Embedding organisational learning in police intelligence: resolving the Learning–Control paradox

Contemporary policing operates in a landscape shaped by transnational criminal networks, rapid technological change, and continually shifting offending patterns (Ratcliffe, 2025). Police intelligence agencies, tasked with anticipating and disrupting these threats, depend on adaptive capabilities grounded in sustained organisational learning (OL) and knowledge development (Walsh, 2020). Yet many police organisations continue to reflect bureaucratic rigidity and resistance to change, limiting the reflective and flexible learning required for effectiveness (Filstad and Gottschalk, 2011). This tension creates a learning–control paradox: structures designed to uphold accountability, hierarchy and standardisation simultaneously restrict experimentation, feedback and knowledge sharing (Ritchie, 2010). In intelligence environments, where information is both strategically valuable and highly sensitive, this paradox is especially pronounced. Errors carry significant operational risk, and organisational cultures frequently privilege individual accountability over collective learning.

These pressures are intensified in South Africa. With a murder rate of 45 per 100,000, among the highest globally (South Africa murder rate surges, 2025), the South African Police Service (SAPS) Crime Intelligence (CI) Division faces severe performance expectations. Cities such as Johannesburg and Pretoria remain unsafe (Numbeo, 2025), and despite reform efforts, public trust continues to decline, with 43.3% of citizens adjusting daily routines to avoid crime (Tshikalange, 2025). References to “unsafe cities” denote residents' exposure to crime and fear of victimisation in urban areas, while “declining public trust” refers specifically to diminishing confidence in the SAPS, as evidenced by national survey data showing persistently low trust levels and declining perceptions of safety (Statistics South Africa, 2025; Human Sciences Research Council, 2025).

Within this demanding environment, OL represents a critical yet underutilised organisational capability. While recent research highlights the value of workplace learning and cultural renewal in improving policing outcomes (Mahamba et al., 2022), limited empirical work examines how learning shapes police intelligence performance. This study addresses that gap through four guiding questions: (1) How do intelligence practitioners conceptualise OL in their operational context? (2) Which organisational and cultural factors support or inhibit OL? (3) In what ways does OL influence operational effectiveness and crime reduction capacity? (4) What practical mechanisms can embed learning within intelligence operations?

Organisational learning refers to how organisations create, store and share knowledge to remain effective (Argote, 2013). In policing, OL is essential as agencies confront shifting crime patterns, rising public expectations and rapid technological change (Filstad and Gottschalk, 2010). Building on foundational organisational learning theory (Argyris and Schön, 1978; Senge, 2006), contemporary policing scholarship views learning as an ongoing, multilevel process involving reflection, feedback, adaptation and institutional memory, rather than training alone (Anand and Brix, 2022). However, hierarchical and performance-driven accountability regimes often constrain learning by encouraging risk aversion and surface compliance over critical reflection (Fan, 2024). International research highlights OL's growing relevance: Fors and Hansen (2014) show that OL improves strategic decision-making in Swedish organised crime investigations, while Filstad and Gottschalk (2010) argue that strong investigative work depends on continuous skill development. Wathne (2012) and Ritchie (2010) likewise document OL becoming embedded in some policing structures. Yet tensions persist between learning-oriented approaches and traditional police cultures. Filstad and Gottschalk (2011) note that bureaucratic, command-driven environments often inhibit the creativity and knowledge sharing OL requires, a concern echoed in more recent policing research that frames learning as politically and institutionally constrained by accountability pressures and legitimacy concerns (Bristow et al., 2021). Nonetheless, other security institutions demonstrate that transformation is possible: the British Army's efforts to embed OL (Catignani, 2014) and Dubai Police's structured, trust-based initiatives (Rowley et al., 2012) illustrate this potential, aligning with contemporary work that emphasises the importance of relational and cultural conditions for learning in high-risk organisations (Edmondson and Bransby, 2023).

Outcomes in policing remain uneven. Sugarman (2010) found OL improved performance in the NYPD, while Antony et al. (2018) showed reflective learning supported renewal in Scotland. Still, cultural and structural constraints remain significant (Filstad and Gottschalk, 2011; Ritchie, 2010). Comparative examples, from Japan, the UK, Canada, Denmark and Belgium, show context-specific but encouraging progress shaped by leadership and organisational incentives. Contemporary evidence-based policing research supports this interpretation, demonstrating that data and research only translate into improvement where organisations possess sufficient learning capacity to interpret, adapt and institutionalise evidence (Bedford and Neyroud, 2021).

Contemporary policing scholarship emphasises that organisational learning extends beyond formal training to include reflection, feedback and the institutionalisation of experience (Staller and Koerner, 2022). Yet learning in policing is often constrained by hierarchical authority, compliance-driven accountability and cultures that discourage error disclosure (Tomkins and Bristow, 2024). These tensions render police learning inherently paradoxical, as mechanisms designed to protect organisational legitimacy may inhibit openness and adaptation. This aligns with broader organisational learning research linking learning to resilience in volatile environments (Evenseth et al., 2022). Professionalising policing, therefore, requires balancing tacit “craft” knowledge with formalised, evidence-based learning to avoid undermining professional engagement and practical impact (Engelmann and Tatnell, 2025; Bedford and Neyroud, 2021).

Intelligence agencies now operate in environments marked by rapid change, uncertainty and growing complexity. These conditions place a premium on adaptability, making OL an essential capability for interpreting emerging threats, updating practices and maintaining operational effectiveness. In policing, this shift is reflected in the widespread adoption of intelligence-led policing (ILP), which positions intelligence at the centre of strategic and operational decision-making (Ratcliffe, 2025). As Lindsay et al. (2022) and Alblooshi and Kassim (2022) note, ILP enhances crime-prevention efforts by grounding interventions in evidence and analysis. Jurisdictions worldwide have reorganised policing around ILP principles in response to increasingly sophisticated and transnational criminal activity. Kostadinovic and Klisaric (2017) as well as Alblooshi and Kassim (2022) describe ILP as a coordinated system of standards and analytical tools aimed at continuous improvement, while Pournara (2020) and Boyko (2024) frame it as a managerial philosophy centred on data-driven, targeted responses.

International practice indicates that OL thrives where collaboration and structured knowledge exchange exist. In the United States, Crosston (2018) and Usowski (2023) highlight strong partnerships between academia and intelligence agencies, fostering reflective practice and continuous learning. The United Kingdom has developed a similar ecosystem through institutions such as the British International Studies Association, the Oxford Intelligence Group, Brunel University and the University of Buckingham (Michael and Kornbluth, 2019; Glees, 2018; Hughes, 2019; The Oxford Intelligence Group, 2023; Brunel University London, 2023; University of Buckingham, 2023). European countries have also built formal pathways for intelligence professionalisation: France's CF2R and the École Française du Renseignement (Dénéce and Arboit, 2010, p. 23), Spain's university-linked “culture of intelligence” (Matey, 2010, p. 748; Michael and Kornbluth, 2019, p. 117; Mendosa, 2022), and Germany's specialised programme at Bundeswehr University (Borghoff et al., 2016, p. 79; National Intelligence University, 2023).

Comparable models worldwide, including the UK's National Intelligence Model, New Zealand's ILP framework, the U.S. Sherman Kent School (Carter and Carter, 2009; Lindsay et al., 2022; Mendosa, 2022; Office of the Director of National Intelligence, 2023), Slovakia's National Criminal Intelligence Model and Israel's Havatzalot Program (Ahronheim, 2019; Usowski, 2023), demonstrate how sustained investment in OL strengthens strategic capability, adaptability and organisational resilience. Yet ILP's effectiveness depends fundamentally on underlying OL processes enabling continuous intelligence refinement. As Walsh (2020) and the Office of the Director of National Intelligence (2023) argue, intelligence leadership and governance must prioritise building effective intelligence communities through sustained learning and development. Without robust OL mechanisms, ILP risks becoming a superficial compliance exercise rather than a genuine capability enhancement.

ILP has become a central strategic framework, placing intelligence at the heart of decision-making and crime prevention (Ratcliffe, 2025). However, implementation research shows that formal accountability pressures can displace adaptive learning, particularly where practitioners face competing legal, professional and political demands (Thomann et al., 2023). Although ILP assumes iterative feedback between intelligence and operations, studies suggest these learning cycles often function unevenly in practice. Organisational silos, weak review processes and limited learning capacity frequently undermine implementation, leaving intelligence products underutilised (Lindsay et al., 2022). Research on transnational crime further indicates that ILP effectiveness depends on continuous organisational adaptation and learning-oriented cultures, rather than static analytical models (Alblooshi and Kassim, 2022; Phythian and Kirby, 2024).

Knowledge management (KM) has gained renewed prominence as policing organisations contend with skills attrition, technological complexity and fragmented information systems. Research consistently shows that police knowledge is largely tacit and experiential, making purely technological KM solutions insufficient on their own (Birdi et al., 2020). Effective KM is instead shaped by leadership support, open communication and alignment between knowledge practices and operational roles (Al Dhanhani and Al Naqbi, 2022). In contrast, siloed structures, mistrust and performance-driven cultures frequently inhibit knowledge sharing, particularly within specialist and intelligence units. Cross-jurisdictional studies further demonstrate that organisational boundaries and incompatible systems constrain knowledge transfer even where crime problems are shared (Birdi et al., 2020). Recent scholarship increasingly links KM to organisational learning, arguing that knowledge only generates value when it is interpreted, debated and applied in practice (Staller and Koerner, 2022; Bedford and Neyroud, 2021). Without such integration, policing organisations risk accumulating information without developing insight, institutional memory or sustained capability.

Organisational culture plays a central role in shaping how learning occurs within police institutions (Schein, 2010; Edmondson and Bransby, 2023). Police culture comprises both cognitive and emotional elements that influence norms, behaviours and everyday practices. It includes formal structures such as policies and training, but also informal attitudes toward suspects, the public and internal conduct (Crank, 2015; Staller and Koerner, 2022). These cultural patterns reflect espoused values that help build identity and cohesion, yet they can also hinder OL. As Filstad and Gottschalk (2011), Bristow et al. (2021), and Tomkins and Bristow (2024) argue, values act as guiding principles within policing. Their study identifies eight commonly espoused values: change, informality, empowerment, relation-orientation, directness, openness, knowledge orientation and stability, but only informality and empowerment show a positive relationship with OL (Filstad and Gottschalk, 2011; Engelmann and Tatnell, 2025). This indicates that despite broad value statements, few genuinely support learning. For organisations like the SAPS CI Division, developing into learning organisations requires reassessing cultural foundations and strengthening openness, shared decision-making and collaborative learning (Edmondson and Bransby, 2023; Evenseth et al., 2022).

Contemporary research shows that accountability regimes shape organisational culture by signalling what is safe to disclose, discuss or withhold. Where accountability systems are perceived as punitive or politically unstable, police personnel tend to limit transparency and constrain learning conversations, even when formal learning policies exist (Engelmann and Tatnell, 2025; Tomkins and Bristow, 2024). As a result, learning is unevenly enacted, and reform efforts often stall, producing the appearance of change while entrenching existing practices (Fan, 2024).

ampere qualitative case study design (Yin, 2018) was conducted to explore OL implementation within the SAPS CI Division in two major metropolitan areas within the country. This approach enabled rich, contextualised understanding of complex organisational phenomena within real-world settings (Creswell and Poth, 2018). This study drew on qualitative data to examine how OL practices were interpreted and enacted within the hierarchical and operationally sensitive environment of the SAPS CI Division.

The SAPS CI Division plays a central role in South Africa's crime detection, prevention and law-enforcement landscape. Its mandate, set out in the Constitution (South Africa, 1996) and the National Strategic Act, 39 of 1994 (South Africa, 1994), spans both analytical and operational intelligence functions. The division gathers, evaluates and interprets intelligence to prevent crime and support investigations (South African Police Service Annual Report, 2024). Operating in a high-crime environment, it faces significant resource and governance pressures. Its specialised national and provincial components (South Africa, 2015) make CI a valuable context for studying organisational learning.

Participants were selected using purposive sampling to ensure the inclusion of information-rich cases. In this study, 30 members of the SAPS Crime Intelligence (CI) Division were purposively selected based on role relevance, operational experience and sustained exposure to OL processes related to crime reduction. Participants represented a wide range of CI functions, including analytical, operational, technological, supervisory and training roles. All had a minimum of three years' experience within CI, ensuring sufficient organisational insight. This purposive sampling approach provided both functional and hierarchical diversity, enabling a holistic understanding of learning practices within the CI environment. Sampling continued until data saturation was reached.

The first author conducted 30 semi-structured interviews between 2023 and 2024, each lasting 45–60 min and recorded and transcribed verbatim. The protocol covered five themes, including OL conceptualisation, learning culture, performance impact, KM practices and alignment with policing strategies. Interviews were held privately to ensure confidentiality. As a serving member of a national law-enforcement intelligence unit, the first author has built trust while maintaining analytical distance through reflexive journaling and peer debriefing. Ethical approval was granted by the relevant university ethics committee and SAPS, with prior informed consent obtained from all participants after they were fully informed of the study's purpose, the protection of their anonymity, the secure storage and use of their data, and the fact that no risks were associated with their participation.

Interview data were analysed using Braun and Clarke's (2021) six-phase reflexive thematic analysis, involving familiarisation, systematic coding, theme construction, review, definition, and reporting. Following reflexive thematic analysis principles, themes were developed and presented based on their analytical significance in illuminating patterned meanings across the dataset and their relevance to the research questions. Following Watkins and Gioia (2015), coding was conducted iteratively. Transcripts were analysed through iterative coding to identify meaningful patterns across the dataset. Codes were clustered and refined into themes that captured shared participant meanings while remaining grounded in the data. Analytical rigour was supported through reflexive memo-writing, clear audit trails, and the use of illustrative data extracts.

The following section discusses the participants' conceptualisations and experiences of OL and its implications for operational effectiveness and crime reduction. Five interconnected themes are presented, illuminating OL conceptualisation, implementation, and impact within SAPS CI. These themes demonstrate both opportunities and substantial challenges in embedding learning within hierarchical law enforcement contexts. Each of the nine themes is presented with exemplary verbatim quotes from participants.

Participants described OL in varied ways, reflecting both informal and formal understandings. Many emphasised practical, experience-based learning as the foundation of professional development. As one participant (CIS2) explained, “The most experienced in that particular field shares the knowledge with the new people. Never mind the ranks.” This framing positions OL as peer-driven, mentorship-based and embedded in everyday operational practice. Others articulated more institutionalised conceptions. CIAS1 described OL as “gathering knowledge and providing that knowledge back in the form of giving courses,” while CIAS4 defined it as “creating, acquiring, and transferring knowledge in the working environment.” CIS5 adopted a more goal-oriented view, describing OL as organisational orientation and training intended to empower members with the skills needed to achieve objectives. Despite these differences, participants broadly agreed that learning is essential for responding to rapidly evolving criminal tactics and technologies.

Mentorship consistently emerged as a key mechanism linking formal and informal learning. CIG3 noted that new members are “first being mentored … to see that, really, this person can cope,” while CIS3 described OL as encompassing “knowledge share, skills transfer, and mentorship.” At the same time, participants depicted a fragmented learning culture. Although some acknowledged the existence of training programmes, others felt these were insufficient. CIG4 argued that existing knowledge transfer was “not enough” given the scale and complexity of crime. This gap between organisational intent and lived experience was recurring. While some participants pointed to available courses and training budgets, others highlighted practical shortcomings, including outdated materials and training that occurred only after operational mistakes had already been made.

Informal learning practices frequently compensated for these gaps. CIS3 observed that experienced members routinely mentor newcomers, while CIA1 stressed the need for internal training to ensure operational competence. However, this reliance on informal learning reflected necessity rather than deliberate institutional design. Structural factors further constrained OL, including discretionary access to training, perceptions of favouritism, and broader issues of mistrust, political interference and weak accountability. Overall, the findings show that OL within SAPS CI is understood through multiple, sometimes competing lenses, yet is widely recognised as critical to operational effectiveness.

Participants described marked unevenness in how OL is implemented across units, with access to formal training, mentoring and skills auditing shaped by discretion, rank and location. Many expressed frustrations at opportunities that rarely materialised in practice. CIA1 noted that although advanced courses exist, “you never see that training,” while CIA2 reported not attending a single course over nine years. Training access was often described as tiered. CIG12 explained that while all members receive basic training, intermediate and advanced development depends largely on individual initiative, a view echoed by CIS5, who argued that the division remains “content with just providing the basic orientation.” By contrast, CIPT1 described more consistent exposure to training, underscoring uneven implementation across units.

Despite dissatisfaction with formal training, mentorship was widely valued. CIG3 emphasised that new members are mentored to help them adapt to operational realities, while CIG26 described structured mentoring and coaching in some units. However, mentoring practices were not consistently formalised. CIG5 described a largely informal system in which individuals must actively seek out experienced colleagues willing to take them “under their wing.”

Structural and cultural barriers further constrained OL. CIS3 characterised CI as lacking a shared vision, while concerns about knowledge retention were raised by CIG8, who noted that skilled members often leave for the private sector without transferring expertise. CIS5 added that unit commanders, although formally responsible for mentoring, are frequently overwhelmed by operational demands, limiting meaningful skills transfer.

Experiences of skills auditing were equally inconsistent. While some units reported annual audits, others questioned their value or had never encountered them. Audits were often described as procedural exercises with little follow-through, and concerns about politicisation and favouritism in training and placement decisions were common. Collectively, these accounts point to a persistent gap between formal OL systems and the learning that occurs at the operational level.

Participants consistently linked OL to stronger operational capability and improved crime-reduction outcomes, although the strength of this relationship was shaped by structural constraints and leadership quality. Several participants expressed this link unequivocally. CIG6 stated that “a worker who is well-trained performs better,” directly associating learning with improved performance and crime reduction. Similarly, CIG8 noted that regular training helps members “learn new trends,” enabling more effective disruption of criminal syndicates.

Specialised knowledge was closely associated with high-risk operational work. CIG4 highlighted the relationship between competence and safety in drone operations, emphasising that errors carry serious consequences. Effective source management was also seen as central to operational success, with participants stressing that how informants are handled directly affects the quality and usefulness of intelligence. CIA1 framed competence in terms of systems literacy, arguing that broader technical knowledge enhances analytical capacity and responsiveness.

Reflective practice emerged as another important learning mechanism. CIPT1 described the value of learning from failure, suggesting that structured reflection strengthens future performance. Others linked OL to collective benefits, including improved teamwork, shared ownership and continuous improvement (CIG13). Briefings and debriefings were frequently cited as practical tools for embedding learning into routine operations, with CIS2 noting that these practices consistently enhance performance.

Participants also described behavioural changes associated with learning, such as improved judgement, deeper critical thinking and increased confidence. However, some questioned whether learning always translates into improved outcomes. CIA1 cautioned that training is not consistently applied, while CIS3 highlighted the demotivating effects of an absent shared vision. Structural constraints, including outdated technology, limited resources and the loss of skilled personnel to the private sector, were also raised. Despite these challenges, enthusiasm for learning remained strong, with many participants expressing a continued desire for further training and development.

Participants expressed mixed views on how effectively OL is embedded within CI's broader policing strategies, frequently highlighting a tension between the rhetoric of continuous learning and the operational constraints that limit it in practice. Some offered positive assessments. CIAS4 noted that CI “does follow an OL-based approach,” suggesting that learning supports crime reduction and can align with intelligence-led policing. Similarly, CIG3 believed OL was embedded through specialised training for CI operatives, while CIG4 linked learning directly to operational impact, emphasising CI's role in supporting investigations across units.

Others were more sceptical, pointing to implementation failures rather than resistance to learning itself. CIPM1 argued that shortcomings lie in execution, while CIG5 suggested that OL initiatives are often misaligned with shifting policing strategies. Concerns about strategic incoherence were common. As CIG8 observed, “You can't be having a strategy of going to point X, but your training goes to Y,” a view echoed by CIPT2. Participants warned that fragmented planning undermines the operational value of learning, even where training opportunities exist. CIS5 further cautioned that without sustained learning pathways beyond basic training, members struggle to keep pace with evolving crime challenges.

ampere deeper structural tension, described as the learning–control paradox, emerged across accounts. CIA1 explained that while technological learning is encouraged rhetorically, it is often constrained by tightly controlled procedures. CIG12 noted that learning remains largely reactive, despite increasingly sophisticated criminal activity. Others highlighted structural gaps, including the absence of dedicated research capacity. CIG1 suggested that access to researchers could help prevent recurring operational shortcomings.

Trust and information sharing also emerged as fault lines. Participants described how corruption, information leakage and interdepartmental competition limit collaboration and constrain learning. CIS1 further noted gaps between classroom learning and field application, while resource constraints restricted access to training. Collectively, these accounts underscore that OL's strategic value depends on alignment between learning priorities, operational realities and organisational direction, a condition not consistently achieved within SAPS CI.

Participants described a range of knowledge-sharing practices within CI, alongside significant barriers that limit their consistency and long-term impact. Several highlighted positive examples of deliberate knowledge transfer. CIPT1 explained that teams routinely share lessons after operations, noting that collective reflection improves productivity and efficiency. CIG9 expressed a strong personal commitment to sharing knowledge, while CIPT2 observed that joint operations between units, such as undercover and surveillance teams, facilitate the exchange and retention of tacit knowledge. These examples illustrate how learning often occurs through everyday collaboration rather than formal systems.

At the same time, participants pointed to substantial barriers. Structural constraints frequently limit cross-unit learning, with CIG6 noting that while knowledge creation does occur, movement between sections is restricted. Knowledge hoarding was a recurring concern, with CIS5 acknowledging that expertise is sometimes guarded due to rivalry or insecurity, resulting in siloed knowledge. CIG5 reflected on the long-term risks of over-reliance on individual experts, warning that dependence on a single source of knowledge can leave future leaders underprepared when those individuals leave the organisation.

Performance measurement systems attracted sustained criticism. Participants described a strong emphasis on output quantity over analytical quality, with evaluations often prioritising compliance rather than development. Several questioned whether current measures meaningfully capture learning or operational impact, noting that assessment tends to operate at senior levels with limited reflection of frontline realities. Despite these concerns, some participants reported tangible benefits from learning, including improved report writing, clearer articulation of intelligence and greater confidence in operational settings. Knowledge retention emerged as a critical vulnerability. Participants highlighted the absence of succession planning and the loss of experienced personnel to retirement or the private sector, taking valuable expertise with them. Calls for dedicated CI training infrastructure and stronger cultural support for learning underscored a shared view that effective knowledge management depends not only on formal systems, but on sustained investment in people, culture and accountability mechanisms that actively support learning.

This study shows that strengthening OL within police intelligence agencies requires more than incremental change; it demands a rethinking of how learning is understood and practised in hierarchical, risk-sensitive and accountability-driven environments. While personnel recognise OL as vital for operational effectiveness and crime reduction, entrenched barriers, including cultural resistance, fragmented structures, uneven training and the learning–control paradox, continue to limit its impact. To address these challenges, the study proposes a Strategic Organisational Learning Implementation Framework as a transformation strategy rather than a prescriptive checklist. Drawing on Senge's learning-organisation principles (Hansen et al., 2020), Argyris and Schön's double-loop learning theory (Kumar, 2024), and contemporary intelligence-led policing concepts, the framework embeds adaptive, reflective and evidence-based learning within command-based settings. Consistent with Senge (2006) and Schein (2010), the findings show that command logics and compliance cultures suppress reflection and error disclosure, constraining genuine organisational adaptation (Argyris and Schön, 1978; Molodchik and Jardon, 2015).

Consistent with OL theory (Argyris and Schön, 1978; Senge, 2006; Argote, 2013), most participants viewed learning as essential for responding to evolving criminal threats, technological change and increasing analytical demands. However, they differed in how they understood OL. Some associated it mainly with formal training and structured courses, while others emphasised informal, experience-based processes such as mentoring, peer learning and on-the-job exposure. This variation reflects what Crossan et al. (1999) describe as weakly shared mental models, where learning occurs without a common organisational understanding of its purpose or mechanisms. Rather than signalling resistance to learning, these differences suggest that OL within crime-intelligence environments is often situational and locally constructed. Units with supportive leadership, stable structures or established mentoring practices reported more consistent learning experiences, while others experienced learning as sporadic, delayed or reliant on individual initiative. Similar patterns are evident in policing research, where learning outcomes are shaped more by local culture, leadership and resourcing than by formal policy alone (Filstad and Gottschalk, 2011; Wathne, 2012).

The findings strongly support the presence of a learning–control paradox, whereby accountability mechanisms intended to manage risk simultaneously restrict the openness required for learning (Filstad and Gottschalk, 2011; Tomkins and Bristow, 2024). Importantly, this tension was experienced differently across organisational levels. Operational members emphasised fear of mistakes, low psychological safety and limited opportunities for experimentation, reflecting research linking hierarchical control to reduced error disclosure and reflection (Schein, 2010; Edmondson and Bransby, 2023). In contrast, some supervisory perspectives viewed control structures as unavoidable in intelligence environments characterised by secrecy and risk. These contrasting accounts highlight that learning constraints are shaped by positional realities within command-based organisations, reinforcing arguments that learning in policing is inherently political and influenced by power, legitimacy concerns and competing priorities (Bristow et al., 2021).

Leadership emerged as a decisive but uneven influence on organisational learning. Some participants described leaders who actively encouraged development, mentoring and reflection, while others pointed to leadership instability, politicised advancement and limited accountability for enabling learning. This variation supports Schein's (2010) argument that leaders shape what is valued, discussed and rewarded within organisational cultures. Rather than a single, uniform police culture, the findings suggest multiple subcultures that produce distinct learning climates. In some units, informal mentoring compensated for gaps in formal training; in others, high workloads, mistrust and resource constraints curtailed learning engagement. Similar patterns have been observed in comparative policing research, where learning cultures vary considerably across units within the same organisation (Ritchie, 2010; Antony et al., 2018).

Most participants linked OL to improved competence, confidence and operational effectiveness, echoing evidence-based policing research that associates learning capacity with better decision-making (Bedford and Neyroud, 2021). Participants described learning as enhancing judgement, analytical depth and responsiveness to emerging crime patterns, consistent with ILP assumptions that intelligence improves through iterative use (Ratcliffe, 2025). However, not all participants viewed this relationship as straightforward. A minority questioned whether learning consistently translates into improved performance, citing failures to apply training, outdated systems and misalignment between learning content and operational realities. These reservations reinforce scholarly cautions that OL outcomes are conditional rather than automatic, dependent on leadership support, relevance, reinforcement and organisational readiness (Molodchik and Jardon, 2015; Lindsay et al., 2022). This finding highlights an important distinction: OL is widely seen as necessary, but insufficient on its own to overcome structural and cultural constraints. Where learning is poorly integrated into practice, its impact on crime-reduction outcomes becomes attenuated, lending support to critiques that ILP risks becoming symbolic if not supported by strong OL processes (Walsh, 2020; Phythian and Kirby, 2024).

Participants widely recognised the importance of knowledge sharing and retention, yet described clear gaps between intent and practice. While some units engaged in effective briefing and debriefing, knowledge transfer was often informal and reliant on individual willingness rather than structured KM processes. This reflects research showing that police knowledge is largely tacit and vulnerable to loss when not institutionalised (Nonaka and Takeuchi, 1995; Birdi et al., 2020). Performance measurement systems further complicated learning outcomes, with participants describing metrics that prioritised output quantity over learning quality, reinforcing compliance rather than reflection. Similar critiques appear in policing scholarship (Staller and Koerner, 2022). Nonetheless, some participants reported noticeable improvements in report quality and analytical clarity following learning interventions, suggesting that OL benefits, while uneven, are evident.

The literature on organisational learning, policing culture, intelligence-led policing and accountability highlights a persistent tension between the formal requirements of control and the informal practices through which learning actually occurs. While existing studies emphasise the importance of learning for policing effectiveness, they also show that hierarchical accountability structures frequently inhibit reflection, knowledge sharing and adaptation. This synthesis positions the learning–control paradox as a central explanatory lens and underscores the need for an empirically grounded framework capable of translating learning principles into implementable organisational practices, particularly within intelligence environments.

The findings suggest that strengthening OL in crime-intelligence environments requires recognising variation as an empirical reality rather than an analytical weakness. The proposed Strategic Organisational Learning Implementation Framework should therefore be understood as a flexible architecture rather than a prescriptive solution. Its emphasis on shared understanding, embedded reflection and learning governance directly addresses the uneven learning climates identified in the data. However, consistent with OL theory, the framework's effectiveness will depend on leadership commitment, cultural reinforcement and alignment with operational constraints (Senge, 2006; Anand and Brix, 2022). By explicitly acknowledging divergent participant perspectives, this revised discussion strengthens the credibility of its conclusions and positions OL as a contested yet strategically indispensable capability within police crime-intelligence organisations.

In response to the challenges identified, OL is reconceptualised for command-driven intelligence environments and proposes a Strategic OL Implementation Framework designed to embed learning throughout the intelligence cycle. The framework does not originate from abstract theory alone but is grounded in the reflexive thematic analysis of participants' lived experiences of learning, control, and operational practice within SAPS CI. The framework shifts performance systems toward recognising learning contributions and relies on secure knowledge platforms to capture tacit expertise (Walsh, 2020; Nonaka and Takeuchi, 1995). It extends current debates on OL, ILP and police professionalisation by offering a structured approach suited to high-stakes, hierarchical settings. The framework comprises five components aimed at moving the SAPS CI Division from reactive, compliance-driven operations to a proactive learning organisation.

The Strategic OL Implementation Framework was developed inductively from the five interconnected themes identified in the findings and is intended to translate empirical insights into a coherent, operational approach for embedding learning within a hierarchical police intelligence environment. Each component reflects recurring patterns in participants' accounts, capturing both the constraints they experience and the learning practices they regard as effective. Rather than treating learning as a discrete training intervention, the framework conceptualises OL as an integrated organisational capability embedded across cultural, structural and operational domains, directly addressing the learning–control paradox identified in the study (Filstad and Gottschalk, 2011; Tomkins and Bristow, 2024).

Structurally, the framewor0k comprises five mutually reinforcing components, each aligned with a specific findings theme. Together, these components provide an organising architecture through which learning can be stabilised, coordinated and sustained within command-based intelligence settings, consistent with core learning-organisation principles (Argyris and Schön, 1978; Senge, 2006).

The first component, shared learning conceptualisation and culture, derives from Theme 1, which revealed fragmented and competing understandings of OL, ranging from formal training to informal mentorship. Participants widely valued learning but described it in inconsistent terms, resulting in uneven expectations and practices across CI units. This component establishes a shared organisational understanding of learning as encompassing reflection, feedback, knowledge exchange and adaptation, rather than training alone (Argote, 2013; Anand and Brix, 2022). In practice, this requires leadership-led articulation of learning principles, embedding OL language in policy and induction processes, and formally recognising mentoring as a legitimate learning mechanism, in line with cultural learning research in policing (Schein, 2010; Edmondson and Bransby, 2023).

The second component, structured implementation and access to learning, responds directly to Theme 2, which highlighted inconsistent training provision, discretionary access and uneven mentoring practices. Participants described advanced learning opportunities as delayed, unevenly distributed or dependent on managerial discretion. This component formalises learning pathways through transparent skills audits, equitable nomination processes and staged development from basic to advanced competencies. Operational responsibility is deliberately located at the middle-management level, where strategic intent intersects with everyday intelligence work, reflecting evidence that learning effectiveness depends on coherent organisational support rather than individual initiative alone (Filstad and Gottschalk, 2011; Wathne, 2012).

The third component, learning embedded in operational effectiveness and crime reduction, corresponds to Theme 3, where participants consistently linked learning to improved judgement, analytical quality and preventive capability. These benefits were most evident when learning was integrated into routine operational practices rather than treated as a separate activity. This component embeds learning within core processes such as pre-task briefings, post-operation debriefings and reflective intelligence reviews, aligning with intelligence-led policing assumptions that intelligence improves through iterative use and feedback (Ratcliffe, 2025; Bedford and Neyroud, 2021).

The fourth component, strategic alignment and learning governance, addresses Theme 4 and the learning–control paradox. Participants described persistent tensions between innovation-oriented learning and rigid accountability structures, as well as misalignment between learning initiatives and shifting policing strategies. These tensions were experienced as practical constraints on applying new knowledge rather than resistance to learning itself. This component re-orients governance mechanisms to align learning priorities with threat assessments and strategic intelligence objectives, without diluting hierarchical control. Instead, governance is reframed to enable reflection and institutional memory within accountability constraints (Ritchie, 2010; Bristow et al., 2021; Walsh, 2020).

The fifth component, knowledge management, measurement and retention, is grounded in Theme 5, which revealed significant weaknesses in knowledge retention, loss of tacit expertise and performance systems that prioritise output over learning quality. Participants repeatedly highlighted how knowledge dissipates through attrition, weak succession planning and compliance-focused evaluation practices. This component integrates KM with OL by formalising knowledge capture, succession-focused mentoring and performance indicators that recognise learning contributions. Such integration is critical given the tacit nature of police knowledge and the risks of accumulated information failing to generate insight (Nonaka and Takeuchi, 1995; Birdi et al., 2020; Staller and Koerner, 2022).

These five components form a thematically coherent, empirically grounded and operationally actionable framework. The Strategic OL Implementation Framework is not intended as a linear implementation model, but as a flexible architecture that can be introduced incrementally according to organisational readiness. By explicitly mapping each component to the empirical findings, the framework demonstrates clear derivation from the data and provides practitioners with a realistic pathway for embedding learning within hierarchical, risk-sensitive intelligence organisations.

While presented at a conceptual level, the framework is grounded in identifiable organisational practices. Responsibility for implementation would primarily reside at the middle-management level, where unit commanders and analytical supervisors translate learning principles into routine practices such as structured briefings, mentoring arrangements and cross-functional intelligence exchanges. The framework further assumes learning to be iterative rather than episodic, requiring learning cycles to be embedded across the intelligence process from tasking and collection to analysis, dissemination and review. Crucially, the framework does not seek to weaken hierarchical control, but to re-orient governance so that accountability mechanisms support reflection, adaptation and institutional memory within the constraints of intelligence work.

Ultimately, the framework adapts OL theory for command-based intelligence settings and provides a practical roadmap for institutionalising learning across cultural, structural and operational domains. As South Africa's crime challenges intensify, embedding OL within CI is not optional but strategically essential.

The findings suggest that strengthening organisational learning within police intelligence environments requires interventions that work with, rather than against, hierarchical accountability structures. In practice, this means embedding learning activities within routine intelligence work instead of treating learning as an add-on. Structured post-operation debriefings and intelligence review forums, for example, can be institutionalised as protected spaces for reflection, provided they are clearly framed as mechanisms for organisational improvement rather than individual fault-finding. Leadership endorsement is critical in this regard, as supervisors signal whether reflection, questioning and error reporting are viewed as legitimate or risky practices.

At an organisational level, intelligence units may also need to recalibrate performance regimes that prioritise output volume over interpretive quality. The findings indicate that learning is constrained where analytical judgement, contextual reasoning and long-term insight are overshadowed by short-term productivity metrics. Aligning performance assessments with learning-oriented indicators, such as analytical development, knowledge sharing and the effective use of intelligence products, would represent a practical step toward addressing the learning–control tension identified in this study. Importantly, such adjustments do not weaken accountability, but reposition it to support adaptive capability and intelligence effectiveness. These practical implications are directly grounded in participants' experiences of fragmented learning cultures, discretionary training access, weak knowledge retention and persistent control pressures, strengthening the applied relevance of the proposed framework for comparable intelligence and policing contexts.

This study is strengthened by its in-depth qualitative engagement with police intelligence officers, allowing us to capture how OL is understood, enacted, and constrained in everyday practice. Access to participants across multiple units and hierarchical levels enabled us to surface diverse perspectives on learning, leadership, and operational realities, and to trace how learning challenges are embedded in organisational structures rather than isolated behaviours.

Despite these strengths, the study has limitations. It focuses on a single national intelligence division, which may shape how learning dynamics unfold and limits direct generalisation to other policing contexts. Access constraints inherent to intelligence work also restricted observation of certain sensitive processes. Finally, our analysis captures learning at a particular moment in time; longitudinal research could illuminate how learning initiatives evolve beyond the study period.

This study set out to better understand how OL unfolds in police intelligence, a setting where hierarchy, accountability, and operational risk are ever-present and deeply consequential. Our study shows that learning in this context is neither absent nor ignored, but systematically constrained by structures designed to ensure control. This tension, described as a learning–control paradox, helps explain why learning is widely valued in principle yet inconsistently enacted in practice. By examining everyday intelligence work, this study shows that learning is shaped by cultural expectations, leadership behaviour, and fragmented infrastructures that limit reflection, experimentation, and knowledge sharing. Rather than treating these constraints as anomalies, this paper argues that they are constitutive features of intelligence organisations. Learning, therefore, cannot be introduced through isolated interventions or training reforms alone; it must be redesigned to operate within hierarchical authority, not in opposition to it.

Our central contribution is the development of a Strategic Organisational Learning Implementation Framework that reconceptualises learning as an operational capability rather than a peripheral support function. The framework translates lived experience into a coherent set of cultural, structural, and operational mechanisms that embed learning across the intelligence cycle, preserve institutional memory, and connect learning directly to crime prevention outcomes. Future research may wish to explore how learning evolves in intelligence organisations, how trust and accountability shape learning trajectories, and how learning systems function under conditions of political pressure or organisational reform. More broadly, it is hoped this study encourages scholars to take seriously the realities of learning in high-stakes, command-driven environments, where adaptability is essential but never simple.

While this study advances understanding of OL in a crime-intelligence context, it also points to clear directions for future research. Comparative studies across police organisations and intelligence units could shed light on how learning practices vary across institutional, cultural and governance settings. Quantitative approaches may complement qualitative insights by examining relationships between learning capacity, accountability pressures and operational performance at scale. Longitudinal research tracking learning-oriented reforms over time would further illuminate how learning evolves, stabilises or erodes within hierarchical policing environments.

Beyond its theoretical contribution, the study highlights the practical importance of designing OL systems that reflect the realities of intelligence work. The findings suggest that learning failures stem less from individual resistance than from structural conditions that privilege control over reflection. At a policy level, this points to the need to treat learning as a governance concern rather than a discretionary training function, aligning accountability frameworks, performance measures and learning objectives. For organisational reform, the proposed framework offers concrete guidance for embedding learning into routine intelligence practices without weakening hierarchical control. For academic research, the study encourages empirically grounded analyses of how learning, accountability and control interact in high-risk public organisations, using comparative, longitudinal and mixed-method designs to refine and test the framework across contexts.