, growing demand for seamless care transitions across services and ever-tightening constraints on financial sustainability that call into question traditional funding mechanisms (Bruin et al., 2020). In parallel, artificial intelligence (AI) is radically transforming health-care work practices, organisational configurations and leadership models, creating both opportunities and tensions for health and care providers (Pavuluri et al., 2024).

Despite continued policy efforts and significant investment aimed at promoting integrated health and social care systems, fragmentation remains widespread, with rural regions or areas characterised by vulnerable demographics experiencing the most pronounced deficiencies. In Portugal, the National Integrated Continued Care Network (NICCN) was established as a strategic response to such challenges. However, territorial inequalities, weak inter-sectoral coordination and limited stakeholder engagement still compromise the network’s operational capacity and strategic effectiveness. These issues are especially acute in regions such as Alentejo, where demographic ageing and geographically dispersed populations exert considerable pressure on the availability and coherence of care services (Ferreira et al., 2021).

The growing incorporation of AI within care systems introduces additional complexity, requiring large-scale reskilling programmes, organisational restructuring and the emergence of leadership capable of steering digital transformation while safeguarding care quality (Keller et al., 2024). The governance implications of AI are now recognised as a critical dimension of future health system resilience, although concerns remain regarding its disruptive potential across entrenched practices and institutional hierarchies (Davenport and Kalakota, 2019). While existing academic literature sufficiently documents AI’s technical affordances, there is a clear lack of research into its broader organisational and human-resource effects, particularly in integrated care settings that straddle multiple administrative and institutional boundaries (Ghosh et al., 2025).

This gap extends to questions of how AI alters power relations, modifies patterns of actor interdependence and redefines leadership roles in decentralised, multi-actor governance contexts typical of modern Integrated Continued Care (ICC) systems (Wagner et al., 2024). Additionally, empirical insight remains limited regarding AI’s implications for workforce planning and collective action strategies, especially in ageing regions faced with persistent resource constraints, where conventional responses may no longer suffice (Silcox et al., 2024).

This study aims to address these research gaps through a strategic foresight analysis of the NICCN in the Alentejo region. It investigates how configurations of institutional actors, their strategic objectives and interdependencies shape the evolution of ICC. Simultaneously, it examines how AI implementation is transforming workforce structures, competency requirements and leadership modalities. By framing ICC as a system of collective action, the research explores how it can be guided towards sustainable and inclusive futures, with an emphasis on understanding the strategic roles of key institutional stakeholders.

The study is organised around six interconnected objectives, each associated with a guiding research question. The first identifies core variables and system components that shape ICC in Alentejo: what are the essential elements underpinning the system? The second characterises key actors and their formal and informal roles: who are the institutional players, and what roles do they fulfil? The third explores how interdependence and power dynamics influence collaboration: how do power relations affect collective governance? The fourth assesses the extent of alignment around shared aims: to what degree do actors coalesce around common objectives? The fifth proposes strategic mechanisms to enhance cooperation: what tools can improve coordination and collective action? Finally, the sixth investigates the effects of AI integration: how is AI reshaping organisational design, workforce functions and leadership approaches?

Methodologically, the study applies Michel Godet’s scenario planning framework (Godet, 1994, 1997; Godet et al., 2000) alongside the Actors, Objectives, Power Relations Method (MACTOR) for analysing actor influence, convergence and strategic alignment. Empirical data were collected via semi-structured interviews with 27 institutional actors operating within the NICCN in Alentejo, encompassing health-care providers, regional administrators, Private Social Solidarity Institutions (IPSS) and coordination structures.

The study offers three principal contributions. Theoretically, it expands the application of strategic action theory by incorporating the systemic effects of AI disruption. Practically, it provides actionable insights into governance and workforce planning strategies in integrated care. Socially, it contributes to the design of equitable care policies tailored to the needs of ageing, resource-constrained regions.

The article proceeds as follows. Section 1 reviews relevant literature on ICC, AI integration, governance and collective action. Section 2 outlines the methodology and analytical instruments. Section 3 presents empirical findings. Section 4 discusses results and broader implications. Section 5 concludes with policy recommendations and directions for future research.

ICC emerged as a response to health-care delivery fragmentation, aiming to enhance quality of life through coordinated, multidisciplinary interventions. In Portugal, the NICCN, established by Decree-Law No. 101/2006, bridges community and hospital-based services, emphasising rehabilitation, autonomy and holistic well-being of service users. Socio-demographic transformations, particularly ageing populations and increased chronic condition prevalence, challenge effective ICC policy implementation (Henriques and Gandra, 2022). Internationally, ICC encompasses health-care and social support system integration, prioritising care continuity and intersectoral coordination (Cresswell et al., 2018). Conceptual diversity persists as terminological variations reflect differing policy traditions and service models (Goddard, 2023). Nevertheless, a shared focus on person-centred care and vertical and horizontal sectoral articulation remains evident (Barry et al., 2021). Within this context, AI integration introduces additional complexity, requiring continuous adaptation of clinical pathways, evaluation frameworks and decision-making processes (Chen et al., 2024). AI serves dual functions in facilitating integrated care: providing predictive insights for patient stratification while enhancing routine task automation (Soenksen et al., 2022). However, AI adoption necessitates new governance models and care provider skillsets (Rees et al., 2023). These technological shifts demand flexible evaluation systems capable of adapting to AI-driven innovations that challenge conventional professional boundaries and decision-making hierarchies (Bartsch et al., 2025).

The NICCN functions as a strategic policy instrument, promoting continuity of care by integrating the health and social security sectors. Its multi-tiered structure supports collaborative networks and personalised care plans aimed at reducing service fragmentation (Decree-Law 101/2006, 2006; Henriques and Gandra, 2022). As an intergovernmental and inter-organisational platform, the NICCN brings together public health, social security and private entities through vertically and horizontally articulated governance models. Integration relies not only on institutional collaboration but also on dynamic interdependencies among stakeholders (KN and Thomas, 2021). Increasingly, these relationships are mediated by digital infrastructures, with AI playing a crucial role in coordination, referral processes and clinical decision-making (Periáñez et al., 2024). Nevertheless, effective AI implementation depends on robust information systems and interoperable technologies (Golden et al., 2024). Ongoing barriers hinder equitable access, resource optimisation and workforce development (Goiana-da-Silva et al., 2024). AI integration intensifies these challenges, requiring reskilling focused on digital literacy, data ethics and collaborative intelligence (Arias Hernández and Rockembach, 2025). Crucially, professionals must be supported in transitioning into AI-supported roles, safeguarding the human-centred ethos of care (Henzler et al., 2025).

ICC systems face mounting pressure because of demographic ageing, multimorbidity and financial constraints (OECD, 2021). Addressing long-term care needs requires integrated solutions that go beyond episodic treatment models (Kelly et al., 2020). Cost-efficiency and service optimisation are vital, particularly in rural areas with limited specialist access (Özkaytan et al., 2024). AI-powered innovations offer opportunities for care continuity and personalised interventions (Johnson et al., 2021) but depend on overcoming interoperability barriers, data security concerns and organisational readiness (Martin et al., 2022). Integrated governance demands alignment between health, social security and community sectors, a persistent challenge globally (Harnett et al., 2020). Human resource governance is pivotal for ICC sustainability (Witter et al., 2020). Portuguese studies highlight regional disparities, discharge coordination gaps and care admission constraints within the NICCN (Costa and Mourão, 2015; Santana et al., 2018; Rainho et al., 2021; Santos et al., 2022). Strategic workforce planning is essential to adapt to AI-led changes (Sutton et al., 2023; Walshe et al., 2024). ICC leadership must shift towards participatory, digitally skilled models (Lee and Cosgrove, 2024), recognising AI’s cultural and managerial implications (Hao and Demir, 2024).

ICC systems function as socio-political constructs shaped by collective action, negotiation and strategic interdependence (Correia de Matos et al., 2024). Theoretical models from Crozier and Friedberg (2000), Martin (2012) and Friedberg (1995) frame ICC as dynamic, power-mediated systems influenced by actors’ strategies in uncertain contexts. Within the NICCN, governance emerges from interdependent relations between central authorities, local providers, private entities and community groups (Santana et al., 2014). The integration of AI is disrupting these dynamics by redefining clinical responsibilities, influencing referral flows and altering decision hierarchies (Guleria et al., 2024). ICC actors must adapt strategies to incorporate AI while upholding principles of equity and social justice (Laein, 2024). Actor mapping is essential to identify shifts in influence, especially as new players – such as tech firms and data analysts – gain prominence (Ryan et al., 2024). Strategic planning must reconcile institutional inertia with innovation, while reskilling efforts should promote collaborative, ethical leadership to navigate AI-related uncertainty (Clauss and Ritala, 2023). As governance evolves, AI strengthens coordination, predictive capacity and interoperability, bolstering ICC sustainability (Dicuonzo et al., 2023).

This investigation used a prospective analysis of the NICCN in Portugal’s Alentejo region to understand change trajectories, actor dynamics and the implications of AI for workforce transformation and strategic adaptation. The study used the French School’s prospective methodology (Godet et al., 2000; Godet, 2007), adopting strategic analysis with triangulated quantitative and qualitative methods for decision support and strategic recommendations.

The central research question examined how ICC, conceived as a collective action system, can be effectively steered towards desirable futures and identified strategic actors’ roles in this process. Six specific objectives guided the research:

1. identifying key ICC variables;

2. characterising principal actors;

3. assessing inter-actor power relations;

4. evaluating actor positioning regarding strategic objectives, examining mobilisation, conflict and alliance potential;

5. proposing strategic recommendations for informed action planning; and

6. examining how AI reshapes organisational structures, workforce dynamics, reskilling imperatives and leadership strategies critical to integrated care systems’ future sustainability.

This methodological approach facilitated a comprehensive understanding of institutional interdependencies, power configurations and strategic alignments while addressing emerging technological transformations affecting integrated care governance and delivery mechanisms.

Strategic analysis involves strategic thinking, environmental scanning and assessment of territorial or organisational dynamics, requiring a forward-looking approach to support future-oriented decision-making (Godet et al., 2000; Guerra, 2006). Strategy is defined as a coherent framework integrating objectives, policies and action sequences, benefiting from methodologies combining qualitative insights with quantitative rigour, enabling comprehensive analysis of means-ends relationships, actor interactions and scenario development (Perestrelo, 2017). The scenario method, developed by Michel Godet and the LIPSOR team within the French School of Prospective, aims to minimise inconsistencies and foster collective understanding. Unlike the more intuitive Anglo-Saxon tradition, this structured, action-oriented methodology is divided into two phases: base construction (system delimitation, structural analysis and actor strategy evaluation) and scenario construction (hypothesis formulation, expert consultation and scenario prioritisation) (Godet, 1997; Saragoça et al., 2017). This study uses the MACTOR, rooted in Godet’s methodology, to examine strategic positioning within the NICCN. Actor strategy analysis facilitates the identification and characterisation of key actors, mapping interrelations and power dynamics to derive strategic recommendations (Perestrelo, 2002; Saragoça et al., 2017). Previous empirical applications of the MACTOR method in policy and governance contexts include studies such as Saragoça et al. (2017) in education planning and Perestrelo (2002) in social policy analysis. These studies demonstrate the method’s utility in mapping actor dynamics and strategic interactions, providing a relevant foundation for its application in the present study.

The MACTOR method systematically assesses strategy divergences and convergences, classifying actors by mobilisation and conflict levels (Godet, 2007). While enhancing strategic reflection, its application demands high stakeholder engagement, effective data collection and rigorous validation. Despite limitations in data acquisition and rationality assumptions, the MACTOR software provides valuable operational tools for scenario-based and independent strategic analysis (Godet, 2007).

It is important to clarify how AI was operationalised within the analytical framework of this study. AI did not function as a computational or data-processing tool embedded within the MACTOR software itself. Rather, AI served as a thematic and analytical lens through which actor strategies, objectives and interdependencies were interpreted and evaluated. Specifically, during the semi-structured interview phase, actors were invited to reflect on how AI integration was already affecting or was expected to affect their institutional roles, decision-making processes, workforce configurations and inter-organisational relationships. These responses were systematically coded and integrated into the MACTOR matrices, notably the actor-objective matrices, namely, the simple position matrix (1MAO), the valued position matrix (2MAO) and the mobilisation matrix (3MAO), as well as the influence-dependence matrices, including the direct influence matrix (DIM) and the indirect influence matrix (DIIM), to assess the degree of alignment or divergence between actors on AI‐related objectives. In this way, AI served as both a substantive theme structuring the content of strategic objectives and as an interpretive frame for understanding emerging shifts in governance and collective action within the NICCN. This clarification reinforces the methodological coherence between the study’s research aims and its analytical execution.

This study used single-case sampling within the Alentejo Health Region, the designated NICCN intervention area. The Alentejo region exhibits pronounced ageing demographics (National Institute of Statistics, 2021) and polarised ICC demand, predominantly serving individuals over 70 years (Regional Coordination and Development Commission of Alentejo, 2023). Participant selection prioritised stakeholder diversity and ICC expertise across national, regional and local coordination levels, encompassing hospital referrals and provider teams. Public and third-sector organisations were included as collective actors rather than individual entities. Selection criteria used maximum variation and convenience sampling (Savitsky et al., 2023), representing all four Alentejo sub-regions and diverse NICCN typologies. MACTOR analysis required actor aggregation based on competency and attribute similarities, following Godet’s (2007) principle that groups constitute single actors when their internal value systems, information networks and relational structures are indistinguishable. In this study, the initial 27 actors were aggregated into 12 macro-typological actors based on four criteria:

1. shared institutional competency domains within the NICCN governance structure;

2. similar hierarchical positioning across national, regional and local coordination levels;

3. equivalent formal mandates as defined by Decree-Law No. 101/2006 and its subsequent revisions; and

4. overlapping strategic orientations identified through documentary analysis and exploratory interviews.

This criterion-based aggregation procedure enhances the transparency and replicability of the analysis.

This study used a mixed-methods approach comprising documentary analysis, exploratory interviews and semi-structured interviews. Initially, official ICC policy documents were collected to identify strategic challenges, objectives and key regional actors within Alentejo. Exploratory interviews with ICC specialists provided contextual insights into actor dynamics and influencing factors, informing the development of the subsequent semi-structured interview guide.

The interview guide incorporated socio-demographic characteristics, strategic challenges, action mechanisms, actor relations and critical success factors, using open-ended questions and MACTOR-weighted scales. Audio-recorded face-to-face interviews were transcribed and subjected to systematic content analysis. Content analysis followed the framework proposed by Cohen et al. (2018), focusing on both documentary and interview data.

The MACTOR method facilitated data structuring, while categorical analysis based on Bardin (2016) addressed open-ended responses through pre-analysis, material exploration and result interpretation, organised around action mechanisms, obstacles and critical success factors. This analytical process aimed to elucidate actor positioning, available resources, barriers and critical success factors for ICC system implementation, integrating quantitative and qualitative data perspectives.

The MACTOR-derived quantitative matrices were triangulated with qualitative content analysis findings (Bardin, 2016) derived from interviews, as well as with insights from the documentary analysis of ICC policy documents. This triangulation approach enhances the robustness and validity of the findings by integrating multiple data sources and analytical perspectives.

The complexity of identifying actors shaping ICC system futures arises from organisational influence over key variables, while AI-driven transformations exacerbate strategic challenges in Alentejo because of geographical dispersion, diverse actor positioning and limited external engagement. Through environmental analysis, 27 actors were initially surveyed via semi-structured interviews, then grouped for MACTOR input into a 12-actor macro-typology, forming the institutional core of ICC provision at national, regional and local levels (Table 1).

AI transforms workforce dynamics, requiring reskilling strategies across organisational structures. Collective actors were classified by activity scope, either (inter)sectoral, concentrating on care provision, or territorial, operating across local, regional and national levels (Figure 1).

Figure 1 illustrates territorial actors operating exclusively within Alentejo, primarily engaged in generalist activities. Meanwhile, AI-driven workforce adaptation strengthens ICC leadership across hierarchical levels. Results reveal a predominance of internal actors and weak engagement from generalist entities, limiting ICC’s external interaction, hindering AI-enabled reskilling and network reinforcement, while the absence of informal caregivers restricts policy formulation, reinforcing an institutional framework where AI-driven organisational shifts redefine actor roles across public and private sectors.

After conducting an internal and external analysis of the Alentejo region’s NICCN, which identified key strategic challenges and influential actors, participants were invited to assess their stance on 6 strategic challenges and 18 related objectives (Table 2). The increasing integration of AI into workforce dynamics has significantly transformed organisational structures, requiring targeted reskilling strategies and adaptable leadership frameworks.

Each strategic challenge, referred to as a “battlefield”, is structured around specific objectives that delineate areas where actors converge, diverge or remain neutral (Godet, 2007). The growing integration of AI necessitates a reassessment of alliances and strategic workforce adaptation. Analysis of the 1MAO matrix (Figure 2) indicates that most actors endorse either the majority or entirety of these objectives, except Primary Health Care Clusters (ACES), which omits “increase Home and Autonomy Promotion Units (UDPA)” from its strategic planning. Differences in prioritisation illustrate strategic differentiation, reinforcing the urgency of AI-driven workforce reskilling.

Analysis of actor-specific strategic actions through the 2MAO matrix highlights varying degrees of involvement and organisational repositioning, reflecting shifts in institutional roles. As AI continues to reshape leadership strategies, organisations must reassess collaborative and competitive frameworks to adapt to emerging governance dynamics.

Interview data reveals strong interdependencies among actors, particularly in AI-driven workforce transformation. Aside from the hospital, local health unit and IPSS, most rely on national network coordination and regional coordination teams (ECR) for political and technical support in strategy development. AI’s integration heightens the need for adaptive leadership and organisational restructuring. The Regional Health Administration of Alentejo (ARSA) plays a central role in supporting strategic objectives, especially in managing AI-enabled workforce reskilling. Local Health Units (ULSs) depend on National Network Coordination (CNRede), while the hospital relies on ARSA to align with ICC objectives. As AI reshapes workforce strategies, CNRede is increasingly reliant on regional and local coordination teams (ECL) for organisational adaptation. All actors depend on the central government, particularly the health and social solidarity ministries, for ICC restructuring, with political will deemed critical for guiding AI-driven reskilling and leadership strategies.

Findings reveal strong interdependence among actors, shaping their strategic positioning. AI-driven workforce transformations further redefine these dynamics, requiring adaptive leadership and targeted reskilling initiatives. Power relations are examined using the DIM and the Direct and DIIM. While DIM assesses direct actor interactions, DIIM integrates AI-mediated influences, reflecting shifts in organisational structures. Influence levels range from 0 (no influence) to 4 (very strong influence, challenging actor existence). The DIM matrix (Figure 3) offers an initial evaluation, distinguishing dominant and dependent actors within the ICC system.

However, direct influence alone fails to encompass AI-related interdependencies. The DIIM (Table 3) reveals how AI affects leadership strategies and reskilling efforts by incorporating intermediary forces. This demonstrates AI’s capacity to disrupt traditional hierarchies and create networked dependencies, requiring new workforce engagement models. The row sum (li) denotes overall influence, while the column sum (Di) represents actor dependency, offering a clearer picture of inter-actor dynamics.

The DIIM matrix analysis highlights Integrated Continued Care Teams (EICC), ARSA, IPSS, Integrated Continued Care Units (UICC) and Local Coordination Teams (ECL) as the most influential actors, while CNRede, District Centres of Social Security (CDISS), ARSA and the ECR show high dependency. These relationships illustrate AI’s disruption of hierarchical structures, requiring new leadership strategies and workforce reskilling. The Influence and Dependence Plan (Figure 4) evaluates motricity and dependence, shedding light on AI-driven workforce transformations.

Dominant actors, such as IPSS, UICC, EICC and Primary Health Care Clusters (ACES), display strong influence with low dependence, with IPSS being especially autonomous. Linking actors, including ARSA, ECL, Hospital Discharge Management Teams (EGAs) and CNRede, show both high influence and high dependence, positioning them at the heart of AI integration and workforce transition. In contrast, actors such as CDISS, the Hospital and ECR show high dependence but low influence, illustrating AI’s disruptive consequences for workforce capacity. ULS, labelled as an independent actor, registers low influence and low dependence, indicating limited AI engagement within ICC structures. Several insights emerge from the influence–dependence analysis. Hinge actors, ARSA, ECL, EGA and CNRede are influential yet deeply dependent, constrained by AI-driven structural shifts. Key actors such as IPSS, UICC, EICC and ACES retain leadership positions while remaining relatively unaffected by AI-led workforce transitions. However, the presence of actors performing both dominant and subordinate roles suggests systemic instability, reinforcing the urgency of strategic AI integration to enhance organisational resilience. Power vectors derived from the DIIM matrix (Ri) quantify each actor’s capacity to exert or absorb influence. As AI reshapes workforce structures, leadership and training strategies must evolve to preserve organisational equilibrium. DIIM power analysis (Figure 5) identifies IPSS, ECL, EGA and UICC as central influencers, with ACES, ECR and the hospital acting as additional structural drivers. Meanwhile, EICC, ARSA, CDISS, CNRede and ULS show limited strategic assertion, revealing vulnerabilities in their ability to manage AI-driven transformations.

Net distance analysis highlights ARSA’s strategic influence, underscoring its reliance on external actors for ICC policy implementation. This dependency reflects AI’s role in redefining inter-organisational coordination and the growing necessity for data-informed decision-making in complex governance systems. Though EICC exhibits low assertiveness, it remains vital in achieving strategic goals, particularly in domiciliary care and community-based responses, two areas deeply impacted by workforce transformation. The MACTOR methodology enables a deeper examination of the net balance of influence, shedding light on AI-related shifts in workforce alignment. Actors with stronger influence than they receive include IPSS (93), EICC (66), UICC (62), ULS (29), ECL (22), ACES (18) and EGA (8). Conversely, actors with negative balances show heightened exposure to external pressures, highlighting the need for adaptive workforce policies. Despite its influence, ARSA remains heavily dependent, requiring structural realignment through AI integration. EICC, in spite of limited power projection, is essential in managing reskilling pathways and community care initiatives. Ultimately, Alentejo’s ICC landscape will be shaped by actors exerting more influence than they absorb, though this does not entirely align with those identified as linking actors (e.g. ARSA and CNRede). The transformative impact of AI on workforce leadership necessitates ongoing strategic recalibration to ensure institutional resilience amidst technological disruption. Building on actor census data and systemic evaluation, the study also analysed strategic alignment using AI-driven adaptation models, contributing to longer-term workforce stability.

Understanding actors’ alignment with strategic objectives is crucial, particularly as AI reshapes workforce structures, reskilling needs and leadership models. ICC development relies on how actors position themselves within shifting organisational settings. Interview data quantify this engagement using two matrices: 1 MAO (simple positions) and 2 MAO (valued positions). The 1 MAO applies a scale of 1 (agreement), −1 (disagreement) and 0 (neutral). The 2-MAO evaluates the relevance of objectives within AI-enhanced ICC, using a 0–4 scale, where 0 signals low relevance and 4 denotes objectives vital to the actor’s existence. AI-driven strategic planning compels actors to reconsider involvement based on agility and skill readiness. While the 2 MAO identifies directly engaged actors, the 3 MAO matrix (Godet et al., 2000) assesses mobilisation and influence. Summing 2 MAO and 3 MAO values quantifies actor involvement and strategic weight. Involvement is drawn from 2 MAO rows; mobilisation, from 3 MAO, reflects AI’s impact on organisational strategy. Analysis of the 2 MAO (Figure 6) highlights CNRede and CDISS as the most engaged, followed by ECR and ARSA, emphasising their strategic focus within ICC development.

AI’s influence on workforce structures requires reassessing mobilisation dynamics (Figure 7). IPSS and EICC, in spite of high mobilisation capacity, exhibit lower direct involvement, reflecting AI-driven shifts in workforce agility and leadership strategies. UICC, with high mobilisation and moderate involvement, underscores AI’s impact on organisational adaptability. ARSA remains highly engaged but ranks fourth in mobilisation. Meanwhile, the hospital is the least involved and mobilised, indicating minimal AI integration within its strategic framework.

The valued positions matrix ranks strategic objectives based on actor convergence, underscoring the central role of AI in fostering intersectoral collaboration and workforce reskilling. Among the highest-priority goals are enhancing cooperation between the health and social security sectors (Objective 15) and reinforcing palliative care responses within ICC (Objective 9). Other key objectives include strengthening domiciliary and community-based services (Objective 1), improving network access and user mobility (Objective 6), advancing intersectoral coordination (Objective 10), ensuring sustained political support for ICC policy (Objective 11) and investing in workforce training and qualification (Objective 14). These reflect an emerging strategic focus on AI’s influence in reshaping organisational models, professional skillsets and leadership. In contrast, lower-priority objectives include expanding paediatric care (Objective 8), increasing home and UDPA (Objective 4), improving coordination between EICC and home support services (Objective 3), increasing care vacancies (Objective 5), qualifying interventions by network actors (Objective 12) and prioritising responses in mental health (Objective 7). These trends highlight a shift towards objectives aligned with systemic innovation, capacity development and adaptive governance. No conflicting objectives were identified, but two clusters emerged:

1. Little mobilising consensuses (low-impact but uncontroversial); and

2. Mobilising consensuses (widely endorsed, high-priority goals).

Figure 8 confirms convergence around strategic priorities, while reduced mobilisation in paediatrics and UDPA suggests limited future investment. AI integration thus reinforces innovation and strategic leadership over conventional service expansion.

The existence of a unified group of objectives indicates a strong likelihood of generalised mobilising consensus, establishing the basis for coordinated strategic actions. Because this group includes nearly all objectives, it is reasonable to expect actors to advance collectively, shaping their strategic direction through AI-driven workforce transformations and organisational restructuring.

The convergence of actors with strategic objectives highlights AI’s transformative effect on workforce dynamics, leadership strategies and reskilling imperatives. The matrix of weighted valued convergences of actors on objectives (3CAA) (Figure 9) quantifies these alignments, promoting AI-driven organisational cooperation. High convergence values signify shared priorities, fostering partnerships and alliances that optimise leadership strategies and enhance strategic mobilisation across sectors.

The 3CAA matrix (Table 4) highlights actor convergence in relation to AI-driven workforce transformation. The strongest convergences among UICC, EICC, IPSS, ARSA, ECL and EGA highlight shared priorities in reskilling, leadership strategies and organisational restructuring, fostering new partnerships that reinforce AI’s role in workforce transformation and strategic collaboration.

The analysis of inter-actor convergence underscores AI’s transformative influence on organisational structures, workforce dynamics and leadership strategies. The strongest convergences occur between UICC and EICC (74.7) and IPSS with UICC and EICC (74.1), highlighting shared priorities in workforce reskilling and adaptation. In contrast, the hospital and CDISS show limited convergence, reflecting slower AI integration. ARSA closely aligns with EICC, UICC and IPSS, while ECL demonstrates strong convergence with EICC and UICC, reinforcing AI-driven workforce optimisation and leadership transformation. EICC plays a central role in strategic alignment, particularly with IPSS and UICC, driving AI-enhanced organisational restructuring, workforce reskilling and leadership evolution within institutional frameworks (Figure 10).

Figure 11 illustrates the strongest 10% of convergent ties among EICC, IPSS and UICC, emphasising AI-driven workforce transformation, reskilling strategies and organisational restructuring, while highlighting the increasing alignment of actors in adapting to evolving leadership models.

The findings emphasise actor roles in AI-driven workforce transformation within Alentejo’s ICC system. Figure 12 shows UICC’s strong convergence with ARSA and ECL, reinforcing AI’s role in reskilling and leadership adaptation. EICC and ARSA emerge as key convergent actors, fostering AI-supported organisational collaboration, while net distance visualises workforce alliances.

The ICC system presents three key characteristics: a diverse array of collective actors requiring AI-driven leadership strategies, an unstable relational structure without clear dominance or subordination, highlighting AI’s role in decentralised workforce models and strong actor convergence fostering strategic cooperation. This alignment strengthens AI’s influence on workforce restructuring, leadership development and organisational adaptation, ensuring future-ready collaboration within Alentejo’s ICC framework.

This strategic analysis examines the principal actors involved in ICC governance within the Alentejo region and the mechanisms shaping their strategic behaviours. Findings indicate that actor strategies are largely influenced by network positioning and the power dynamics inherent in interdependent relationships (Guise et al., 2024).

The application of the MACTOR methodology has proven instrumental in characterising actor behaviour, providing a structured approach to mapping strategic objectives, identifying areas of convergence and assessing alliance-building potential (Ryan et al., 2024). The ARSA has emerged as a pivotal entity in resource management and performance auditing, collaborating with UICC, EICC, ECL and IPSS. Collectively, these actors ensure population needs are met while facilitating care articulation. Power analysis identifies ARSA, the CNRede and the ECR as both influential and dependent actors. Their centrality within the governance framework represents both an asset and a constraint, reinforcing systemic interdependence (Bacelar-Silva et al., 2022). Within the NICCN, power is exercised through actors’ ability to shape strategic exchanges and influence policy operationalisation (Martin, 2012). Diverging strategic priorities were observed. At the macro level, sustained collaboration between the Ministries of Health and Social Security remains essential for the long-term sustainability of the NICCN, aligning with Costa and Mourão’s (2015) analysis. At the micro level, reinforcing coordination across typologies, care levels and sectors is critical, particularly in expanding domiciliary care, enhancing palliative care provision and improving accessibility and mobility, areas identified as systemic gaps (Rainho et al., 2021).

Additionally, training and capacity-building for actors are essential for enhancing system resilience, with prior literature demonstrating the centrality of human resource governance in health system efficiency (Witter et al., 2020). Qin et al. (2023) further emphasise talent retention and workforce planning as key policy levers. This analysis must also be considered within the broader discourse on AI in workforce transformation, an increasingly relevant theme in integrated care policy (Cresswell et al., 2020). AI functions not merely as a technological advancement but as a transformative force reshaping workforce structures, organisational models and leadership paradigms (Kaplan and Haenlein, 2020). Within ICC ecosystems, AI introduces both challenges and opportunities (Henzler et al., 2025), particularly in decentralised decision-making processes through predictive analytics, automated triage and optimised resource allocation (Kuziemski and Misuraca, 2020). In the NICCN, AI could empower traditionally dependent actors, such as ECL, by enhancing real-time planning, task delegation and service coordination (Khatri et al., 2023). However, these advances require agile leadership capable of integrating data-driven methodologies without undermining human-centred approaches (Meskó et al., 2018).

AI also reconfigures professional roles, requiring hybrid competencies combining clinical expertise with digital fluency (Cresswell et al., 2018). Given regional challenges such as demographic ageing and workforce skill mismatches, tailored reskilling strategies must be implemented (Cohen et al., 2020). Strategic Objective 14, concerning workforce training, must incorporate digital transformation readiness, ensuring AI enhances rather than disrupts service provision (Smith, 2021). From an institutional perspective, AI integration demands adaptive leadership (Lee and Cosgrove, 2024) capable of interpreting algorithmic insights and translating them into equitable, actionable policies (Roberts et al., 2021). IPSS and UICC, already involved in strategic planning, are well positioned to assume such leadership roles, provided they incorporate AI governance frameworks (Simon et al., 2024). These frameworks must promote ethical AI utilisation, data transparency and safeguards ensuring AI augments rather than replaces human-centred care (Arias Hernández and Rockembach, 2025).

Empirical findings highlight persistent fragmentation between health and social sectors, where structural and operational inefficiencies, particularly in service integration, resource allocation and informal caregiver inclusion, continue to hinder effective coordination, in spite of consensus on their significance (Kim and Kim, 2022). Literature has long advocated for functional and territorial integration (Barr et al., 2024; Correia de Matos et al., 2025). AI has the potential to bridge these gaps by harmonising data flows and facilitating cross-sectoral service planning (Secinaro et al., 2021). However, without foundational interoperability and shared standards, digital tools risk reinforcing existing silos (Levy et al., 2021). Disparities in strategic objective prioritisation highlight a lack of robust negotiation mechanisms, reinforcing Crozier and Friedberg’s (2000) assertion that sustainable collective action requires transforming individual agendas into shared commitments. ICC’s hybrid governance model, composed of public, private and third-sector actors, necessitates structured negotiation and participatory frameworks.

However, inclusivity remains undermined by the absence of formal engagement mechanisms for informal caregivers, aligning with Reynaud (2004), who argues that collective negotiation legitimises strategic planning, as further emphasised by Belrhiti et al. (2024), who advocate for actor inclusion in policy design and implementation. Despite consensus around key objectives, effective coordination remains contingent upon resource availability, particularly financial sustainability, a challenge highlighted by Clauss and Ritala (2023). System instability is evident as core actors, including ARSA, ECL and CNRede, oscillate between leadership and dependency, reflecting a dynamic balance of power. ICC’s future hinges on its strategic direction, alliance-building capacity and adaptability to demographic and digital transformations. Proponents of domiciliary care continue to resist re-institutionalisation, reinforcing Clarkson et al.’s (2018) position that care models must align with both user needs and system efficiency. Ultimately, consolidating ICC in Alentejo necessitates enhanced collaboration between the Health and Social Security sectors, investment in interoperable digital infrastructures, and regionally adapted reskilling policies prioritising AI readiness. These measures align with systemic integration models (Huang et al., 2020), advocating for participatory governance, collaborative leadership and technological adaptation, ensuring AI serves as a strategic enabler for sustainable, coordinated care (Moore et al., 2023).

This study investigated the NICCN in the Alentejo region, examining how continued care can be directed towards collectively desirable futures and the strategic role of key stakeholders in this process. The findings confirm that ICC operates as a dynamic system of collective action, where actors assume interdependent strategic roles in shaping and executing care policies. The analysis underscores the importance of effective articulation across governance levels and collaboration between actors in strengthening the NICCN. Key stakeholders, including the ARSA, EICC, UICC, ECL and IPSS, are pivotal, not only in implementing policy but also in forging alliances that support network cohesion.

The integration of AI into ICC systems carries significant implications for organisational configurations and workforce composition (Henzler et al., 2025). Our findings indicate that AI adoption demands thorough reskilling throughout the care continuum while introducing strategic pressures for more collaborative and adaptive governance models (Price et al., 2023). As algorithmic tools increasingly guide clinical and managerial decisions, shifts in power structures within the NICCN are likely (Davenport and Kalakota, 2019). The study also highlights the necessity of advancing integrated care models that prioritise domiciliary services, bolster community-based interventions and expand access to palliative and mental health care (Sánchez-Cárdenas et al., 2021). Future leadership will need to balance technological advancement with human-centred values, particularly by involving informal carers and local support networks in AI-enhanced care environments (Larsson et al., 2025).

This research contributes theoretically by framing the NICCN as a collective action system shaped by strategic interdependencies among governance actors. Strategic action theory clarifies how demographic vulnerability and technological transition influence ICC decision-making. The MACTOR method enables the mapping of power asymmetries and stakeholder alignment, informing debates on AI-driven change in health-care governance. Practically, the findings highlight the need to improve coordination across governance levels and integrate resources more effectively. Effective AI implementation requires workforce reskilling and leadership approaches that align innovation with ethical care values. Key actions include strengthening domiciliary and outpatient services, enhancing digital training and supporting informal caregiver involvement. Ethical oversight is vital to ensure fairness in algorithmic decisions, especially for vulnerable populations. Addressing both physical and digital access barriers is also critical to reduce regional inequalities in care delivery.

This study presents some limitations that should be acknowledged. First, its cross-sectional design limits the ability to capture the evolving and dynamic influence of AI on integrated care systems over time. Second, the analysis focuses exclusively on institutional actors, excluding the perspectives of users and informal caregivers, which could provide additional insights into care dynamics. Third, the reliance on self-reported data may introduce response bias, as actors’ positions reflect subjective perceptions and strategic interests. In addition, the MACTOR method relies on assumptions of actor rationality, which may not fully capture informal or emergent behaviours within complex governance settings. Data collection may also be constrained by the willingness of institutional actors to disclose strategic positions, particularly in politically or organisationally sensitive contexts. Furthermore, the reliance on expert elicitation introduces a degree of subjectivity in the assessment of influence and positioning. Finally, the regional scope of the study may constrain the broader applicability and generalisability of the findings. Future research should integrate user experiences, enable comparative analyses across different regional contexts and examine emerging policy models. In addition, longitudinal studies are needed to explore the evolving impact of AI on health-care structures and workforce dynamics.