Business Analytics has evolved from being a purely technical matter, where the emphasis was only on analysing the growing volumes of data, to being regarded as a social phenomenon that involves human capabilities. Previous researchers (Ongena 2023) have shown that business analytics in the public sector has been a focal point since its regular application. However, there are underlying complex challenges associated with its application, and if not addressed, these can have a negative impact (Fredriksson et al. 2017). Literature analysis demonstrates that previous studies (Al-Ajmi & Al-Busaidi 2024) have been primarily focused on business analytics in developed economies. There is a need for scholarly work in the context of developing economies, which have received less academic attention (Maroufkhani et al. 2019; Sabharwal & Miah 2021). Therefore, this study seeks to contribute to the body of knowledge by answering the following research questions:

The study employed the PRISMA framework (Page et al. 2021) and structured this article accordingly. The articles considered eligible for review were published within the past 8 years (2018–2025). One of the criteria was to include only journal and conference papers that have been peer reviewed and published. In addition, the practical gaps identified in the literature (Oesterreich et al. 2022b; Phaladi & Ngulube 2024) include the lack of frameworks for knowledge retention in public enterprises, specifically in business analytics. Such an absence of frameworks results in knowledge loss, poor service delivery and an inability to improve taxpayer compliance. Despite the South African private sector’s advancements in business data analytics (Majam & Jarbandhan 2022), the public sector remains lagging. The situation is attributed to challenges such as the scarcity of data analytics skills, inadequate skills acquisition and retention capabilities, and limited knowledge sharing among professionals in the business analytics domain, as separately advocated by Shava and Vyas-Doorgapersad (2022) and Majam and Jarbandhan (2022). There is a variety of articles published already in different sectors; however, this article includes the public sector as part of the eligibility criteria, with a more focused look at the data analytics teams.

Five different databases were used to search for articles, including AIS eLibrary, AJIS, Web of Science, Scopus and ScienceDirect. These databases were selected because they have high-impact quality articles (Mohammed & Ozdamli 2024). After applying the inclusion and exclusion criteria, 42 articles were selected, as shown in Figure 1.

A total of five search queries were executed in the different databases listed in section 2.3. The vital part of search strings was to identify the keywords aligned to this research and combine them appropriately to yield the best results:

The study employed inclusion and exclusion criteria to select articles that were more closely aligned with the research objectives and to determine the scope of the systematic literature review. Six inclusion criteria were used, with the publication period being solely restricted from 2018 to 2025. To ensure that quality articles were selected, only journal and conference papers that have been peer reviewed and published were included. To streamline the area of study, the third criterion was to select articles from the Information Systems field. The fourth criterion was to select articles which reported on Business Analytics, Knowledge Sharing and related subjects. The area of study is organisations from public enterprises, mainly government organisations and other related areas; therefore, this criterion was used to eliminate articles that focused on different areas. Lastly, all articles which were included were written in English only. The exclusion criteria were then based on all the articles which does not meet the inclusion criteria. Table 2 documents the inclusion and exclusion criteria.

The researcher utilised computer software called NVivo to facilitate the data coding process. NVivo enabled the researcher to select aspects of data that were required to analyse, code phrases that summarised the portions to be analysed, group codes together to form categories and finally build themes from the categories of information (Hilal & Alabri 2013; Mezmir 2020). The inductive method was employed for the application to the code, enabling data analysis that aided in answering the research questions. The software is user-friendly and robust. In the current study, the coding process was conducted systematically, eventually forming a pattern from the developed themes (Rambaree & Faxelid 2013). It helped to synchronise data for in-depth analysis, supporting what Creswell and Creswell (2017) echoed, that NVivo aids in coding data efficiently through rigorous probing and analysis. The first research question is ‘How can information systems capabilities be enabled to improve business analytics in public enterprises?’ Different themes emerged; however, through the utilisation of the sub-categories of the research questions, more meaningful themes aligned with the research started to appear, and this allowed for the grouping of these themes based on the information systems capabilities factors.

The second research question is ‘How does trust influence knowledge sharing in business analytics in the public enterprises?’ The themes presented a lot of challenges linked to trust, which hindered knowledge sharing.

The third research question is ‘How can knowledge sharing be enabled to improve business analytics in public enterprises?’ The initial codes were derived from the research question. These codes were then refined by focusing on the themes that presented knowledge-sharing factors, which aided in revealing the enabling factors that can result in improved business analytics.

Business analytics research areas have been studied for several years. It is deemed relevant for this study to have proper methods for scope limitation and to include only articles aligned with the study. Additionally, the researchers used an 8-year period to mitigate selection bias. The bulk of the articles used for the study were from Europe and the US, with only a few articles (five) from South Africa. This revealed the existing gap in Business Analytics in Africa, which indicates the necessity to delve further into and research this area in Africa. Although the study spanned across different geographical scopes, the generalisation was strengthened through deriving findings from diverse contexts, which will result in applicability across different regions, instead of focusing on one country.

Information systems capabilities can improve business analytics by enabling organisations to manage the end-to-end process of capturing, applying and leveraging knowledge and expertise attained internally. However, human factors, such as skills and knowledge, play an important role in creating value in business analytics (Kristoffersen et al. 2021). As reported by Vidgen, Shaw and Grant (2017), managers encounter several challenges in creating business value. Such challenges include the lack of a clear data and analytics strategy, the right people with competencies to instil a data-driven cultural change, and alignment between business analytics capabilities and the organisation’s business strategy. Several scholars have investigated how organisations can improve their data competencies by incorporating expertise from interdisciplinary teams and other analytical domains (Shirani 2016; Vijayasarathy & Jetley 2025). These studies confirmed the vital role played by human factors, such as skills and knowledge, in creating value in business analytics. This includes the organisation’s capacity to attract and retain skilled employees, such as data scientists and data engineers, who can create value from the data. However, sourcing employees who possess both technical and business skills is difficult. A resource is defined as a specific asset or know-how. On the other hand, capabilities are defined as the skills acquired through the organisation’s processes that enable it to utilise the assets and competences comprising a set of skills and a group of technologies (Wamba et al. 2024). The lack of data skills within various data teams, combined with a deficiency in leadership skills and capabilities, leads to the failure of most big data projects (Gürlek & Çemberci 2020), as noted by Gartner (Persaud 2020). This has a negative influence on organisations’ drive to serve their clients efficiently and provide quality data. On the contrary, as much as some technical data and human skills are scarce, they can be sourced on the market if organisations provide the right incentives (Ferraris et al. 2019). Klee, Janson and Leimeister (2021) also revealed another challenge for business analytics, showing that organisations depend on systematic data analytics competencies to use data appropriately and make strategic decisions, including determining the most suitable approach between inductive and deductive methods. Thus, if these data competencies are not available, then significant data synergies cannot be recognised without them, which has a negative impact on business value. Furthermore, organisations face challenges in managing business analytics as they strive to establish a data-and information-driven framework to generate business value (Vidgen et al. 2017). These challenges include the lack of a clear data and analytics strategy, the right people with competencies to instil a data-driven cultural change, and alignment between business analytics capabilities and the organisation’s business strategy (Vidgen et al. 2017). In alignment with the above findings, this study is extending to look closely at the information systems challenges in business analytics and the impact on the public enterprises. Human capital and institutional logic facilitate the coordination of different technologies through data science and domain-specific skills (Di Vaio et al. 2022). Therefore, IT capabilities, IT human capabilities and IT knowledge management capabilities are deemed important for organisations to attain a competitive advantage (Muazu & Abdulmalik 2021). Retention, motivation, development, and attraction of employees with the right data analytics skills should be emphasised for achieving successful organisational performance (Gurlek 2020). Enhancement of generation, usage, sharing, and continuous knowledge protection requires IT knowledge management capabilities to enable organisations to manage the end-to-end process of capturing and applying knowledge and expertise attained internally. This knowledge becomes a distinctive competence once it is retained as explicit knowledge, assisting in providing a competitive advantage (Muazu & Abdulmalik 2021). Data scientists must possess these competencies to enhance their performance and become innovative in an ever-evolving global data environment (Hattingh et al. 2019). It is also vital to note that the data scientist cannot possess all the competencies; instead, they complement each other to perform beyond expectations and improve the organisation’s competitiveness (Hattingh et al. 2019). Competency extends beyond knowledge and skills to include attitudes, behaviour, personal and abilities traits and work habits (Persaud 2020). As demonstrated by Goswami et al. (2025), both analytics competence and analytics culture are positively related to ambidexterity (Goswami et al. 2025). Therefore, for an organisation to reap benefits from analytics competencies derived from its technical resources, an analytics culture needs to be established (Al-Ajmi & Al-Busaidi 2024; Ghasemaghaei 2019). Market competence can be enhanced through the effective use of analytics, facilitated by analytics expertise and a culture that supports analytics.

The incorporation of business analytics into business operations yields immediate positive effects (Klee et al. 2021), as this strategy can facilitate the formulation of data-driven models. Therefore, an established framework is needed to highlight the significance of these skills in enhancing organisational performance. Business analytics initiatives require professionals with both technical and managerial skills, as data-analytical thinking is essential for data scientists; nevertheless, research suggests that managers can also benefit from it (Mikalef et al. 2018). Insights generated from business analytics lead to innovation, improved productivity, growth, enhanced service delivery, cost control and increased customer value (Persaud 2020). However, the shortage of skills, the lack of knowledge and experience on deployment and the conversion of insights into greater value challenge the realisation of the anticipated value (Johnson et al. 2021; Persaud 2020). Organisations seek all four types of competencies to be part of a job role, as reported by Persaud (2020). Therefore, employees are expected to possess both the social competencies and meta-competencies. Although cognitive and functional competencies are essential, they are often considered insufficient for making hiring decisions. This study argues that all four competencies are essential, and teams working in business analytics environments must consider them, as they are complementary to one another. As such, team members need to integrate and share all the knowledge and expertise they possess to allow for this knowledge to be utilised for improved business analytics. Furthermore, business analytics knowledge is a tacit resource, and this knowledge also serves as a driving force to aid in achieving organisational competitive advantage. Skills shortages, inability to recruit and retain SMEs in these public enterprises in South Africa are some of the prominent challenges identified, and it is recommended that organisations invest in data analytics training programmes for improved business analytics (Bag et al. 2021; Ochuba et al. 2024; Sutherland 2020). In addition, skills and capabilities, including data analytics and programming, can be socially complex if there is no coordination and knowledge transfer among experts with in-depth knowledge. Bag et al. (2021) report that the biggest challenge experienced by organisations is the hiring of SMEs who possess these IS capabilities. Such skilled personnel are hired at a high cost, and management needs to motivate them by altering policies to extend the salary bands (Bag et al. 2021). Findings reveal that skills shortages and the transferability of knowledge are key considerations for the success of significant business analytics initiatives.

Trust within teams is one of the factors that influence knowledge sharing (Agbejule, Rapo & Saarikoski 2021; McAllister 1995). However, trust has been defined in different ways by the researchers; McEvily and Tortoriello (2011) found more than 129 definitions of trust. For this study, interpersonal trust will be the key focus since knowledge sharing between two business analytics experts aids in the retention of specialised knowledge. Interpersonal trust relationships can be categorised into two forms: cognitive-based trust and affect-based trust (McAllister 1995). Cognitive-based trust is more about an individual’s rational decision or belief to trust based on previous knowledge of the reliability and dependability of the other team member. In contrast, affect-based trust is based on the emotions, care and concern of the other team member, and vice versa; it is primarily a reciprocal relationship built over a more extended period. An exploration of the impact of cognition-based and affect-based trust on employees’ willingness to share and use tacit knowledge revealed that both types of trust have an influence (Holste & Fields 2010). Affect-based knowledge has a greater impact on employees’ willingness to share tacit knowledge. In contrast, cognition-based trust plays a more significant role in their desire to utilise tacit knowledge (Holste & Fields 2010). Building on prior research, Rutten, Blaas-Franken and Martin (2016) investigated the levels of knowledge sharing between coworkers in high- and low-trust scenarios, focusing on both affect-based and cognitive-based trust, and comparing explicit and implicit knowledge. Findings revealed a significant difference in the level of knowledge sharing between high- and low-trust environments, with the difference being notably higher in affect-based trust and implicit knowledge (Chua, Thinakaran & Vasudevan 2023; Khan et al. 2022; Rutten et al. 2016). Trust is the foundation of a supportive work environment where employees can share knowledge and insights about their work tasks. Management plays a key role in enabling workplace spirituality (Khan et al. 2022). This is also applicable to data analytics teams, where trust plays a crucial role in their decision-making regarding whether to share knowledge or not (Maja & Letaba 2022). Functional experts possess critical data knowledge and competencies, and they rely on a trust-based culture within their teams to share and solve problems that require human intervention (Vasilyeva & Richardson 2022). Trust enables this open exchange of knowledge; thus, the positive mindset bestowed by trust enables the knowledge bearer to view the knowledge shared as a benefit rather than a cost (Issac et al 2022; Ghasemaghaei 2023). This also concurs with Wah et al. (2018), who reiterated that employees are likely to be more open and willing to engage in innovative behaviour if they perceive trust in their line manager and ultimately become open to exchanging information and sharing knowledge. In the context of this study, based on the Social Exchange Theory, a business analytics specialist will be willing to share knowledge if trust is demonstrated in the social relationship, thus minimising complexity. However, an attitude of helping and collaboration will be shown to other team members. Leadership needs to influence knowledge sharing in teams by creating a transparent environment at the individual level, which fosters trust among team members (Khoreva & Wechtler 2020). Trust among team members is emphasised, so that the knowledge bearer can understand the benefit for the team and the organisation of sharing the knowledge. On the contrary, where there is no trust, the risk of inaccurate knowledge being shared is present. At the same time, trust is subject to how the knowledge bearer perceives it towards others and decides whether it poses an improvement or risk in the workplace.

Facets of knowledge sharing and information systems capabilities are considered for recommending a business analytics implementation framework. Employees possess different data analytics skills within organisations. Still, as a result of a lack of a proper knowledge-sharing framework, there are limited opportunities to share these skills and utilise business analytics effectively. Shabbir and Gardezi (2020) affirmed that big data might not perform as expected if the data quality does not allow for true replicability of models and if there is no clarity on the sources that cause instability in the models. These deep insights require knowledge sharing within data teams, so that employees with the necessary skills to ensure data quality can assist in pinpointing the specific sources that cause silo instability (Akter et al. 2019b). However, data teams face interpersonal challenges, including knowledge sharing (Didas 2023; Gupta et al. 2023). Understanding these challenges is of significant value, especially during new product development (Simsek et al. 2019). Resultantly, senior executives need to be tasked with building significant data initiatives that integrate data and analytical capabilities already existing within the organisation. However, this can result in cultural challenges, such as resistance to knowledge sharing because of organisational silos and disputes arising from the implications of analytical insights (Issac et al. 2022; Memon, Qureshi & Jokhio 2020). Organisations that have a strong emphasis on knowledge sharing have a competitive advantage and are more likely to succeed in open innovation pursuits; however, Singh et al. (2021) caution that for organisations operating in a dynamic business environment, knowledge can quickly become obsolete. Hence, it is essential to adopt innovative policies and practices. Through the lens of a community of practice, Mosala-Bryant and Hoskins (2017) revealed that knowledge sharing was perceived as necessary, and knowledge was shared through various means, including meetings, presentations and the Department of Public Service and Administration’s website. However, the study’s findings concur with those of previous scholars in reporting a lack of institutional repositories or knowledge portals to store all the knowledge shared during meetings or presentations. However, contemporary studies (Akter et al., 2019a) argue that knowledge sharing in business analytics is enhanced through digital technology. Such knowledge can be stored in knowledge repositories, such as Wikis and SharePoint, for easy accessibility and enhanced decision-making. Recently, Deng, Duan and Wibowo (2023) found that digital technology facilitates knowledge sharing, resulting in improved decision-making and enhanced employee job performance. An interesting study in the knowledge domain was conducted by Mahura and Birollo (2021). The findings reveal that IT project teams more commonly practised informal knowledge transfer methods in the public sector to enhance knowledge retention. However, the private sector practices formal knowledge transfer to comply with the organisation’s requirements. Therefore, a formal knowledge transfer method to enable knowledge retention is recommended. Another challenge faced by the public sector is the lack of knowledge retention of business domain knowledge gained from data analytics teams (Abeysekera 2021; Hagen & Hess 2021). A closer look at the causes of enterprise tacit knowledge loss in State-Owned Enterprises in South Africa showed that knowledge management capacity was hindered by both voluntary and involuntary employee turnover in the absence of knowledge retention investment strategies (Phaladi & Ngulube 2024). Building on Phaladi and Ngulube (2024), this study seeks to comprehend how knowledge is shared and retained within business data analytics teams. Moreover, this study aims to provide an understanding of the influence of knowledge sharing among data analytics teams on the improvement of business analytics in public enterprises based on the outcomes from previous studies.