Beyond the Hype: Cultivating a THINK With AI™ Culture in Your Organisation with Mark Garrett and Stefan Steffen

Executive Summary

This webinar outlines key aspects of Artificial Intelligence (AI) implementation within organisations, exploring its multifaceted impact on talent transformation, team collaboration, and work management. Mark Garrett and Stefan Steffen address the challenges and opportunities presented by AI, including the governance of generative AI, compliance with privacy standards such as email archiving, and the necessity of human involvement in collaboration and decision-making processes. Furthermore, the webinar emphasises the importance of knowledge workers and the implications of AI adoption on project efficiency, particularly in agile development, while also examining potential biases in AI-generated content and prospects for future adoption. By understanding these dynamics, organisations can effectively navigate the complexities of AI integration and leverage its benefits for sustainable growth.

Webinar Details

Title: Beyond the Hype: Cultivating a THINK With AI™ Culture in Your Organisation
Date: 30 September 2025
Presenter: Mark Garrett and Stefan Steffen
Meetup Group: DAMA SA Big Data
Write-up Author: Howard Diesel

The Implementation of AI in the Organisation with Mark Garrett and Stefan Steffen

Howard Diesel opened the webinar and indicated that Mark Garrett would primarily lead the presentation, sharing slides. At the same time, Stefan Steffen contributed context regarding his role at the global AI innovation league for the League of Companies.

The League of Companies is a collective of experts specialising in various aspects of digital transformation, with a heightened focus on artificial intelligence (AI) this year. One of their key partners is Lucid AI, led by CEO Lucid Mark, who has developed the “thing with AI” framework, currently being implemented in South Africa and beyond. To contextualise the AI landscape, it is compared to the evolution of mobile phones, which have become essential tools for navigation and communication, integrating AI capabilities that enhance daily life.

AI is increasingly becoming integrated into every aspect of our lives, both professionally and personally. As it continues to evolve over the next few years, it is crucial to consider how to integrate AI effectively into our businesses and daily routines. This integration can enhance productivity by enabling individuals to achieve better results and identify new ways to solve various problems. The presentation will highlight these themes, showcasing the significant implications of AI technology and its diverse applications across different contexts.

The Challenges of AI in Society and Industry

Mark shared that the focus of the presentation would be on the implementation and challenges of artificial intelligence (AI) in the business context. Additionally, the session would explore the role of AI in enhancing productivity and problem-solving across various sectors, paralleling the evolution of AI with the mobile phone revolution.

Concerns surrounding AI data governance and the perceptions of generative AI are pivotal issues in the ongoing discussion about AI adoption. Many users express resistance to embracing AI due to fears of being perceived as cheating, which highlights the need for educational systems to adapt and incorporate AI constructively.

By fostering a supportive environment that encourages the responsible use of AI technologies, society can facilitate a broader shift towards integration and acceptance of these innovations. Ultimately, addressing these concerns can pave the way for a more informed and collaborative future in AI utilisation.

The perception of generative AI as a form of cheating has created hesitancy in its adoption within educational settings. Many individuals believe that using such tools undermines personal thinking and creativity, presenting a significant challenge for universities attempting to change this mindset. Some educators are advocating for a more supportive approach, encouraging students to embrace generative AI rather than shy away from it.

It was noted that a professor from the University of California, San Francisco, emphasised the importance of engaging with these technologies to avoid potential disappointment in the future. This perspective aligns with broader discussions in education about the relevance of traditional methods, such as long division, in an era where technology, including calculators, is integral to learning and understanding. Ultimately, fostering an open attitude toward generative AI may better prepare students for a technologically advanced world.

The current state of AI is a significant focus in the speaker’s daughter’s educational experience, who attends high school and engages in diverse subjects. In her English class, she would explore classic literature, such as “Othello” and “A Streetcar Named Desire,” while facing strict policies regarding the use of AI. On Mondays and Wednesdays, she transitions to her computer science class, where students use a program called Replit to practice coding effortlessly, allowing them to create, debug, and publish an AI-based application by the end of the semester. This dual exposure illustrates the integration of AI in both humanities and technical fields within her curriculum.

Mark then highlighted the contrasting dynamics within organisations regarding certain topics. While some subjects remain taboo and are forbidden on one side of the hall, others are viewed as expected and necessary on the opposite side. This phenomenon is prevalent in companies globally, regardless of their size.

The Impact of Generative AI in Data Science and Its Governance

Generative AI has evolved from a phase of hype to one of strategic advantage, marking a fundamental shift in how we utilise technology. The debate around its legitimacy, often framed as “cheating,” can be likened to using tools like Excel for bookkeeping or a skill saw for construction—once considered unconventional, these practices are now standard due to their impact on productivity. The proper application of AI not only enhances efficiency but also enables individuals to allocate more time for cognitive thinking and collaboration with colleagues, a trend that is becoming increasingly evident in everyday work environments.

The integration of AI as a thought partner and tool in the data science field is reshaping compliance and user interactions with data. Participants in the ongoing discussion are encouraged to share their personal experiences and journeys with AI within their companies and client engagements. A particularly insightful contribution comes from Cal Newport, author of “Deep Work,” who highlights research on students’ use of ChatGPT and generative AI.

Mark shared that many students leverage these tools as sounding boards for their ideas, often instructing ChatGPT to refrain from adding new content. This approach underlines the tool’s potential to enhance critical thinking rather than replace it, showcasing AI’s valuable role in fostering deeper analytical skills. Such discussions underscore the profound impact that AI can have on the data science landscape and its potential to catalyse more advanced thinking and collaboration.

Engaging in conversation can significantly enhance the process of tackling complex tasks, such as writing a thesis. Having a discussion partner enables individuals to identify gaps in their understanding, resulting in increased clarity and enhanced thought processes. However, many individuals struggle to utilise AI as an effective thought partner due to limited practical knowledge, often shaped by its portrayal in social media and news outlets. Recognising the importance of data governance serves as a crucial initial step toward effectively harnessing the benefits of AI in these contexts. Ultimately, fostering meaningful conversational engagement can dramatically elevate the quality of academic and creative endeavours.

The governance challenges associated with AI tools such as Microsoft Copilot and ChatGPT in corporate environments are significant. Despite Microsoft Copilot being powered by GPT-5.0, many users gravitate towards ChatGPT for its enhanced functionality, resulting in the phenomenon known as “shadow AI,” where employees utilise unapproved applications that may jeopardise data privacy and security.

This is particularly concerning when external tools, such as Gamma for slide generation, are used outside of the established governance framework. To address these risks, solutions are available, including specialised products starting at $60 per seat annually, designed to curb shadow AI activities and improve management of these challenges.

Figure 1 From Hype to Strategic Advantage

Figure 2 The GenAI Revolution is here.

Figure 3 The tipping point

Figure 4 AI’s Business Impact

The Impact of AI on Talent Transformation and Team Collaboration

The role of artificial intelligence (AI) in business and personal life significantly enhances productivity and problem-solving across various sectors; however, its implementation also poses challenges. Led by Mark, with contributions from Howard and participants such as Stefan and PG, the discussions drew parallels between AI’s evolution and the mobile phone revolution, focusing on key challenges including data governance and societal perceptions of AI, particularly generative AI, which some view as “cheating.”

Mark underscored the importance of adapting educational systems to embrace AI, as exemplified by the high school experience of his daughter, who navigates strict AI usage policies in humanities while engaging in coding applications in computer science. He also emphasised the need for open discussions within organisations about AI, aiming to overcome taboos surrounding its use.

In the past, teams would spend significant time reading lengthy documents, like a 400-page report, and then attempt to summarise key points before developing a project plan. This process often led to confusion and second-guessing about overlooked details. However, with the integration of AI, teams can now analyse extensive documents more efficiently. AI provides essential insights, allowing team members to ask questions and explore the content interactively. This shift enables better collaboration and a clearer understanding of strategies, freeing teams from the cumbersome task of reading and summarising large volumes of information.

The integration of AI into business solutions presents significant opportunities for enhancing collaborative efforts and utilising external resources. However, a primary challenge arises from the difficulty many neurotypical individuals encounter in translating their roles into effective standard operating procedures (SOPs) for AI use.

In contrast, Mark shared that he found that individuals on the autism spectrum, particularly those with Asperger’s, often possess a unique ability to articulate their processes with clarity, ultimately providing valuable insights that can optimise AI workflows. By recognising and leveraging these strengths, businesses can maximise the impact of AI technologies in their operations.

Leveraging AI for Effective Work Management

Leveraging AI technology can significantly alleviate the overwhelm individuals face when catching up on work after a week-long vacation. In today’s fast-paced environment, managing emails, Teams communications, and meeting notes can feel daunting; however, AI can streamline this process by scanning inboxes, generating meeting summaries from Teams recordings, and presenting insights for human review.

This approach not only reduces the time typically spent on these tasks—once taking several hours—but also enhances decision-making and action planning. To fully harness the benefits of AI, it is crucial to cultivate a workplace culture that embraces and understands its role in boosting productivity.

There appears to be a disconnect among team members regarding their understanding of the culture and processes, particularly in relation to the transcription of meetings. Many team members are reluctant to have meetings transcribed, which could aid in clarity and communication.

This leads to frustration as individuals struggle with an overwhelming volume of emails, often skipping important information. Furthermore, the challenge of preparing relevant emails for effective meeting participation contributes to feelings of being behind and overwhelmed, resulting in a sense of futility regarding work-life balance, such as the value of taking vacations.

Figure 5 After (Human and AI)

Figure 6 The Excitement Vs the Reality

Figure 7 The GenAI Iceberg

Figure 8 Navigating “AI Overwhelm” and Defining Strategy

The Opportunities in AI Implementation and Change Management

Mark emphasised the impactful implementation of artificial intelligence (AI) in both business and personal life. Participants, including Stefan from Telcom BCX and PG, showcasing his musical talents, contributed to discussions led by Mark and Howard. These conversations explored AI’s transformative role, comparing its rapid evolution to that of mobile phones, and underscored the need to integrate AI across diverse sectors to boost productivity and enhance problem-solving capabilities. Ultimately, the insights shared during the discussions highlighted the significant potential of AI to reshape various aspects of our daily lives and work environments.

Key concerns included Data Governance, the perception of generative AI as a form of cheating, and the need for educational systems to adapt and effectively integrate AI technologies. This was exemplified by the experiences of a high school student navigating strict AI usage policies in English and computer science classes, reflecting the broader societal shift towards AI adoption. The speaker emphasised the contrasting dynamics within organisations regarding AI discussions, indicating a flexible presentation style to facilitate engagement and understanding.

The process of integrating new technologies, such as AI, often leads to an initial decrease in labour productivity, similar to hiring a new employee. During the initial stages, it is essential to ensure that data is properly organised, governed, and clean, as the team becomes familiar with the tools and processes.

Measuring the return on investment (ROI) is crucial, even if it results in a temporary decline in productivity due to the time spent on training and adaptation. This learning phase requires defining roles, interviewing candidates, and extensive onboarding, paralleling the effort needed to train AI systems or personnel in new skill sets before they become truly productive.

To enhance productivity and trust in AI results, it’s vital to address ethical considerations, particularly those related to bias in model training. South Africa serves as a key example, where Western models often carry inherent biases that must be acknowledged, especially when comparing them to Asian models, which face similar data bias challenges. Furthermore, establishing a framework—referred to as AI RTiSM—is essential for understanding the integration of these elements. This process is complex, akin to managing a simultaneous office move, implementing a new CRM system, and merging with another company, emphasising the significance of effective change management and the potential opportunities that arise from navigating these challenges.

Figure 9 Implementation Hurdles and Resource Gaps

Figure 10 Measuring ROI and Managing Risks

The Implications of AI Implementation in the Organisation

The implementation of AI technologies often faces substantial challenges, resulting in a high rate of project failure. Reports from organisations like PMI and McKinsey suggest that anywhere from 80% to 97% of AI initiatives do not achieve their intended return on investment. A notable example is a major bank that has invested heavily in Microsoft’s Copilot yet struggles with user adoption.

This situation reflects a historical trend seen during technological transitions, such as the shift to desktop computers, which required managers to adapt to new roles that included direct engagement with technology. The current resistance to AI tools like Copilot primarily arises from fear, uncertainty, and difficulties associated with change management, factors that are frequently overlooked during deployments. Addressing these issues is essential for maximising the potential benefits of AI in organisations.

User engagement is vital for the successful implementation of technological solutions, as these initiatives often falter without active participation from those involved. The ongoing evolution of technology reveals that past efforts by professionals to integrate tools may have been unsuccessful due to a lack of user involvement. However, with the rapid advancements in technology, it is now possible to significantly enhance workflows in just a few months. Consequently, it becomes essential for users to remain informed about these improvements and to embrace adaptability, as relying on outdated methods can lead to inefficiencies and hinder progress.

Adaptability and curiosity are critical AI skills often overlooked in business environments, where employees and leaders are incentivised to manage risk and maintain stability rather than embrace change. This creates a gap between understanding the importance of AI and effectively implementing it.

Research highlights that 93% of AI pilot projects fail, and of the successful ones, only 30% leveraged in-house talent due to a lack of dedicated AI leadership and strategists. To overcome these challenges, companies must focus on empowering employees to think critically about how AI can enhance their processes, encouraging them to identify opportunities where AI can streamline operations and improve efficiency.

Understanding AI and its capabilities is crucial for organisations before embarking on large-scale AI projects. Many individuals face challenges with basic data tools, such as Excel, which makes it difficult to comprehend more advanced concepts like data warehouses or AI-driven data validation. Consequently, organisations risk wasting resources on AI tools, including multiple licenses for products like Copilot, without possessing the foundational knowledge required for effective utilisation. Therefore, conducting a readiness assessment is essential to identify the tipping point at which organisations can successfully integrate and leverage AI technologies to drive growth and efficiency.

Figure 11 The “Knowing-Doing Gap” in GenAI

Understanding the Importance of Knowledge Workers in the Organisation

To address the needs of a company with fewer than 500 knowledge workers, it is crucial to distinguish between knowledge workers and front-line workers. For instance, in a hospital with 5,000 employees, if 4,000 are solely using medical terminals without engaging in knowledge work, they can be excluded from this discussion.

Similarly, in industries like construction and mining, many workers may interact with technology but do not qualify as knowledge workers. Therefore, when focusing on the core group of knowledge workers, it is crucial for leadership to thoroughly understand the organisation’s direction and methodology to support and guide these employees effectively.

Figure 12 Accelerating Innovation and Efficiency

Figure 13 Think with the AI Framework

Compliance of Generative AI in Organisation

The THINK framework emphasises the importance of education when integrating generative AI into business, addressing common marketplace failures due to a lack of understanding among employees. Key issues include the need for comprehensive knowledge of compliance, data privacy, and effective data consolidation, such as centralising files within SharePoint for enhanced interaction with tools like Copilot. Additionally, successful adoption requires a cultural shift led by leadership, ensuring accountability and engagement with new technologies. To align AI initiatives with company goals, organisations should identify strategic objectives and start with small, manageable projects that demonstrate the value of AI integration.

Implications of E-mail Archiving and Privacy in the Context of AI

The issue of email archiving raises significant concerns regarding the privacy of former employees and the potential misinterpretation of their statements after they leave the company. Understanding the context of these comments is crucial, especially when evaluating archival information for purposes such as government audits.

Leadership must navigate the complexities of utilising this information constructively while being mindful of societal shifts in understanding and morality over time, as evidenced by historical injustices against Indigenous populations. Ultimately, the conversation emphasises the need for caution and consideration when using archived data, particularly since the original contributors are no longer present to clarify or defend their past remarks.

The concept of “human in the loop” emphasises the necessity for individuals to engage actively with AI-generated information rather than relinquishing responsibility for interpretation. Decision-makers must contextualise data, interrogate its relevance, and consider the evolving nature of information. As such, companies must establish clear guidelines for handling archived data and integrating this into their employee documentation and use cases. Ultimately, while AI can assist in summarising and providing insights, the responsibility for final decision-making rests with the individuals who must consider the surrounding context and implications of their choices.

The effective use of new tools is crucial for preserving institutional knowledge and enhancing decision-making processes within an organisation. To ensure critical information remains accessible, especially for new team members and during personnel changes, meetings will now be recorded and transcribed for future reference.

This approach encourages thorough inquiry and challenges assumptions, promoting engagement among colleagues in evaluating decisions made under pressure, which are often based on unreliable data. Additionally, the distinction between personal and corporate data is emphasised, highlighting concerns regarding an over-reliance on AI tools. Ultimately, this call for a more critical approach underscores the importance of not accepting information at face value to improve organisational resilience and decision-making quality.

The Role of Human Involvement in AI Collaboration and Decision-Making

The training framework discussed consists of two key ratios: 10-80-10 and 50-20-30, which illustrate the collaborative effort between humans and AI. The 10-80-10 model emphasises that for any task, 10% of the effort should come from the human. In comparison, 80% is handled by AI, with human involvement being critical in the final analysis and decision-making process to ensure accuracy against potential discrepancies in the data.

Meanwhile, the 50-20-30 model suggests that for a task requiring one hour, 50% of the time (30 minutes) should be allocated to upfront preparation, understanding objectives, and designing the approach, while the remaining 20% is dedicated to interacting with AI—engaging in back-and-forth dialogue, asking questions, and exploring ideas collaboratively.

The growing integration of AI in our workflows necessitates a proactive approach to responsibility, emphasising the importance of user verification and validation of information. Relying solely on AI outputs can lead to a dangerous dependency, much like how GPS can render individuals directionally impaired.

To mitigate this risk, individuals should prioritise the development of critical thinking skills by engaging with information more deeply and asking insightful questions. By fostering an understanding of AI tools, users can enhance their analytical capabilities, positioning AI as a valuable thought partner in their decision-making process. Ultimately, this balanced approach not only safeguards against cognitive decline but also empowers users to make informed decisions in collaboration with technology.

The necessity of critical thinking and active engagement when utilising technology, especially AI, cannot be overstated. Like teaching a child, it is crucial not to take answers at face value but to question and understand the underlying processes. Our interactions with it heavily influence the efficacy of AI; it can serve as a powerful tool or become limited due to our approach to it.

For instance, a person who seeks validation from ChatGPT without probing deeper illustrates how a lack of critical inquiry can diminish the effectiveness of AI. Ultimately, fostering a mindset of scepticism and curiosity is essential for maximising the benefits of technological advancements.

Reframing conversations is essential for fostering understanding and effective communication among participants. By acknowledging insightful questions and answers, a more productive dialogue emerges, as demonstrated by RJ and Paul, who navigate technical difficulties with both humour and openness.

This interaction underscores the key takeaway that the effectiveness of a system or idea relies heavily on the support it receives, highlighting that its intelligence is intricately connected to the quality of input and guidance provided. Ultimately, fostering an environment of collaboration and constructive dialogue can lead to deeper insights and successful outcomes.

Figure 14 Think with AI

Bias and Context in Image Generation

The importance of context in AI responses is pivotal for accurately capturing cultural nuances and ensuring diversity in representation. During the discussion, it became clear that while AI can be sensitive to various issues, it requires structured guidance to navigate complex societal landscapes, particularly in South Africa. A year ago, efforts to incorporate a diverse array of individuals primarily resulted in representations of darker-skinned individuals, raising concerns about the need for broader racial diversity.

This highlighted the unique biases inherent in certain AI models and tools used for these representations. Mark then reflected on their Canadian perspective, noting the differences in diversity-related challenges across regions. However, recent advancements in these tools have improved our ability to achieve the desired representation, emphasising the need to understand both their capabilities and limitations. Ultimately, fostering diversity in AI requires ongoing dialogue and adjustment to ensure all voices are genuinely represented.

The Challenges and Prospects of AI Adoption

The adoption of AI technology is expected to accelerate rapidly due to competitive pressures, contrasting with the slower historical adoption of previous technologies, such as the electric light bulb and computers. This shift is driven by a growing awareness of the potential losses associated with inaction.

Key to this transition will be the emergence of inspiring case studies that demonstrate AI’s capability to address significant societal issues, such as poverty. While there is great enthusiasm, success will require patience, clear objectives, and a realistic understanding of AI’s capabilities, alongside the continued necessity for human involvement in these projects. Current assessments suggest that transformative advancements may be about a year away, necessitating a grounded approach to implementation.

The integration of AI within companies necessitates a patient and aware approach to technology adoption. Many organisations still depend on outdated data platforms and legacy systems that are not widely understood, highlighting an urgent need for modern solutions. To effectively implement AI, it is essential to create a comprehensive strategy that encompasses governance, compliance, and privacy laws, while also identifying quick wins through smaller use cases. Additionally, investing in the training and nurturing of personnel is crucial to maximising the technology’s potential and facilitating a smooth transition into new systems. Ultimately, a thoughtful and thorough approach will help organisations leverage AI effectively and ensure a successful transformation.

The importance of comprehensive training in understanding AI is highlighted, emphasising the nurturing aspect that enables effective tool selection. For instance, a client used Copilot for six months and recognised the need for a specific tool to assist with contracting and legal reviews. By understanding their goals, we could guide them in evaluating suitable options. This analogy is likened to asking for a specific type of hammer, illustrating how clarity in requirements leads to the identification of the right technology, tools, processes, and systems to achieve desired outcomes.

Figure 15 Your Practical next steps

Figure 16 What you’ll gain from the Training

Building AI Habits: Adoption and Implementation

To successfully integrate AI tools like Copilot in a Microsoft environment, it is essential to foster a culture of sharing and support, especially for newcomers. Start by creating a dedicated AI implementation team, consisting of enthusiastic champions who can lead by example and highlight the time-saving benefits of these tools.

Rather than attempting to tackle the project on a large scale initially, focus on a small pilot group that can demonstrate the value of AI adoption. This approach encourages open communication and collaboration, helping to drive a cultural shift within the organisation as team members share their positive experiences and insights with colleagues.

Figure 17 Q&A Session

Leveraging AI in Agile Development and Project Efficiency

Traditional Agile practices often face significant challenges due to inefficiencies caused by excessive ceremonies and lengthy meetings. Participants in a recent discussion highlighted their frustration with the time wasted in these processes, illustrating the issue with the example of a renewable energy construction company where 75 out of 100 employees are involved in extensive meeting schedules. Some employees are dedicating up to 16 hours a week to updating Gantt charts, which hinders productivity and progress.

Stefan proposed that generative AI could play a crucial role in streamlining these processes, leading to smaller, more effective meetings and fostering a culture that integrates and articulates project information more efficiently. Ultimately, embracing such innovations could alleviate the burdens of traditional Agile practices and enhance overall team effectiveness.

The goal is to automate the transcription and summarisation of discussions, enabling Scrum Masters and stakeholders to track action items and updates easily. By leveraging AI to extract key points from conversations, individuals can efficiently access relevant information without needing to attend entire meetings.

This allows team members to focus on specific topics of interest, enhancing communication and context. Additionally, the reporting system will ensure that critical information is effectively rolled up to those responsible for updating charts and project metrics.

The Impact of AI and Mood Coding on Product Development and Design

AI has the potential to revolutionise the design process by significantly reducing the time required to transform concepts into interactive prototypes. By enabling designers and stakeholders to visualise their ideas quickly, AI facilitates a rapid transition from initial thoughts to tangible products within hours, rather than months.

This technology empowers individuals, especially those who struggle to convey their vision, to experiment and iterate on their designs more effectively, thereby enhancing creativity and communication. Furthermore, the integration of AI promotes efficient information flow among team members, ensuring that critical insights from designers, architects, and security experts are shared promptly, minimising delays caused by prolonged meetings. Ultimately, this streamlined approach not only fosters innovation but also enhances collaboration within design teams.

The prevalence of failures in AI implementation often stems from overly ambitious projects, highlighting the need for a more practical approach. During the recent discussion, participants emphasised the importance of experimenting and gradually building skills, much like how one would use a search engine to find information.

Organisations like Lucid and the Global AI Alliance (GAIL) are dedicated to supporting businesses in maximising resources, reducing workforce stress, and enhancing job satisfaction by providing affordable training sessions priced around $150. Ultimately, the goal is to transform conversations about AI into actionable strategies that foster improvement and effectiveness in the workplace.