Latest / 1 More Paper / Guiding principles for the development of adaptive self-regulating guidelines for the use of generative artificial intelligence in the business customer experience — Kanganga & Mtotywa (2026)
Transcript
- You’re listening to “Guiding principles for the development of adaptive self-regulating guidelines for the use of generative artificial intelligence in the business customer experience,” by R.J. Kanganga and M.M. Mtotywa. Published in 2026. ISSN: 2331-1975 (Online) Journal homepage: the linked source Guiding principles for the development of adaptive self-regulating guidelines for the use of generative artificial intelligence in the business customer experience Rufaro Joylyn Kanganga & Matolwandile Mzuvukile Mtotywa To cite this article: Rufaro Joylyn Kanganga & Matolwandile Mzuvukile Mtotywa (2026) Guiding principles for the development of adaptive self-regulating guidelines for the use of generative artificial intelligence in the business customer experience, Cogent Business & Management, 13:1, 2620163, DOI: 10.1080/23311975.2026.2620163 © 2026 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group Published online: 01 Feb 2026. Submit your article to this journal Article views: 931 View related articles View Crossmark data Citing articles: 1 View citing articles Artificial Intelligence, Digitalization, and New Technologies | Research Article Guiding principles for the development of adaptive self-regulating guidelines for the use of generative artificial intelligence in the business customer experience Rufaro Joylyn Kanganga and Matolwandile Mzuvukile Mtotywa Rhodes Business School, Faculty of Commerce, Rhodes University, Makhanda, South Africa ABSTRACT. Generative Artificial Intelligence (GenAI) is reshaping business customer experiences. However, the benefits of adopting GenAI are competing with the inherent shortcomings like biases and hallucinations. Therefore, to mitigate the risks of GenAI adoption, the study develops guiding principles to achieve adaptive self-regulation in the use of GenAI within business customer experiences. The theories undergirding the research are Albert Bandura’s Social Cognitive Theory (SCT) of self-regulation and Cybernetics. This theoretical triangulation overlaps on how information is processed to self-regulate GenAI. Cybernetics focuses on the use of feedback (information) to effect control and maintain equilibrium. Whereas Albert Bandura’s SCT of self-regulation highlights the role of information about internal, environmental, social or collective goals and standards in influencing decision making. This decision making is informed by making observations or garnering information from self-monitoring, making a judgement and taking self-reactive measures to self-regulate. The research gap and rationale behind adaptive self-regulation is the systemic lag between the development of GenAI and the adaptive self-regulation. To this end, the high costs and elongated time periods involved in promulgation of GenAI regulatory legislation create a gap. This gap in South Africa specifically meeting the demands for adaptive legislation for GenAI business customer experience is not abating. Therefore, the principles of Cybernetics and Bandura’s SCT of self-regulation undergird this research towards developing guiding principles for adaptive self-regulation of business customer experience. Hence, the views of South African technology professionals were garnered in line with the interpretivist paradigm adopted for this research. Through an exploratory qualitative approach, semi-structured interviews were conducted. The findings confirmed that GenAI offers efficiency, time-saving and actionable insights, but also presents gaps, including biases, hallucinations, a lack of oversight, and privacy. Most of the participants recommended safe and trustworthy GenAI with a hybrid deployment (human involvement and oversight). No new ideas and themes emerged once saturation was reached. Saturation was attained after 21 interviews. The guiding principles that emerged from the findings were also congruent with principles from the Organisation for Economic Cooperation and Development (OECD) AI recommendations on AI. These were first adopted in 2019 and revised in 2023 and 2024 to reflect new developments. 1. Introduction. ARTICLE HISTORY GenAI; business customer experience; guiding principles; self-regulation; adaptive SUBJECTS Artificial Intelligence; Information Technology; Business, Management and Accounting Generative Artificial Intelligence (GenAI) is a subset of Artificial Intelligence (AI) systems that can create autonomous content when prompted. Artificial Intelligence (AI) is a wide set of computational simulations of human intelligence or intellectual capability, for decision making, learning or inference. Therefore, the difference between AI and GenAI is that AI simulates human intelligence while GenAI generates new content by learning patterns and relationships within large datasets. GenAI leverages computational systems to autonomously generate novel content, including but not limited to images, faces, voices and music. Some of the transformational benefits of business customer experiences include efficiency, automation, speed, assisted decision making and productivity. Business customer experience refers to the set of emotions evoked in a customer as a result of their interaction with a business. To facilitate this, policymakers and managers need to have a clear understanding of the inner workings of data-driven decision-making algorithms. The use of such GenAI models entirely without guardrails has risks involved, for example hallucinations, deepfakes, copyright infringements, bias, privacy concerns and others. These risks are also exacerbated by the regulatory gaps of GenAI governance, including the lag between the pace of the development of GenAI and its simultaneous adaptive regulation. According to the Oxford dictionary, the word adaptation where adaptive comes from refers to the process of altering something to adapt to changing conditions. Therefore, to stay abreast with these nefarious consequences of GenAI, immediate, agile, adaptive interventions are required. One of the first intergovernmental recommendations on AI were approved by the Organisation for Economic Cooperation and Development (OECD) and G20 leaders in 2019. The goal of the recommendations is to ensure the responsible and trustworthy AI in accordance to human rights and values. To remain relevant, adaptive and aligned to the development of AI and GenAI the recommendations have so far been revised in 2023 and 2024. In South Africa, a regulation gap for GenAI existed. Only in 2024 many years after AI and GenAI adoption, the first-ever National AI Policy was promulgated. However, South Africa’s policy is for AI and there is still a gap for policy on GenAI in business customer experience to protect consumers and businesses. Furthermore, several countries are only recently promulgating legislation on AI and not distinguishing or promulgating individual laws to govern GenAI. Additionally, some of these policies are not adaptive, or agile and moving at the same pace with the development of newer GenAI. Therefore, there is a risk of national policies becoming obsolete in a short space of time. Hence, adaptive self-regulation initiated by businesses emerges as a pragmatic alternative to national regulation. State-led regulation can often become constrained. This is due to several reasons, including significant fiscal burdens and suffering from limited enforcement capacity due to the influx of businesses adopting GenAI. Ultimately, self-regulation leverages the agility and domain-specific expertise of industry actors to establish and maintain ethical standards. Moreso, GenAI governance is pertinent to confront the risks in GenAI business customer experience. Thus, businesses are balancing the adoption of GenAI for their business customer experiences while maintaining ethical and safety measures. Hence, to self-regulate GenAI, it is imperative to understand the implications, risks, limitations, capabilities and potential of GenAI. Against this background, the study aimed to develop guiding principles for adaptive self-regulation guidelines for the use of GenAI in business customer experiences. This was investigated according to three research objectives; to determine the role, use and benefits of GenAI in the business customer experience; to explore the gaps and shortcomings encountered in GenAl business customer experience; and to identify the steps towards developing adaptive self-regulation principles for the use of GenAI in the business customer experience. 2. Review of the literature. 2.1. Theoretical contribution: social cognitive theory of self-regulation and cybernetics theory. This study was undergirded by drawing insights from Bandura’s Social Cognitive Theory (SCT) of self-regulation and Cybernetics Theory. Notably, Bandura’s work focuses on motivating self-regulation by attaining specific standards and goals whereas Cybernetics is centred on maintaining homeostasis. Similarly, both theories are premised upon receiving information and regulating the state of a living organ or machine. Although both theories recognise that when the information or feedback is negative there is a need to self-regulate to align with standards the regulation behaviour is different. Bandura recognises that self-regulation decisions and goals are motivated by the level of self-efficacy of the decision maker. However, in Cybernetics the feedback informs how to control the system to restore equilibrium. Bandura’s SCT recognises that goals, or standards can change influenced by external and internal features and the environment whereas in Cybernetics the goal is stability or equilibrium. Hence for this research elements from both theories are pertinent. GenAI models must maintain a state of stability and not crash (Cybernetics) while being self-regulated when new goals and standards have emerged (Bandura’s SCT of self-regulation). Ultimately, Bandura highlights a triadic causation where self-belief (self-efficacy) shapes a perception of whether an environment can be controlled, which influences their subsequent choices. Similarly, Cybernetics focuses on human-machine interactions and the role of communication and control in self-regulating the system through feedback. Thus, the overlap between Bandura’s SCT and Cybernetics is primarily centred on feedback and control. GenAI systems can simulate the cognitive skills and intelligence of humans by self-reproducing from learning patterns, a phenomenon also acknowledged by Cybernetics. Interestingly, Norbert Wiener also notes that machines can learn patterns of behaviour and self-reproduce, which is similar to the foundational principle of GenAI. However, both theories recognise that various extenuating factors can impede control. Wiener highlighted that too much feedback can negatively influence control. Similarly, Bandura noted that low self-beliefs or self-efficacy in the capacity to change the environment can also negatively influence the self-regulation goals and behaviour of the managers. Together the above theoretical insights, through the integration of Bandura’s SCT and Cybernetics, provides a nuanced understanding of achieving adaptive self-regulation of GenAI to mitigate the risk associated with the technology. 2.2. GenAI within business customer experience. The research of business customer experience and GenAI is still in its formative stages since GenAI technologies are still fairly new; for example, ChatGPT only came out in 2022, and Meta AI was just released in 2024. Contextually, GenAI models used in business customer experience can be self-regulated by garnering businesses through feedback from customers. GenAI-enabled business customer experiences encompass two dimensions: hedonic (memorable, entertaining and novel) and recognition (feeling valued, respected and safe) aspects. 2.3. Role use and benefits of GenAI in business customer experience. The role, use and benefit of GenAI models spans across various domains of industry and commerce. This includes banking, entertainment, mining, and more, to enhance customer experiences. GenAI in business customer experiences has been linked to several advantages. These include empowering enterprises to refine their feedback communication strategies, gauge customer satisfaction with precision, and foster more empathetic connections. GenAI provides various strategic opportunities such as personalised email campaigns, dynamic website content, and customised advertising, thus fostering enhanced customer loyalty. GenAI models in business customer experiences support the timely delivery of products and services. Furthermore, GenAI is beneficial in customer relationship management, offering personalised outputs. Notwithstanding, this underscores the need for the security of customer personal details. Additionally, GenAI provides insights for data-driven multitier marketing campaigns. Similarly, GenAI is beneficial in large data analysis. A notable example is CarMax, a used vehicle retailer that is leveraging GenAI to generate concise text summaries for its car research pages. However, such information shared with machines can also lead to hacking and the compromise of businesses. Ultimately, GenAI benefits are competing with the innate gaps of GenAI, which are further explored below. 2.4. Gaps and shortcomings of GenAI in business customer experience. Several gaps emerged from the interaction of customers with GenAI. Key issues include hallucinations, privacy concerns, system downtime, biases, compromised learning development and absence of human intervention and deepfakes. These warrant the need for ethical scrutiny and adaptive self-regulation in GenAI adoption for business customer experiences. The 2023 ChatGPT ban in Italy, due to privacy issues, illustrates the dangers of unregulated GenAI use. Despite its promise in areas like healthcare cost reduction, GenAI carries significant risks, including data misuse. Privacy concerns are especially common during sensitive interactions; thus, customers often prefer in-person service due to fears of data misuse. System downtime and failures to update or correct responses also risk damaging customer trust and business continuity. Biases arise when models are trained on limited or skewed data, leading to distorted outputs. Hallucinations occur when GenAI generates convincing but fabricated outputs. Learning development is at risk as students increasingly rely on GenAI, exposing them to misinformation, plagiarism and reduced critical thinking. Despite GenAI enhancing customer service, the absence of human oversight in high-stakes interactions (e.g. banking) can increase risk, such as financial loss from deepfakes. Deepfakes are AI-generated media that mimic real individuals, often used to defraud or manipulate. Understanding these limitations informs the need for safe, trustworthy, human centric GenAI deployment and adaptive self-regulation strategies. 2.5. Practices in self-regulation of GenAI. A culture of self-led GenAI regulation is critical for responsible business customer experiences. The rationale behind self-regulation is that it is a more scalable approach, to rollout at organisational level, can take less resources and time to implement. Globally, several nations have introduced policies or guidelines on GenAI. An intergovernmental set of principles governing AI were welcomed by the G20 in the Osaka Summit in 2019. These principles highlighted the responsible and safe stewardship of trustworthy AI. Collaboration among nations in GenAI governance is growing. BRICS+ countries (Brazil, Russia, India, China, South Africa, UAE and others) have also developed GenAI guidelines. Scholars such as Gans (2018) and Hirsch et al. (2021) advocate for GenAI regulation that serves the public good, not just corporate interests. 3. Methodology. This exploratory qualitative research design is grounded in the interpretivist paradigm. The interpretivist approach was adopted to understand the lived experiences of the technology experts and their insights on GenAI self-regulation. In this investigation, a nonprobability sampling method was used, where a sample of 21 business professionals in South Africa was purposively selected and interviewed for this study. The criteria used were that they needed to have experience working in a technology company doing business in South Africa, be familiar with GenAI and have a technology background (Table 1). To ensure a robust data collection strategy a new information threshold measure was rigorously applied. This involves a qualitative measure to determine whether enough evidence exists. In this study, the acceptable threshold was set at ≤5% and achieved at 21 participants (Figure 1). This benchmark is similar to the statistical analysis method of using a p-value of <0.05 or <0.01 when determining evidence on whether to accept or to reject a null hypothesis. Similarly, it is noteworthy that although this benchmark is widely referenced it may not entirely guarantee saturation but rather it provides a replicable and transparent way of calculating data saturation that other research scholars may comprehend. Hence the lower this new information threshold the lower the ratio of undiscovered themes after the threshold has been attained. In this case after 21 interviews responses were similar and had plateaued, no new information were emerging from the participants. For example, the most common themes, such as biases, lack of human oversight and efficiency, were now recurring. In other similar research studies, 12–20 participants were considered an adequate sample. Semi-structured interviews were conducted using a hybrid between online and in-person, with online interviews conducted and recorded on the Zoom Platform. This approach ensured an accurate capture of participants’ responses, allowing for detailed enquiry into each focus area in the research questions. The interview guide used is in Appendix A. The interview transcripts were analysed using a step-by-step structured approach adapted from Naeem et al. (2023). The steps include: (i) transcription and identification of quotes, (ii) selection of key words, (iii) coding, (iv) themes. Company involved in the conceptualisation stages of GenAI (Source: Authors). Initially, 175 in vivo (direct quotes from participants) codes emerged. These were organised for relevance, consistency and clarity to ensure they capture the true meaning of the findings. These were consolidated, generating 22 unique consolidated codes which formed seven code groups or themes. This grouping and subsequent data analysis were conducted using computer-assisted software, Atlas.Ti 22. The study further advanced trustworthiness through extensive involvement with the design and collection process. There was also data triangulation, which ensured the adequacy and relevance of the sample. The purposive sampling strategy eliminated any threat to validity, minimising biases and increasing the transferability of the findings. The study also ensured transferability to similar settings, confirmability and dependability by providing brief accounts of all steps taken to collect the data. Participants were engaged with the same semi-structured question to ensure objectivity. 4. Findings of the study. The empirical analysis resulted in seven themes that were role use and benefits in GenAI, gaps of GenAI, enablers of GenAI self-regulation - self-monitoring GenAI, feedback error correction, judgement of monitored processes, self-reaction and control of the GenAI processes, and self-efficacy subfactors. 4.1. Research findings on the benefits and gaps of GenAI. Participants articulated their awareness of the benefits and gaps of balancing the use of GenAI technologies. For example, Participant 15 highlighted that through GenAI use customers are saving time ‘...you can do some mobile banking where you are using AI on the platforms to help you through the processes, it’s like giving back time to people’. These sentiments were also shared by the other participants, and the findings show that not only are customers saving time, but businesses save time and focus on core business. Technology businesses and clients are also reaping the benefits. For example, Participant 5 highlighted that ‘it saves time. Sometimes I was taking eight hours to write code, but when I use AI, it takes me two hours’. Similarly, Participant 2 noted that ‘we outsource our coding in most cases..., I am sure they are using generative AI for them to meet deadlines...’. However, the former raises ethical risks regarding the compensation of businesses and people for GenAI deliverables. Another common trend in the findings related to the use and benefits of GenAI is the efficiency associated with the adoption of these technologies in the customer experience. Participant 18 highlighted that ‘there is potential for improved efficiency by learning from feedback’. Additionally, Participant 7 also highlighted that GenAI helps bring solutions more easily. One technology professional expressed the prowess of the GenAI model in summarising reports on business customer experience. Participant 1 explained that ‘we can just upload your CSV files there and then go and analyse’. The participants also highlighted that GenAI models can be used to monitor sales performance in the business by providing information on the performance of their customer experience, which provides actionable insights. However, concerns were continuously raised on security and privacy concerns. A common gap that participants identified was the biases associated with the adoption of GenAI models. These include training data biases, human biases and GenAI learned biases. Participant 19 noted that, ‘...the bias in that training dataset could creep in...’. Therefore, the findings suggest that it is not uncommon for GenAI models to contain biases. Furthermore, Participant 2 believed that ‘there’s a hallucination factor where generative AI kind of likes to make up its own thing...’. Another participant stressed the importance of limiting access to sensitive information and protecting customer privacy. Even more worrying for businesses is that not only are their customers a prime target, but their business data is also. Participant 10 noted that ‘... the exchange of information through the GenAI models for business customer experiences can attract hackers’. Therefore, the role of human oversight and security checks cannot be overemphasised. It is clear from history that technology cannot fully escape human beings. This was echoed by participant 18, who noted that, ‘There will always be a need for human intervention, as there will be concerns about ethical safeguards and transparency’. 4.2. Developing principles for achieving adaptive self-regulation guidelines. In line with the interpretivist philosophy of this research, the perspectives from the participants’ lived experiences provided input into the development of guiding principles of adaptive self-regulation. Congruent to Bandura’s SCT of self-regulation the findings in the participants’ insights underscored the significant role of self-monitoring. Contextually, self-monitoring involves careful observance and awareness of the GenAI and its ecosystem. Participant 19 noted that adequate self-monitoring leads to enhanced efficiency by learning from feedback and adjusting. To add to this, Participant 11 recommended that for this to be effective, businesses should adopt an agile approach where they monitor their results. Participant 8 also underscored the relationship between self-monitoring and ‘getting the best performance’. Equally, this oversight, according to participant 9, is ‘very important’ for businesses to stay abreast of their GenAI business customer experience performances. A majority (75%) of the participants, when asked if they had mastered their ability to monitor and identify gaps, felt that they were self-efficient. Feedback and error correction were also highlighted. Continuous feedback allows customers and businesses to continuously engage with GenAI models in an agile manner to eliminate gaps that may arise. Because GenAI models are predominantly trained on data, it is possible that some of the data will be outdated, and the models will also become outdated. Hence for adaptive self-regulation to occur regular updates play a critical role. Participant 10 highlighted that, ‘there are tools that help you determine if maybe a certain function has broken or maybe it needs an update’. Therefore, tools can be used to automate and keep track of the regular updates of the GenAI models. It is also clear from the experts’ perspectives that there needs to be some form of knowledge of standards and acceptable practices of GenAI to initiate discernment and judgement over the performances. Participant 5 highlighted that, ‘So to be able to make a judgement, you have to understand what the biases are, what it means’. Therefore, participants cited the importance of organisational policies in determining when the monitored performances were beneficial or a cause for concern. Hence, participants also identified that benchmarking against industry standards becomes relevant. Proactive measures, as opposed to reactive ones, were common among the participants in terms of self-reaction to GenAI risks. One of the participants identified quality assurance testing as a self-reaction tool. Participants recognised the significant role that testing plays before, during and after the deployment of GenAI customer models. Proactive testing aids in identifying the cause of anomalies and quality assurance, to ensure the integrity of the GenAI, which facilitates the self-regulation process. Furthermore, other participants described the role of proactive measures where safety is embedded at the inception of the GenAI through prompt engineering and safe by design models. Furthermore, the participants highlighted that customers also had the onus to restrict the information that they shared and to self-regulate the models by not compromising their personal information. On sharing information, Participant 15 was of the view that, ‘My starting point is to be as conservative as possible and I operate on a need-to-know basis’. Thus, highlighting the importance of clarity and self-awareness of what is being shared to avoid the exchange of sensitive data with GenAI. About 75% of the participants felt that they were self-efficient when asked if they had mastered their ability to monitor and identify gaps. This is particularly important, as businesses need to be confident in their ability to adapt to self-regulation. This belief also partially influences how experts perform other self-regulation sub-functions (self-monitoring, judgement and self-reaction). Similar to informative feedback in Cybernetics theory, participants recognised the role of knowledge where GenAI administrators are trained for self-regulation. Participant 3 expressed that GenAI models and humans must interact as a system to achieve optimal business customer experiences. However, Participant 7 recognised that there are limitations to what models can do versus what a human is able to achieve in terms of customer experiences. They emphasised that ‘it lacks the human factor’. Equally, they highlighted that there are questions or issues that GenAI cannot resolve. Hence, the findings advocate for a dual holistic approach where GenAI and humans collaborate to manage and adaptively self-regulate business customer experiences. 5. Discussion – conceptual framework and adaptive guidelines. 5.1. Conceptual framework. The conceptual model in Figure 2 illustrates how businesses grapple with balancing the benefits and the gaps or shortcomings in the use of GenAI in business customer experiences. The model is grounded in Bandura’s Social Cognitive Theory of self-regulation and Cybernetics Theory, which jointly inform five adaptive self-regulation mechanisms: feedback and error correction, self-monitoring, self-judgement, self-reaction and self-efficacy. This conceptual model focuses on the contribution of two theories to achieve agility and adaptivity in adaptive self-regulation of GenAI in business customer experience. This model theoretically guides how businesses adapt to this ever-evolving technology landscape. For example, in universities that need to quickly adapt to self-regulating GenAI technologies, they need to maintain the integrity of their qualifications. These conceptual guidelines serve as a mediator to manage and resolve the balance between GenAI’s benefits and risks. Theoretically, the aspect of adaptive self-regulation is closely related to the acquisition of knowledge. In this case, the acquisition of knowledge on the real-time GenAI performance leads to the alteration of the subsequent behaviour of the model. Contextually, self-monitoring the performance of the GenAI model is critical towards achieving efficiency, mitigating risks and adaptively influencing subsequent behaviour. 5.2. Guiding principles for adaptive self-regulation guidelines. Guiding principles are foundational for effective guideline development, ensuring consistency, transparency and utility. They function as heuristic tools to inform decision making. Ultimately, 3 principles were developed based on existing GenAI governance principles, the research findings and the two theories undergirding the research (Table 2). Principle 1: Principle of safe GenAI design. This principle is at the design stage and underscores the safe design and engineering of GenAI algorithms that promote fairness, bias mitigation and transparency. This builds on the next principle of transparency and clarity through self-monitoring. GenAI inherently involves security threats and vulnerabilities, which businesses must manage to promote fairness and accuracy. Hence, the principle emphasises safe design of GenAI at inception, ensuring integrity and preventing unauthorised access throughout the GenAI lifecycle. Self-regulating systems must proactively address these concerns without relying solely on external enforcement at the design stage. Principle 2: Principle of clarity and transparency through self - monitoring. The next step once GenAI design is safely engineered, is self-monitoring the design to ensure clarity and transparency. Human oversight and GenAI self-monitoring systems promotes safety by providing clarity and transparency on the performance of GenAI. Self-monitoring helps maintain control over GenAI output, eliminating errors and fostering transparency, accuracy and data responsibility. It enables real-time awareness and enhances business accountability in customer-facing applications. In terms of transparency, the engineering behind the design of the GenAI and clarity on how the model is trained is critical. This not only promotes accountability but reduces any risks while enhancing efficiency, innovation and competitiveness. Additionally, clarity and transparency on the GenAI preserve human rights and promote the explainability of the functionality of the GenAI to the customers. This builds trust as customers gain clarity on how personalisation capabilities and features are derived from their data. While GenAI enhances personalisation and other features concerns remain over privacy and ethics remain. Hence, this principle emphasises the role of continuous human self-monitoring to bridge any knowledge gaps and supporting adaptive self-regulation. Principle 3: Principle of human-centredness. Humans not only play a critical role in ensuring the safe design and deployment of GenAI but also in self-monitoring, assessment or judgement of the system performance and subsequent error correction. It is humans that also raise a flag when a GenAI has negative performances and also that lead to the rectification of error and malperformances. Research participants recognised that humans cannot be eliminated from any of the GenAI processes. In fact, other the research participants and the OECD recommended that GenAI should be built with humans in mind preserving human rights and fostering human inclusion through hybrid approach to GenAI deployment. Thus GenAI-human strategies where recommended where the merits of both in business customer experiences are leveraged. For example, where a GenAI has limited comprehension on customer’s requirement there is an option to connect with a customer service consultant. Furthermore, humans are valuable towards integrating benchmarking, testing and other self-regulation measures. The research findings also highlighted that GenAI cannot substitute human empathy contextual understanding and expertise. However, risks exist in terms of biases and human error. OECD also recommends working with humans to empower them to ensure that they are trained and aware of the technologies and not completely displaced by technology. 5.3. Practical application of the guiding principles. Several case studies support the practical application of certain guiding principles by businesses. Businesses can apply these principles as a reference to support the development informed self-regulation goals and behaviour by learning from best practices and standards. Each principle can be operationalised by policymakers in the drafting of GenAI governance and interventions by organisations. Furthermore, these guiding principles can be applicable in informing future research that may want to build upon similar extant literature. Developers can also find practical application of these principles in informing the quality of GenAI tools they develop. For example, by developing, explainable AI (XAI), training modules and secure Application Programming Interfaces (APIs). Ultimately, the practical implications of these principles provide foundational elements towards the development, governance and maintenance of GenAI. This promotes transparency, accountability and agility in the self-regulation of GenAI in customer-facing operations. 5.4. Implications of the study. 5.4.1. Implications for management. The implications of these principles for management lie in customising the adoption of GenAI to suite their unique business circumstances. This involves change management, decision making, training and goal setting. For example, the OECD recommends building capacity and preparing for the technology. It was also recommended to train staff and equip them for the transition and any customised roadmap should be focused on preparing managers and teams for the transformation that GenAI brings. In practise, managers are always improving their after-sales support, processing orders and other tasks despite gaps such as privacy breaches and sometimes customer dissatisfaction. Therefore, by building upon these principles, managers can structure their self-regulatory interventions undergirded by research. Such structured approaches ensure consistency, rigour and transparency in governance of GenAI. The research also contributes to informing policymaking. Ultimately, the research findings corroborate existing literature, demonstrating a significant concordance between the theoretical foundations and practical implications. This dual approach strengthens the role of the academic body of literature and practise. Furthermore, this triangulation demonstrates to management the role of research in the governance of GenAI. This involves consulting both technology practitioners and extant research to bolster a wider understanding considering both human and GenAI interaction. This creates a holistic and robust approach to GenAI adaptive self-regulation. 5.4.2. Implications for policymakers. Policymakers consult research and other sources to develop policies and also need to review and stay abreast of the latest trends in their legislation or policies. Furthermore, in line with best practices, research participants also recommended a human-centred approach. Furthermore, a multidisciplinary approach towards achieving adaptive self-regulation was recommended. This includes other areas outside of technology, such as psychology. To policymakers, this approach widens the lens of the understanding of GenAI while eliminating biases. Furthermore, the South African National AI Policy also converges with Bandura’s theory by reinforcing that policymakers adopt a human-centered approach to adaptive self-regulation that promotes human oversight. The participants provided credence to the role of benchmarking and standardised approaches in policymaking based on the lived experiences of the research participants. Lastly, as GenAI continues to evolve and become increasingly ubiquitous, policymakers must prioritise the development of effective adaptive self-regulatory guidelines. By triangulating the literature review with the research findings, this study provides nuanced guidance to policymakers on achieving adaptive self-regulation of GenAI. Thereby contributing to the advancement of knowledge in this field. 5.4.3. Theoretical contribution of the study. The theoretical contribution of this research is the expansion of research on the triangulation of Albert Bandura’s self-regulation SCT and Cybernetics, focusing on the role of adaptive and agile self-regulation of GenAI. This area of theory is also critical for understanding the implications of the moral and ethical experiences of business customers engaging with GenAI. Lastly, the theoretical contribution is towards the understanding of GenAI technology benefits and gaps, and where GenAI models are lagging in terms of self-regulation in business customer experiences. Ultimately, the research provides a continuation of existing literary works on how businesses can become more adaptive and agile in their GenAI self-regulation approach. 5.5. Limitations and future studies. Notwithstanding the study has a few limitations. Firstly, the study’s scope on the use of GenAI was limited to technology professionals/experts with no customer input on GenAI customer experiences. The former could be another area of future research. Moreover, for the scope of the research, the participants are also GenAI customers and provided insights from both the business and customer point of view. Secondly, the study’s focus on GenAI adoption in business customer experience may not reflect other areas of GenAI adoption, such as education. The rationale behind contextualising the research in South African business customer experiences is that at the time of the inception of the research (early 2024), there were gaps in terms of AI regulations in this country. Notwithstanding their efforts, it was clear that the country needs such knowledge and research. The National AI Strategy was only launched recently in August 2024. 6. Conclusions. The research study demonstrated the critical need for adaptive self-regulatory guiding principles to mitigate gaps and shortcomings while maximising the benefits associated with the use of GenAI. The technology continues to influence all aspects of our business and society and is rapidly becoming unavoidable owing to its ability to streamline and automate business processes. However, businesses are continuously grappling to optimise and leverage the benefits of GenAI without experiencing unintended harms and gaps associated with its use. This study contributes to the expansion on research on the of adaptive self-regulation of GenAI models, an area still in its formative stages, as this technology is still new. This study demonstrated the critical need for adaptive self-regulatory approaches to mitigate risks obsolete regulatory measures. The empirical research yielded guiding principles for adaptive self-regulation of GenAI in business customer experience. The findings of the study support the intersection of Albert Bandura’s SCT of self-regulation and Cybernetics theory where feedback and control play a significant role. As GenAI continues to evolve and become increasingly ubiquitous, researchers, practitioners and policymakers must prioritise the development of effective adaptive self-regulatory guidelines. This ensures the safe, transparent and responsible use of this technology, especially to protect customers. Acknowledgments. Rufaro Kanganga is involved in the conceptualisation of the paper, methodology, software, formal analysis, writing – original draft and review, and editing. Matolwandile Mtotywa is involved in the conceptualisation of the paper, methodology, supervision, writing – original draft, review and editing and the final approval of the version to be published. All authors agree with all aspects of this study. Authors’ contributions CRediT: Rufaro Joylyn Kanganga: Conceptualization, Formal analysis, Methodology, Software, Writing – original draft, Writing – review & editing; Matolwandile Mzuvukile Mtotywa: Conceptualization, Methodology, Supervision, Writing – original draft, Writing – review & editing. Disclosure statement There are no potential conflicts of interest associated with this study. Funding There was no funding available for this research work. About the authors Rufaro Joylyn Kanganga holds a Master of Commerce from the Rhodes Business School with distinction for her research on the adaptive self-regulation of GenAI in business customer experience. She also holds a Bachelor of Commerce in Information Systems Honours degree with distinction. Her Honours research focused on the use of QR Codes in Real Estate for property viewing. Currently her research focus is in the area of safe and ethical use of disruptive technology such as AI and governance. She actively pursues and aspires to participate in informing policy development and governance not only for AI technologies but for all disruptive technologies. Rufaro’s career has evolved from Real Estate, to Information Systems and more recently to Governance. She has served as a consultant in the Governance department of an international development partner for many years and looks foward to consulting for governments and boards. Dr. Matolwandile Mzuvukile Mtotywa works within the Operations and Decision Sciences field, integrating operations management with analytical, cognitive, and computational decision layers. He teaches quantitative decision-making and value creation through artificial intelligence, and he coordinates the Postgraduate Diploma in Business Analysis. A double PhD holder in Science and Commerce, his research focuses on decision science, data science, and business analysis at the nexus of business, society, and technology, especially fourth and fifth-industrial revolution technologies. He has more than 20 years of academic, executive, and consulting experience, is an author, a PhD and Master’s supervisor, and an editorial board member. Data availability statement The data that support the findings of this study are available upon reasonable request. Interested readers may contact the corresponding authors. 2. In your own view, what are the current benefits of using GenAI in business customer experience in South Africa? Explain. 3. In your own view, what are the gaps or shortcomings in the use of business customer experience in GenAI applications? Explain. 4. In view of the identified gaps, what do you suggest can be done in terms of the GenAI self-regulation? 5. In your own opinion, advise the use of GenAI in coding, and have you also used it? For example, ask GenAI to write code for an application or load code for correction in GenAI tools such as chatbots. Explain 6. Have you developed any GenAI business customer application that was adaptively self-regulated? Explain. 7. What is your opinion of adaptive self-regulation of GenAI? Explain. 8. In your view, have you mastered the ability to monitor GenAI processes? In other words, are you confident in your ability to keep a close eye on any GenAI business customer experience application and immediately pick up irregularities that contradict regulation measures? Can you immediately address or mitigate any risks effectively? 9. How good is your judgement in terms of classifying whether an incident should be regulated, especially where it concerns GenAI processes, irregularities such as biases, inconsistencies, copyright infringements, security breaches, and any other undesirable effects? 10. Are you self-monitoring the performance of GenAI applications that you interact with through continuous learning on GenAI to stay aware of any new developments or threats and engage in new materials? 11. Are you able to take steps that show your initiative to self-regulate the use of GenAI effectively? 12. Any other comments related to the subject matter under discussion?