Manufacturing AI From A Cultural & Cognitive Lens

Pooling learnings that Subrata Sen and I have compiled working with major operators trying to leverage the promise of Artificial Intelligence across Manufacturing Operations. Have a read, and let us know how your experience compare

*****************

Over last 2 years, AI has been a central topic of my discussions at industrial talks, as a panelist and in numerous workshops - with topics ranging from the potential benefits (primary focus) to concerns about job displacement (usually the last question). This broad discourse indicates that AI is poised to eventually impact all industries and functions. Within a manufacturing setting, this post examines the trajectory of technological shifts and industrial adoption over the past two decades, highlighting entrenched resistance while forecasting a trajectory for the current AI discussions. The recurring theme of fearing obsolescence and a loss in competitive advantage provides a view into leadership’s difficulty in adapting to new technological eras. It suggests that a significant barrier to AI adoption could be the persistence of outdated business models, presenting major constraints to leadership options and providing some clues from history on where to start.

?

Evolving Industrial Data and… Cognitive Biases

To assess AI's impact, understanding the evolution of industrial data is crucial. For over 50 years, the process industry has used data primarily for compliance and safety, driven by regulations. Initially, data’s role was to ensure operational licenses. The concept of "Data is the New Oil," highlights that data, like oil, must be refined and processed to be valuable. Recent debates have centered on data ownership, with consensus now recognizing that data generators (owner/operators) retain ownership, which is essential for training AI models. We segment the industrial data and technology evolution into four epochs. Each epoch marked an evolutionary stage on data and technological maturity, with the final epoch being AI.

• 2000-2015: Era of Visibility

From 2000 to 2015, the focus in industrial data environments shifted towards improving visibility and reporting. This era emphasized developing systems and tools for better data access and transparency, driven by the need for enhanced compliance, operational efficiency, and faster decision-making. Key advancements included sophisticated data visualization tools and dashboards that enabled real-time monitoring and insights into performance metrics, helping businesses identify trends and anomalies. There was also a significant push towards integrating data from multiple sources for a comprehensive operational view. Additionally, data storage solutions advanced, accommodating large volumes of data and supporting detailed reporting. Regulatory requirements played a crucial role, as companies aimed to meet compliance standards through accurate data capture and reporting. Overall, this period marked a transition towards greater data transparency and accessibility, laying the groundwork for more advanced data analytics and intelligence, and setting the stage for the integration of artificial intelligence in later years. This optimistic era had placed its bet on data, but could not deliver on the value. Practitioners still carried the day with their expertise and experience.

• 2015-2019: Era of Bold Initiatives

With a buzz around standardization, organizations shifted their focus to derive value from balanced scorecards, to capture and integrate business flows across functions. Application vendors, lacking standardized mandates, controlled data models and constructs within their systems. This led to large, expensive transformation projects aimed at integrating data and business processes. These multi-year programs were costly and complex, with few organizations able to successfully maintain and align them with evolving business demands. In the latter part of this era, data analytics experienced significant growth. However, many of these efforts failed due to fundamental issues in manufacturing data availability, quantity and quality, along with the requirement to satisfy a cross-functional set of stakeholders. This era is marked by an abundance of?Siloed Mentality

• Definition:?A prevalent organizational mindset in which departments function in isolation, hindering collaboration and integration, and thereby obstructing comprehensive AI adoption and cross-functional insights.
• Characteristics: Disjointed communication, absence of unified objectives, and resistance to technologies that necessitate inter-departmental coordination.

·?????? 2020 - 2023: Era of Myth Busters

During the pandemic, manufacturing facilities had to quickly adapt to remote data sharing, collaboration, and decision-making due to enforced social distancing. Historically, the industry had been reluctant to adopt remote operations, preferring in-person collaboration for decision-making. Improved connectivity, remote collaboration tools, and cheaper sensors facilitated rapid transformation, enabling safe operations with minimal on-site staffing. As markets recovered, manufacturing facilities ramped up on-site staffing but not to the pre-pandemic levels. This new model emphasized increased data usage, connectivity with remote experts, and consolidation of workforce roles across various sites. The industry also faced the challenge of losing experienced talent through retirements and exits, prompting a focus on enhancing application intelligence to capture and utilize this knowledge. The shift towards cross-functional insights and optimization began to break down traditional organizational and technological silos. Although the digital transformation was uneven, with some operators embracing new technologies more readily than others, it became clear that digitization and sensorization of plants were essential. The transformation's benefits were not uniformly felt, with digital laggards experiencing less impact from the changes. This era is marked by a few key mindsets:

• Risk Aversion: Definition: A strong emphasis on avoiding risks related to new technologies, resulting in a reluctance to embrace AI solutions. Characteristics: Preference for traditional methods, reluctance to explore innovative technologies, and a focus on maintaining stability rather than seeking progress.
• Compliance-Driven Culture: Definition: A mindset focused on adhering to regulatory and compliance standards, often at the expense of adopting innovative technologies like AI. Characteristics: Prioritization of compliance over innovation, concern about potential regulatory issues, and hesitancy to implement AI due to perceived compliance challenges.

·?????? 2023 – Future: The AI Era

Universal appeal of ChatGPT has transformed human-machine interactions by enabling natural language conversations, making technology accessible and user-friendly for everyone. This advancement allows individuals to retrieve and interpret information without needing expert-level knowledge or extensive training. This represents a significant democratization of data. The next phase of this technology—where machines can think, learn, and provide context for actions—promises to be revolutionary. It will empower less experienced practitioners to tackle complex issues independently, reducing reliance on a shrinking pool of experts. This shift could lead to fewer generalists on the shop floor and pave the way for autonomous systems. Though full realization of this potential might take years, the current focus should be on building trust in AI recommendations and preparing for this transformative change. This era is currently characterized by several distinct mindsets:

• Legacy Thinkers: Definition: Individuals or organizations that cling to traditional methods and resist change, viewing new technologies as risky or unnecessary. Characteristics: Reliance on past successes, reluctance to adopt innovative approaches, and a preference for maintaining established practices.
• Short-Term Focused: Definition: Emphasizes immediate results and short-term gains over long-term strategic planning, avoiding investments in technologies that might not provide quick returns. Characteristics: Preference for incremental improvements, avoidance of large-scale transformation, and skepticism towards AI due to perceived risks and uncertainties.
• Fear of Disruption: Definition: A mindset driven by concerns that AI and other disruptive technologies could threaten existing jobs, processes, and market positions. Characteristics: Resistance to change due to worries about job security, potential obsolescence of skills, and disruption of established workflows.
• Technology as a Cost Center: Definition: Viewing technology investments, including AI, primarily as expenses rather than opportunities for creating value and gaining a competitive edge. Characteristics: Focus on minimizing costs, reluctance to invest in technological innovation, and underestimating the strategic benefits of AI.
• Overconfidence in Human Expertise: Definition: Belief in the superiority of human judgment and expertise over AI-driven insights and automation, leading to resistance to AI adoption. Characteristics: Valuing human decision-making over data-driven insights, underestimating AI capabilities, and hesitance to integrate AI into decision-making processes

?In Conclusion

Over time, technology advancements has led to adaptive changes in Enterprise operating models, but these adjustments were typically minor and did not significantly alter established business models or cultures. To fully realize AI’s potential, companies must reassess their cultural and cognitive frameworks. Maintaining the status quo could lead to significant competitive disadvantages. Therefore, a deliberate shift in organizational culture, spearheaded by senior leadership, is essential. Transformative initiatives of this nature necessitate the establishment of a dedicated Transformation Office, ideally led by the CEO or COO. Leading organizations are already investing proactively in AI infrastructure and talent. To remain competitive, others should consider initiating similar transformative efforts.

Sandeep Chandran

Industrial AI

?#Leadership #BusinessGrowth #ContinuousLearning #AI #DecisionMaking #ExecutiveDevelopment #CorporateStrategy #Innovation #Csuite #ValueCreation #Technology #OrganizationalGrowth #ChangeManagement #ReflectiveLeadership #StrategicThinking #LearningCurve #CompanyGrowth #BusinessStrategy #Adaptation #CausalAI