Original Research

The influence of cultural lag on integrating artificial intelligence in software development

Roxi E. Francke, Kenneth N. Ohei

Received: 06 Feb. 2026; Accepted: 29 Apr. 2026; Published: 01 Sept. 2026

Copyright: © 2026. The Authors. Licensee: AOSIS.
This work is licensed under the Creative Commons Attribution 4.0 International (CC BY 4.0) license (https://creativecommons.org/licenses/by/4.0/).

Abstract

Background: Material culture evolves faster than non-material culture when new technologies emerge, creating a cultural lag that misaligns society, hinders artificial intelligence (AI) integration, and causes resistance, potentially leading to failed integration efforts.

Objectives: The study applies Ogburn’s theory to the modern context of AI, identifying material and non-material cultures that influence its integration and proposing strategies to reduce cultural lag.

Method: The research is grounded in an interpretivist worldview, employing an inductive, qualitative approach rooted in phenomenology. Semi-structured interviews served as the primary data collection method, and thematic analysis supported by ATLAS.ti provided a systematic approach to analysing the data.

Results: The findings revealed that the company’s material culture reflects a multi-tool AI environment. However, non-material cultures (infrastructure limitations, operational constraints, subscription costs, and procurement choices) created a cultural lag. Strategies to reduce this lag were proposed, including leadership communication, structured enablement and operational improvements.

Conclusion: This study established the need to explore material and non-material cultures influencing AI integration and to develop strategies to reduce cultural lag. While AI can transform industries, resistance because of cultural lag often hinders integration, underscoring the need to understand these cultures.

Contribution: The study contributes a strategy model that can assist software development companies and other industries in addressing the gap between technological innovation and cultural adaptation. The findings seek to benefit software developers, IT professionals, business leaders and academics by facilitating smoother technological transitions and improving AI implementation in the industry.

Keywords: artificial intelligence; technology integration; cultural lag; software development; material and non-material culture.

Introduction

Social value

‘Technology may change rapidly, but people change slowly’ (Norman 2002:xiv). While Artificial intelligence (AI) has the potential to transform industries, integrating such technologies often faces resistance because of cultural lag. The phenomenon in which technology changes faster than society is a significant source of social dysfunction (Cruz 2016:65–75). This consequently results in job losses, disillusionment with new technologies, ineffective technological solutions, and broader implications for the economy and societal acceptance of these technologies (Autor 2015:3–4; Griesmyer 2018; Spencer 2021:6–7).

The term ‘Cultural lag’ is a concept introduced nearly a century ago by William Ogburn in 1922, which describes the gap that forms when material culture (technology and tools, etc.) often changes at a faster pace than non-material culture (values, beliefs, social norms, etc.) (Ogburn 1922).

Scientific value

While previous studies have explored technological changes in industries, they have often failed to consider the complex relationship between technological innovations and the cultural environments in which they are introduced (Ciniselli et al. 2024; Lee 2018). A significant research gap in our understanding remains regarding how both material and non-material cultures influence the integration of AI within the selected software development company.

This study proposes to address the critical research gap by applying Ogburn’s Cultural Lag theory to the modern context of AI integration. The study intends to explore the influence of cultural lag on the integration of AI within a software development company in the Western Cape, South Africa, to identify the specific material and non-material cultures that influence AI integration, and strategies to reduce the potential cultural lag, fostering smoother technological transitions and more effective integration of AI in the company. The study holds significance in this era of technological advancement and innovation with AI. The findings could inform stakeholders in the selected software development company, software developers, and policymakers on how cultural lag evolves and on strategies to mitigate its influence not only within this software development company but across various software development industries facing similar disparities.

Literature review

Ogburn (1922) argued that modern culture develops in two aspects: material culture (technology, tools, inventions) and non-material culture (values, beliefs, ethics, social norms). Ogburn (1922:274) reasoned that material culture diffuses more easily within society because tangible objects are adopted into daily routines more readily than non-material aspects, such as languages and beliefs, which are ‘less readily changed’. This temporary misalignment between rapidly advancing material culture and more slowly evolving non-material culture is what Ogburn (1922:199) describes as cultural lag.

The relevance of cultural lag in modern technological contexts has been demonstrated across several scholarly works. Brinkman and Brinkman (1997:616) illustrate how industrial machinery (material culture) outpaced common-law negligence (non-material culture) until workmen’s compensation laws closed the gap. McMahon (2013) explores cultural lag in the adoption of e-textbooks in university classrooms, showing that e-textbook technology advanced faster than student study practices, faculty teaching methods, and institutional support.

More recently, Prasetyo et al. (2022) examine cultural lag during Indonesia’s 2020 regional elections, where digital campaigning advanced rapidly but candidates and the public were unprepared for the digital transformation. These studies demonstrate that cultural lag continues to manifest as a structural feature of technological transitions, where misalignment between material and non-material culture creates adjustment issues, resistance to change, and a mismatch between technological capability and user readiness (Beitinger 2020; Blut & Wang 2020:649, 654; Laumer & Eckhardt 2011; Ogburn 1957:167–171).

Closely related to cultural lag is the literature on how cultural dimensions shape whether technology is integrated or resisted. Hofstede’s (2011) Cultural Dimensions theory identifies six dimensions, power distance, uncertainty avoidance, individualism versus collectivism, masculinity versus femininity, long-term versus short-term orientation, and indulgence versus restraint, that influence how societies perceive and adopt new technologies. Lee, Trimi and Kim (2013:24) and Gallego-Álvarez and Pucheta-Martínez (2021:198) argue that cultures with weak uncertainty avoidance are more agile and willing to experiment with new technologies, while high uncertainty avoidance cultures tend to delay adoption. Zhang and Maruping (2008) and Lee et al. (2013:27) further argue that individualistic cultures are more likely to assess technology independently, while collectivist cultures rely on group cohesion and trusted early adopters. These dimensions provide a useful complement to Ogburn’s temporal account of cultural lag.

A substantial body of recent literature examines AI integration into the Software Development Life Cycle (SDLC). Imam (2024) and Arshad, Butt and Iqbal (2025:74 062) describe how AI is being integrated into each stage of the SDLC, from requirements gathering, where machine learning algorithms automate elicitation and classification, to deployment and maintenance, where AI-powered tools enable predictive analytics, automated rollback, and real-time observability. Soni et al. (2023:2–3), Liljavirta (2025), and Tiwari et al. (2024:43) similarly observe that AI enhances developer efficiency through automated code generation, test case creation, and resource optimisation. Baqar and Khanda (2024:2, 13, 18) and Odeh (2024:125, 132) note that AI automation streamlines repetitive tasks, accelerates iteration, and frees developers to focus on creative and strategic work. At the industry level, Achumie et al. (2022:13–15) highlight AI’s capability to assist companies in creating a competitive advantage, while Mir et al. (2020) and Althati, Tomar and Shanmugam (2024:220–221) identify data abundance and infrastructure as enablers of AI integration. Challoumis (2024:135) emphasises the importance of upskilling development teams as another factor for successful AI integration, and Korinek and Vipra (2025) emphasise that successful AI integration requires ecosystems rather than isolated efforts.

Despite these advantages, the literature also identifies significant challenges to AI integration. Kusak (2022:210) describes the ‘garbage in, garbage out’ problem, where biased or incomplete training data produces inaccurate or discriminatory outputs. Thalpage (2023:31–34) identifies the ‘black box’ problem, where opaque AI models hinder accountability and trust. Kayode (2023:6) and Moch and Oberdieck (2024:28–29) note a shortage of skilled professionals capable of managing complex algorithms, while Laumer and Eckhardt (2011) and Khandii (2019) emphasise that resistance to AI is often driven by perceived threats to job security and shifts in workplace power dynamics. Challoumis (2024:135) argues that progressive companies recognise their future success as linked to ongoing employee development and continuous upskilling.

While this body of work establishes the technical, economic, and cultural dimensions of AI integration, a clear research gap remains. The technical literature on AI in the SDLC tends to address productivity and tooling questions, while the cultural literature on technology adoption tends to address dispositions and dimensions in isolation from particular technologies. Few studies connect the two through an explicit theoretical lens that explains why these patterns recur. Ciniselli et al. (2024) and Lee (2018) note that technological innovations are typically introduced into organisations without sufficient consideration of the cultural environments in which they are received. This study addresses this gap by applying Ogburn’s Cultural Lag theory, supplemented by Hofstede’s Cultural Dimensions theory, to the integration of AI within a software development company in the Western Cape, South Africa, in order to identify the specific material and non-material cultures shaping AI integration and to develop strategies for reducing the resultant cultural lag.

Theoretical framework

The literature review provides a foundation built on four tenets, as illustrated in Figure 1, which emanate from the study’s title. Certain discrete understandings of the four tenets were explored, and together they provide a theoretical foundation for the research based on Ogburn’s Cultural Lag theory. In this section, the information technology and social science literature is reviewed to inform the four tenets. Each of these four tenets is explored individually to understand current perspectives, and a proposition is made for how they come together at a theoretical level.

FIGURE 1: An outline of the literature review.

Figure 1 includes all tenets, starting with tenet 1, the ‘Cultural Lag’. It establishes the theoretical lens of the study, defining and introducing Ogburn’s theory, and distinguishing between material and non-material culture. Tenet 1 defines cultural lag as the temporary misalignment that happens because material culture, technologies, tools, and machines advance faster and diffuse more easily than non-material culture, beliefs, morals, customs, and laws, which are less readily changed. It further explains how these two misaligned aspects cause lag and helps explain why society might struggle to keep pace with rapid technological changes associated with AI.

Tenet 2, ‘The Influence of Culture on Technology Integration’, uses Hofstede’s Cultural Dimensions theory to explain how culture influences whether societies perceive, integrate, or resist new technologies and to compare cultures to understand how cultural dimensions influence the integration of technology in societies.

It argues that technology integration is influenced by power distance, uncertainty avoidance, individualism versus collectivism, masculinity versus femininity, long-term orientation, and indulgence versus restraint, which together influence access, risk tolerance, decision-making, and willingness to integrate new tools.

Tenet 3, ‘AI integration in Software Development’, explains that AI is being integrated into the SDLC to automate operations, accelerate development, and enhance productivity. It further highlights advantages that include enhanced productivity and efficiency, software quality, alongside challenges including heavy data dependency, a shortage of skilled professionals, issues with trust and transparency, and perceived threats to job security.

Finally, tenet 4, ‘AI integration in the Software Development Industry’, shifts the focus to the business level, demonstrating the importance of AI integration in the industry and showing that AI is revolutionising software development practices, which helps companies to gain a competitive advantage, improve cost-effectiveness, and leverage AI-powered technologies to drive revenue growth. It also discusses enabling factors that help explain to why some companies are at the forefront of AI integration, while others trail behind. For example, companies integrating AI for automation, prediction, and personalisation outperform manual approaches, improve cost-effectiveness by reducing errors and accelerating development, and use AI-powered technologies such as predictive analytics, chatbots, and virtual assistants to drive revenue growth and customer engagement.

Research methods and design

Study design

This interpretivist study employed a qualitative research method to gain a thorough understanding of individuals’ lived experiences, perceptions, and attitudes towards AI within the selected company, gathering detailed insights rather than mere statistics and numbers (Creswell 2009; Kumar 2011). In this study, as illustrated in Figure 2, this method was chosen for its ability to explore complex phenomena in real-life settings, particularly when the boundary between the phenomenon (the influence of the potential cultural lag on AI integration) and the context (the selected company within the software development industry) is not clearly defined. This blurring between the phenomenon and the context presents a valuable opportunity for in-depth exploration of the influence of the potential cultural lag on AI integration in a software development company in the Western Cape, South Africa.

FIGURE 2: The intersection of the potential cultural lag and artificial intelligence integration in a software development company.

Setting

This study adopted a single-case study research approach centred around a selected software development company in the Western Cape, South Africa. This qualitative approach allowed for a detailed exploration of a phenomenon within its real-world context. By focusing on a single case, the study thoroughly explores the context and the specific individual experiences of how employees within the company interact with AI, the challenges they encounter, the hesitations they face, and the factors that drive or hinder the process. Therefore, the study gains a much richer and more detailed understanding of the specific cultural aspects involved in AI integration. The single-case research strategy is a recognised qualitative research design (Baxter & Jack 2008:544; Yin 2017), particularly suited to studies in which the aim is to develop a context-rich understanding of a phenomenon rather than to produce findings statistically generalisable to a population. Baxter and Jack (2008:544) argue that a single-case study allows for a detailed exploration of a phenomenon within its real-world context, where the boundary between the phenomenon and the context is not clearly defined. Consistent with this orientation, this study aims to develop a transferable theoretical model of cultural lag and AI integration, grounded in the lived experience of one organisational setting, that future research may extend across multiple cases.

Criteria for selection

The sampling frame for this study includes Information Technology (IT) professionals directly involved with AI technology within the company, as well as senior management and decision-makers. Additionally, the study includes other relevant staff members who, while not directly working with AI, are still impacted by it. The study employed a set of criteria for inclusion in the study: (1), participants must be employed by the selected software development company and (2) must have been with the company for at least 6 months, (3) participants with knowledge or experience of AI within the software development company were selected, ensuring the data are rich and directly relevant to the research questions. Through a non-probability sampling technique, participants were purposively sampled. Eight participants from the company gave their consent to participate, and interviews were conducted. Data saturation was observed after the seventh interview, with no new concept emerging.

Nonetheless, the additional eighth participant was interviewed to verify the thoroughness of existing codes and capture any divergent perspectives. Ultimately, data collection stopped once no new insights emerged, and comparable instances were repeated. This approach helped to strengthen the credibility and confirmability of the findings.

Data collection, ethics and procedure

Because the study is qualitative in nature, interviews were deemed appropriate as a data collection technique.

Specifically, semi-structured interviews were conducted via Microsoft Teams, which was convenient and compatible with the company’s standard technology stack. These interviews were transcribed in real time, with the transcripts serving as the primary data source. To ensure the study’s validity and rigour, the interview protocol was pre-tested with a single participant. Feedback from this pre-test helped to refine the interview guide to ensure clarity, relevance, and alignment with the study’s objectives. The study employed a basic interview guide to steer the conversation and cover all important topics related to the research questions (Creswell 2009). The interview instrument was designed to align with the study’s aim and was based on the four tenets identified from the literature review. The interview instrument was shared with participants beforehand so they could become familiar with the topics to be discussed.

Data procedure

The study employed thematic analysis, based on the framework of Braun and Clarke (2006), to analyse the information collected from the interviews. Thematic analysis was chosen for its systematic method used to identify themes and patterns of meaning throughout the data, making it particularly suitable for studies that explore complex phenomena and subjective experiences (Braun, Clarke & Weate 2016). The data analysis began with the researcher organising and preparing the transcripts, including cleaning the data for accuracy and consistency by removing irrelevant information. Once the familiarisation stage was concluded, the researcher employed ATLAS.ti (ATLAS.ti Scientific Software Development GmbH, Berlin, Germany) to manage and organise the research materials digitally. The researcher then started to analyse the codes and considered how different codes may combine to form an overarching theme.

The next stage involved the researcher reviewing and refining the identified themes. Thereafter, the researcher defined and further refined the themes by identifying the ‘essence’ of what each theme is about (as well as the themes overall) and determining what aspect of the data each theme captures. Having established a set of fully developed themes, the researcher conducted the final analyses and composed the analytical narrative. This narrative included data extracts to illustrate the prevalence of each theme and built an argument related to the research questions as well as the final themes, which provided a thorough understanding of how material and non-material cultural factors influence the integration of AI in the software development company. These were contextualised within and supported by the existing literature. The thematic analysis provided a rigorous approach that directly addressed the study’s aim. The analysis was driven by the information gathered, allowing patterns and themes to emerge naturally without forcing preconceived ideas. This approach laid a foundation for contributing to the theoretical and practical understanding of the phenomenon.

Researcher reflexivity and bias mitigation

Reflexivity is essential in qualitative research, as researchers cannot fully detach themselves from their study (Palaganas et al. 2017:426–427). The researcher acknowledged that her background in information technology and personal assumptions about AI may have subtly influenced data analysis, potentially making her more sensitive to specific narratives shared by participants. To manage this influence, reflexive notes were maintained throughout the study, allowing the researcher to remain aware of her own biases, particularly during thematic coding. For example, INT1’s frustration that ‘when the AI tools aren’t integrated into your databases or into your systems, it doesn’t actually know who you are as a company’ (INT1, Head of Business Development, 3 years) was initially coded as ‘Resistance to Complexity’, but after re-evaluation and member checking it was recoded as ‘Cost and Procurement Barriers’, emphasising integration challenges rather than user frustration.

In addition to reflexivity, the study employed several practical strategies recommended by Ahmed (2024) to establish trustworthiness through credibility, dependability, and confirmability. Credibility was strengthened through member checking, in which interviewees were sent their transcription and a summary of the main points from their interview, and any amendments were incorporated to ensure the transcript accurately reflected their responses (McKim 2023:45). Dependability was supported by maintaining an audit trail covering the interview process, audio recordings, transcription notes, and analysis, with ATLAS.ti used to organise transcripts, track theme development, and ensure consistency. Confirmability was supported by the reflexive notes described above, which captured personal assumptions and helped to ensure that the findings were not unduly shaped by the researcher’s biases. Further bias-mitigation procedures included pre-testing the interview protocol with one participant to refine wording and remove leading questions, sharing the interview guide with participants in advance to reduce the influence of in-the-moment framing, supervisor review of transcripts and emerging codes, and the conduct of an additional eighth interview after data saturation was reached at the seventh, specifically to test for divergent perspectives and confirm code thoroughness. Together, these procedures support the credibility, dependability, and confirmability of the analysis.

Discussion

Demographic analysis

Having outlined the methodological approach employed in the study, which included the research design and data collection procedure, this section provides a detailed analysis of the empirical data and presents the findings based on interviews with participants in the selected software development company. The thematic analysis aligned with Ogburn’s Cultural Lag theory as the underpinning lens.

The participants’ demographics were analysed, providing a clear overview of the eight participants’ details.

The participants were anonymised and are referred to as INT1 to INT8 (Interviewee 1 to Interviewee 8). There were eight participants in the study, and Table 1 presents their demographics and illustrates the diversity of their expertise, ranging from senior executives to technical and creative employees, and reveals varying levels of engagement with and attitudes towards AI integration. This variety strengthens the validity of the study by capturing a broad range of experiences and perspectives in the company.

Thematic analysis

Material cultures influencing artificial intelligence integration

The following findings are key to addressing the material cultures that influence the integration of AI in the company. Figure 3 illustrates the findings that are organised into five sub-themes.

FIGURE 3: The key findings in relation to the material cultures influencing artificial intelligence integration.

Firstly, the findings reveal that the software development company operates a multi-tool AI environment, where AI extends beyond a single model and is embedded across workflows. All eight participants reported using AI for drafting, summarisation, coding support, and deliverables, such as:

‘I use ChatGPT to write proposals and emails.’ (INT1)

‘ClickUp AI will give you an executive summary.’ (INT2)

and

‘you can link it to AI and AI can summarise that meeting … Fireflies.ai Notetaker.’ (INT6)

Artificial intelligence is also accessed through native platform features, for example:

‘I use Copilot in the Microsoft suite.’ (INT3)

This aligns with Imam (2024:30) and Arshad et al. (2025:74 062), who discuss that AI tools are integrated across SDLC stages to automate operations, accelerate development, and enhance productivity.

Secondly, while most integration in the company was plug-and-play, one participant described a more advanced custom application programming interface (API) configuration, indicating the company’s interest in more extensive integration beyond basic tool use. Thirdly, infrastructure constraints limited effective integration. Infrastructure bottlenecks were common, with participants reporting resource strain and intermittent connectivity:

‘It can take so much from my machine … I do sometimes have network issues.’ (INT5)

and

‘the hardware hasn’t caught up yet.’ (INT7)

This reflects a cultural lag in which tool availability outpaces supporting infrastructure, consistent with McMahon (2013) and Ogburn’s (1957) adjustment issues.

Fourthly, operational constraints of AI tools, such as rate limits, occasional timeouts, hallucinations, inaccuracies, network connectivity issues, bugs, and performance differences between tools, influenced participants’ workflows. This aligns partly with Kusak’s (2022:210) ‘garbage in, garbage out’ and Thalpage’s (2023) ‘black box problem’, highlighting trust and transparency concerns.

Finally, cost pressures influenced procurement choices, as only a minority preferred premium tiers to overcome access, privacy, and functionality limits, while others questioned whether AI benefits justified the expense. This aligns with Soni et al. (2023) and Baqar and Khanda (2024), who frame AI’s economic value and argue that AI can automate high-volume, low-value tasks, freeing resources for complex projects and strategic initiatives, and that it can streamline development, enabling faster, more efficient deployment.

Non-material cultures influencing artificial intelligence integration

The following findings are key to addressing the non-material cultures that influence the integration of AI in the company. Figure 4 illustrates the findings that are organised into nine sub-themes.

FIGURE 4: The key findings in relation to the non-material cultures influencing artificial intelligence integration.

The findings reveal that non-material cultures strongly influence how AI is perceived and integrated in the software development company. Artificial intelligence use was described as varying by role and task, with four participants noting differences between strategic and operational use and suggesting that junior or admin roles may be less comfortable than senior developers. This observation aligns with Tenet 3, which focuses on AI use across the SDLC, and Tenet 4, which frames AI integration at an industry level. The findings also revealed employees’ personal initiative, as integration was often situational rather than habitual, with about half of the participants using AI based on their roles and tasks, but not depending on it daily.

In addition, integration patterns revealed that the company culture supports experimentation, yet six participants stated that systems are not well set up for AI integration because of limited infrastructure, a lack of workflow alignment, and systems lagging behind. The company’s experimental culture encourages a bottom-up approach, illustrated by one participant:

‘I introduced ChatGPT for proposal development myself; it wasn’t a top-down approach.’ (INT1)

These observations are consistent with Hofstede’s Cultural Dimensions theory, suggesting traits of low uncertainty avoidance and an individualistic culture that enables employees to experiment freely with new technologies.

The findings also revealed that attitudes towards AI integration were mixed. Five of eight participants were initially sceptical because of unfamiliarity, a lack of skills, or trust issues, while the remaining three participants were optimistic. The findings also indicated that participants often expressed enthusiasm alongside concern, for example:

‘I love it. I think it speeds up processes’, but also found ‘a general fear that AI is going to replace everyone’s jobs.’ (INT1)

These findings align with Beitinger (2020) and Blut and Wang (2020), who argue that technology advances faster than users can learn it effectively, and that discomfort and distrust reduce readiness for integration.

Furthermore, trust and transparency emerged as situational, with six participants emphasising that AI should not be relied on in sensitive areas and that full end-to-end generation is ‘not good practice’. This reflects Thalpage’s (2023) ‘black box problem’, highlighting the need for transparent systems where accountability is essential.

In addition, the findings revealed the role of human verification. Five participants described AI as generating initial drafts or templates, but humans retaining responsibility for fact-checking and quality control. One participant stated:

‘I will actually probe the AI even further … to test whether I am getting a factual response.’ (INT3)

While another explained:

‘I cross-reference those answers … sometimes I go to YouTube, Google.’ (INT5)

The observations support the existing literature on verification testing, as noted by Arshad et al. (2025) and Liljavirta (2025), and the importance of transparency, as cited by Thalpage (2023).

The findings further revealed that leadership training was largely absent, with seven of eight participants reporting no leadership-led AI training, despite a desire for the company to become more AI-led. This aligns with Power, Schoenherr and Samson (2010) and Zhang and Maruping (2008), who note that individualistic cultures may not prioritise team-based development. Leadership communication was similarly limited. Seven participants were unclear about expectations, with one stating:

‘No, I don’t think there’s any expectation.’ (INT8)

Yet participants also acknowledged a supportive tone:

‘There’s this openness about AI … But there’s no sort of expectation that you have to use AI for everything.’ (INT8)

This reflects weak uncertainty avoidance cultures that impose fewer rules and remain open to innovation.

In addition, the companies’ shared framing of AI showed that all participants viewed AI as a supportive tool or partner rather than a replacement for human judgement, consistent with Soni et al. (2023), who frame AI as a development assistant boosting productivity. Finally, the majority of the participants believe that a cultural lag exists in the software development company, where AI advances faster than the company’s skills, readiness, and processes. This aligns with Ogburn’s cultural lag theory and with Beitinger (2020) and Blut and Wang (2020), who stress that simply introducing new tools or systems does not mean that users are prepared to integrate them into their work or daily lives.

Strategies to reduce the potential cultural lag

The following findings are key to addressing the strategies to reduce the potential cultural lag in the integration of AI in the company, as illustrated in Figure 5.

FIGURE 5: The key findings in relation to strategies for reducing the potential cultural lag.

Structured enablement emerged as a key strategy to reduce the cultural lag in the software development company, with participants emphasising the need to implement learning in practice by incorporating updates on AI tools, prompts, and use cases into recurring meetings. Two participants proposed turning AI learning into a team habit, while another recommended peer learning to improve skills. This aligns with the findings of Challoumis (2024:135), who stresses that upskilling and investment in training demonstrate commitment to employee empowerment and successful AI integration.

The findings also identified onboarding materials as an important enablement strategy. One participant suggested simple learning resources and AI guidebooks to help new employees adopt AI more effectively.

This reflects traits of an individualistic culture, consistent with the study of Lee et al. (2013:27), who found that individuals in such cultures seek information from direct and formal sources independently.

Practice and iteration were also emphasised, as comfort levels improved through repeated use and prompting refinement. Participants found that detailed context and precise wording improved output quality and speed, reinforcing Challoumis’s (2024:135) argument that continuous skill development is essential for integration.

Mentorship was highlighted as another structured enablement strategy. Participants noted the absence of a strong AI mentor to guide the company into the AI space, with one recommending an AI mentor to increase confidence and awareness of AI. This aligns with collectivist traits described by Lee et al. (2013:27) and Zhang and Maruping (2008), where individuals rely on early adopters and trusted users when integrating new technologies. Korinek and Vipra (2025) similarly argue that AI integration cannot be achieved in isolation and requires partnerships and ecosystems.

Leadership communication was identified as the second strategy. The findings revealed that although the company is open to AI use, leadership has provided limited formal guidance. Participants reported unclear goals, expectations, and policies related to AI, and expressed a desire for leadership to establish a clearer vision, ethical boundaries, and frameworks governing AI use. This aligns with the literature on uncertainty avoidance, suggesting that while low uncertainty avoidance cultures support experimentation with fewer rules, employees still value basic guidance from leadership on AI use.

Operational improvements formed the third strategy. Participants emphasised selecting appropriate tool tiers, noting that safer paid subscriptions reduce usage limits and improve data handling. This aligns with the findings of Soni et al. (2023), Tiwari et al. (2024), and Baqar and Khanda (2024), who frame AI’s economic justification through time and labour savings. Finally, resource and infrastructure alignment was also identified as essential to reduce the cultural lag, as technical readiness remains constrained by hardware performance and intermittent connectivity. The findings suggest that upgrading hardware, improving networks, and integrating AI tools with core systems are necessary to reduce friction and ensure successful AI integration. This aligns with the results of Mir et al. (2020) and Althati et al. (2024:220–221), who argue that AI integration depends on necessary infrastructure, processing capacity, and system readiness.

Ethical considerations

Ethical approval was obtained from Cape Peninsula University of Technology’s Faculty Research Ethics Committee (No. 217194540/2025/2) prior to the commencement of the data collection phase. Participants were fully informed about the nature and aim of the research and provided written informed consent to participate voluntarily. Confidentiality was strictly upheld. Data were securely stored and accessed only by the supervisor and the researcher. Audio recordings were stored in a secure location for 5 years, after which they will be permanently deleted. The study ensured that participants were treated fairly and equitably, and that selections were made without bias.

Results

The study identified both the material and non-material cultures influencing the integration of AI in the software development company, and it proposes strategies to reduce the resultant cultural lag. The distinct material cultures existing in the company are characterised by a multi-tool AI environment that integrates AI into various workflows. In addition, the company offers both basic and advanced API-level AI integration; however, infrastructure constraints, operational constraints, costs, tiers, and procurement choices hinder the effective and full utilisation of these AI tools. In terms of the distinct non-material cultures, it became evident that while the company adopts a bottom-up, role-sensitive approach to AI integration within a supportive culture, its infrastructure readiness and a lack of clear policies, training, formal communication, and leadership guidance present constraints. It was also found that as familiarity and skills improved, participants became less sceptical, leading to greater acceptance of AI integration; however, the capability gap remains the main challenge to integration across the company.

The cultural lag that exists within the company can be attributed to the misalignment between these two cultures.

The study, therefore, proposes strategies to reduce this lag during the integration of AI. These strategies include leadership communication, structured enablement, and operational improvements. Leadership communication comprises clearly communicating goals, expectations, and policies related to AI from leadership, while maintaining the company’s supportive and non-coercive approach. This could involve leadership introducing clear policies and communication channels that provide structured guidance on approved tools, ethical use, data protection, and other AI-related best practices. A second strategy is structured enablement, which includes learning through practice, providing onboarding materials, encouraging practice and iteration with AI tools, and appointing mentors to raise confidence and awareness. The final strategy is operational improvement by upgrading its tools through selecting tool tiers that address limitations in access, privacy, and functionality, and prioritising the alignment of resources and infrastructure. Together, these strategies can reduce cultural lag and support smoother and more effective AI integration across the company.

Situating these findings within the wider empirical literature reveals important points of convergence with prior case-based research on AI integration in software contexts. Imam (2024) and Liljavirta (2025), in their case studies of AI integration across the SDLC, similarly document the integration of AI across SDLC stages, including how AI tools support requirements gathering, design, code generation, testing, deployment, and maintenance. Mir et al. (2020) and Althati et al. (2024:220–221) identify data abundance and infrastructure as enablers of AI integration, both of which surfaced as constraints in the present findings. Challoumis (2024:135) similarly argues that progressive companies link successful AI integration to ongoing employee development, reinforcing the structured enablement strategy proposed by this study.

Beyond software-specific work, the recurrence of cultural lag across very different technological transitions, including the e-textbook adoption studied by McMahon (2013), the digital campaigning examined by Prasetyo et al. (2022), and earlier industrial transitions analysed by Brinkman and Brinkman (1997:616) suggests that the misalignment between rapidly advancing material culture and more slowly evolving non-material culture is a structural feature of technological transitions rather than a feature unique to any single organisation or technology. This study contributes to this conversation by framing these recurring patterns through Ogburn’s Cultural Lag theory, thereby offering a coherent theoretical explanation for why the same constraints surface across diverse contexts and supporting the analytical generalisability of the proposed Strategy Model.

Discussion of key findings

As mentioned previously, according to Ogburn (1922), modern culture develops in two aspects: material culture (technology, tools, inventions, etc.) and non-material culture (values, beliefs, ethics, social norms, etc.).

The authors observed that these two aspects do not evolve at the same pace, with material culture typically advancing more quickly than non-material culture. As shown in Figure 6, this disparity, termed cultural lag, highlights the temporary misalignment between these cultural aspects and the challenge societies face in aligning their behaviours with innovation (Ogburn 1922:199).

FIGURE 6: Cultural lag.

The findings of this study led to the development of a modified version of Ogburn’s depiction of Cultural Lag theory, as illustrated in Figure 7. This modified version, ‘A Strategy Model to Reduce Cultural Lag for AI Integration’, is specifically designed for the integration of AI in the selected software development company in the Western Cape, South Africa. Unlike the original depiction of Ogburn’s Cultural Lag theory, this model introduces three strategies to reduce the cultural lag, which is particularly relevant in this context. This modified model contributes to expanding the study’s theoretical framework and serves as a practical tool to aid in reducing cultural lag. The model could potentially be applied beyond this single software development company to the software development industry and other industries experiencing similar misalignments.

FIGURE 7: A strategy model to reduce cultural lag for artificial intelligence integration.

Figure 7 indicates that as time progresses, material culture advances rapidly, supporting Ogburn’s (1922) argument that material culture typically evolves faster than non-material aspects of society. The material culture is represented by the green solid curve, which rises sharply during the initial phases of AI integration because of the company’s readiness to use AI tools.

The non-material culture, including employees’ attitudes, leadership training and communication processes, is represented by the dashed red curve. Non-material culture initially lags behind material culture as employees deal with distrust, skill gaps and the lack of formal policies.

Over time, however, the dashed red curve is likely to accelerate upward as the company implements strategies such as leadership communication, structured enablement, and operational improvements. The three strategies are presented in chronological order, where each one builds on the previous: firstly, leadership communication aligns the company; secondly, structured enablement then empowers employees; and thirdly, operational improvements formalise the AI integration process.

The distance between the two lines represents the cultural lag, which Ogburn (1922:199) describes as a temporary misalignment between these cultural aspects and the challenge societies face in aligning their behaviours with innovation. Over time, the two curves can converge, symbolising the closing of the cultural lag where technological innovation and company culture evolve harmoniously. This alignment improves AI integration, supporting Ogburn’s theory by suggesting that alignment is both possible and important through deliberate adaptation of non-material culture to match technological advancement.

Strengths and limitations

The study’s limitations include a relatively small sample size of eight participants from a single software development company located in a specific geographic region, the Western Cape, South Africa. The sample size limits the generalisability of the findings to a broader population of software development companies. A larger sample, consisting of more software development companies, may have provided a broader understanding of the varied experiences and perspectives within the industry. Moreover, because this study focuses on the Western Cape region, it may not fully represent the experiences of all software development companies in other parts of South Africa, especially those with different technical and cultural contexts.

This study’s single-case study research strategy limits the ability to generalise the findings to other software development companies. The replication of this single case may prove challenging if the organisational context, technology landscape, or culture changes significantly over time. By recognising these limitations, the study’s findings should be interpreted with caution, and further research is necessary to validate and extend these findings.

The boundedness of the case is a methodological strength as much as a limitation: it permits the depth of engagement necessary to surface the cultural dynamics that broader survey-based designs may miss. Strengths of the study include its alignment with Ogburn’s Cultural Lag theory as an underpinning lens, the use of a pre-tested interview protocol, the collection of data until saturation followed by a confirmatory eighth interview, the application of member checking to support credibility (McKim 2023), and the systematic application of Braun and Clarke’s (2006) thematic analysis framework supported by ATLAS.ti.

Implications or recommendations

Future research could build directly on this study by conducting multicase comparative studies that examine the influence of cultural lag on AI integration across multiple software development organisations, ideally spanning different sizes, sectors, and cultural contexts. Such studies would enable testing and refinement of the proposed Strategy Model in varied settings. In addition, future research could explore conducting longitudinal studies to track changes in AI integration over time. Furthermore, quantitative studies and mixed-method studies could be undertaken to gain a more comprehensive understanding of AI integration.

Finally, further research could focus on the ethical issues arising from the influence and integration of AI in the software development industry.

Conclusion

At its core, this study explored the integration of AI within the software development industry. The study adopted Ogburn’s Cultural Lag theory as its theoretical lens, and the findings echo Ogburn’s concept that the company’s material and non-material cultures advance at different speeds, creating a lag that hinders technological integration.

The research resulted in the development of A Strategy Model to Reduce Cultural Lag for AI Integration. This model identifies the material and non-material cultures influencing AI integration and proposes strategies to reduce the potential cultural lag. Sustainable and effective integration relies on aligning material innovations with non-material values, beliefs, and practices. The findings aim to benefit stakeholders, including software developers, IT professionals, business leaders, investors, academics, researchers, policymakers, and educational institutions, by facilitating smoother technological transitions and improving the overall effectiveness of AI implementation in the industry.

By applying this theory, the study confirms that Ogburn’s Cultural Lag remains a relevant theory for understanding digital transformation in the modern context of AI, contributing to the expanding body of knowledge in this field. Thereby, the identified model can be applied beyond this single software development company to the wider software industry and other industries experiencing similar misalignments.

This allows companies to identify their material and non-material cultures and develop strategies to reduce cultural lags, enabling smoother and more effective technological integration.

Acknowledgements

This article is based on research originally conducted as part of Roxi E. Francke’s master’s thesis titled ‘The Influence of Cultural Lag on Integration of Artificial Intelligence in a Software Development Company in the Western Cape, South Africa’, submitted to the Department of Information Technology, Faculty of Informatics and Design, Cape Peninsula University of Technology in 2025. The thesis was supervised by Kenneth N. Ohei. The original thesis is publicly available at: https://etd.cput.ac.za/entities/publication/7976d367-e286-418e-ab50-582482334d43. The thesis was reworked, revised and adapted into a journal article for publication. The author affirms that this article complies with ethical standards for secondary publication, and appropriate acknowledgement has been made of the original work.

The authors would also like to acknowledge all those who contributed to this study.

The authors, Roxi E. Francke and Kenneth N. Ohei, declare that they have no financial or personal relationships that may have inappropriately influenced them in writing this article.

Roxi E. Francke: Conceptualisation, Data curation, Investigation, Methodology, Software, Visualisation, Writing – original draft, Writing – review & editing. Kenneth N. Ohei: Project administration, Supervision, Writing – review & editing. All authors reviewed the article, contributed to the discussion of results, approved the final version for submission and publication, and take responsibility for the integrity of its findings.

The authors received no financial support for the research, authorship, and/or publication of this article.

The data that support the findings of this study are available on request from the corresponding author, Roxi E. Francke. The data are not publicly available because they contain information that could compromise the privacy of research participants.

The views and opinions expressed in this article are those of the authors and are the product of professional research. They do not necessarily reflect the official policy or position of any affiliated institution, funder, agency, or that of the publisher. The authors are responsible for this article’s results, findings, and content.

References