Beyond the Algorithm: The Human Touch in Business Simulation

Author: Denis Avetisyan

New research highlights the critical role of human instructors in maximizing learning outcomes when artificial intelligence is integrated into business simulation games.

This paper examines how the SECI model explains the interplay between AI-driven knowledge combination and human-led socialization in experiential learning environments.

Despite growing enthusiasm for artificial intelligence in education, the specific roles AI can play in fostering complex experiential learning remain underexplored. This paper, ‘AI Combines, Humans Socialise: A SECI-based Experience Report on Business Simulation Games’, investigates the integration of generative AI into a business simulation designed for engineering students. Findings reveal that AI primarily supports the $\text{SECI}$ model’s ‘Combination’ phase – efficiently synthesising explicit knowledge – while crucial processes like tacit knowledge acquisition and social learning continue to rely on human interaction and instructor guidance. How can instructional designs be refined to better leverage AI’s strengths while preserving the essential role of human mentorship in developing practical wisdom and competency?

The Dichotomy of Knowing: Explicit Knowledge and Its Limitations

Historically, institutions of higher learning have largely functioned as conduits for explicit knowledge – the facts, figures, and formulas easily articulated, documented, and measured through standardized assessments. This emphasis on codified information stems from a long-held belief that knowledge is best conveyed through systematic instruction and objective evaluation. Curricula are frequently structured around lectures, textbooks, and exams designed to verify comprehension of defined concepts. While undeniably important, this traditional approach often prioritizes what can be easily taught and tested, sometimes overshadowing the development of more nuanced and practical understanding. The very structure of academic disciplines encourages the fragmentation of knowledge into discrete, assessable units, reinforcing the primacy of explicit content in the educational process.

Formal education systems frequently concentrate on the conveyance of explicit knowledge – facts, figures, and established principles easily documented and evaluated. Yet, a significant dimension of competence resides in tacit knowledge, encompassing the nuanced understanding gained through experience, the ability to make sound judgments based on intuition, and the application of practical wisdom to navigate complex situations. This often-overlooked realm of learning isn’t readily transferred through textbooks or lectures; instead, it develops through practice, observation, and mentorship. Consequently, individuals may possess theoretical understanding without the capacity to effectively apply it, highlighting a critical gap between knowing what and knowing how – a deficiency that can impede problem-solving and innovation in real-world contexts.

The disparity between learned theory and practical application often leaves students unprepared for the complexities of real-world challenges. While higher education excels at imparting explicit knowledge – facts, formulas, and established procedures – it frequently overlooks the development of tacit understanding. This implicit knowledge, honed through experience and intuition, is critical for navigating ambiguous situations where textbook solutions are insufficient. Consequently, graduates may possess a strong grasp of what to do, but struggle with how to do it effectively when faced with novel problems, incomplete information, or unpredictable circumstances. Bridging this gap requires educational approaches that prioritize experiential learning, critical thinking, and the cultivation of practical wisdom alongside traditional academic rigor, ensuring students are not merely knowledgeable, but truly capable.

Experiential Learning: The Genesis of Tacit Understanding

Experiential learning is a process wherein tacit knowledge – the skills, insights, and understandings difficult to articulate – is developed through a combination of direct experience and reflective practice. Unlike explicit knowledge, which is easily documented and transferred, tacit knowledge is deeply embedded in individual experience and requires active engagement to cultivate. This process moves beyond theoretical understanding by providing opportunities for learners to apply concepts in practical situations, analyze the outcomes of their actions, and refine their understanding through iterative cycles of practice and reflection. The resulting tacit knowledge base enables more nuanced and effective performance in complex, real-world scenarios where standardized procedures are insufficient.

Business simulation games function as controlled environments where students apply managerial principles and make decisions with incomplete information, replicating the uncertainty inherent in real-world business scenarios. These games typically require participants to manage virtual organizations, encompassing functions like production, marketing, and finance, and to compete against other teams or AI opponents. This practice allows students to develop crucial skills including strategic planning, resource allocation, risk assessment, and collaborative problem-solving, all within a safe context where the consequences of errors are minimized. Performance is often evaluated through key performance indicators (KPIs) such as profitability, market share, and return on investment, providing quantifiable feedback on decision effectiveness and promoting iterative learning.

Research indicates that integrating Artificial Intelligence (AI) into business simulation games substantially improves their educational efficacy. Specifically, our findings demonstrate that AI can facilitate knowledge combination – the process by which participants integrate previously disparate information to formulate effective strategies. This is achieved through AI-driven mechanisms that provide dynamic feedback, personalized learning pathways, and adaptive challenges within the simulation environment. The AI doesn’t replace managerial decision-making but rather augments it by highlighting knowledge gaps and suggesting potential insights, leading to improved learning outcomes and a more robust understanding of complex business principles.

Augmenting Simulation with Artificial Intelligence: A Precision Approach

Several artificial intelligence tools, including NotebookLM, ChatGPT, Gemini, and ElevenLabs, are being integrated into simulation environments to function as pedagogical assistants. These tools utilize natural language processing to respond to student inquiries within the simulation, offering explanations of concepts, clarifying rules, and providing hints related to strategic decision-making. The AI assistants can be programmed to deliver tailored guidance based on a student’s specific actions and expressed knowledge gaps, effectively personalizing the learning experience. This capability extends beyond simple question answering to include the provision of contextualized support and the facilitation of deeper exploration of simulated scenarios.

AI-driven simulation platforms are now capable of generating dynamic conversational agents that facilitate realistic interactions with students. These agents utilize large language models to process student input and formulate responses mimicking human dialogue, allowing for open-ended questioning and scenario-based learning. Crucially, these agents provide immediate feedback on student decisions within the simulation, evaluating actions against pre-defined criteria or established best practices. This feedback isn’t limited to simple correct/incorrect assessments; agents can offer explanations of reasoning, suggest alternative approaches, and highlight the consequences of specific choices, all within the context of the simulated environment. The system’s ability to deliver this immediate, personalized feedback loop is a key component in accelerating skill development and knowledge retention.

Personalized support within AI-powered simulations demonstrably increases student engagement and facilitates exploration of a broader spectrum of strategic choices, thereby accelerating the development of tacit knowledge-the understanding gained through experience. Research indicates this approach augments instructor capabilities by providing scalable assistance, but does not obviate the need for human educators to cultivate competencies and phronesis, a practical wisdom involving ethical judgement and skillful action. While AI can offer immediate feedback and guidance, the development of nuanced understanding and ethical reasoning continues to require human mentorship and critical evaluation.

The SECI Model: A Framework for Knowledge Dynamism

The SECI model offers a compelling explanation for how knowledge is not simply transferred, but actively created within organizations and individuals. This framework posits that knowledge exists in two forms: tacit – the kind difficult to articulate, residing in experience and intuition – and explicit, easily documented and shared. Knowledge creation begins with Socialisation, where individuals learn through observation and shared experiences. This tacit knowledge is then made explicit through Externalisation – articulating insights into concepts and models. The Combination stage integrates these explicit knowledge components with existing information, fostering new knowledge systems. Finally, Internalisation occurs when this combined knowledge is absorbed and becomes part of an individual’s tacit knowledge base, completing the cycle and enabling further innovation. Understanding this cyclical process is crucial for effectively managing and cultivating knowledge within any learning environment.

Artificial intelligence offers distinct advantages across each phase of knowledge creation as outlined by the SECI model. The process begins with Socialisation, where AI can observe and record the actions of experts, capturing nuanced behaviours often unstated in formal training. This observed data then supports Externalisation, converting implicit, experiential knowledge into explicit, codified forms, such as best-practice guides or decision-making algorithms. Crucially, AI excels at Combination, rapidly synthesizing diverse information sources – data, codified knowledge, and observed patterns – to generate novel insights and solutions. Finally, through personalized learning pathways and simulations, AI facilitates Internalisation, allowing individuals to integrate this newly combined knowledge into their own understanding and skillset, thereby completing the cyclical process of knowledge creation and refinement.

AI-powered simulations offer a unique opportunity to expedite knowledge development by actively supporting the cyclical SECI model, fostering both explicit and tacit understanding. These simulations don’t simply deliver information; they create environments where learners can combine existing knowledge – the ‘Combination’ stage – in novel ways, drawing on observed behaviors and codified insights. Recent research demonstrates the practical application of this within business simulation games, where AI effectively synthesizes player decisions and market responses to generate new strategic insights. This accelerated combination of knowledge isn’t merely about faster learning; it cultivates a deeper, more nuanced understanding, ultimately leading to more effective and adaptable skill sets.

Cultivating Phronesis: The Zenith of AI-Enhanced Learning

The enduring purpose of education extends beyond the accumulation of facts and figures; it fundamentally centers on cultivating phronesis, a concept often translated as practical wisdom. This isn’t merely intellectual knowledge, but rather the capacity to discern the ethically sound course of action in complex situations, informed by deeply held values and sound moral judgment. Developing phronesis requires an ability to apply learned principles to real-world challenges, navigating ambiguity and considering the consequences of decisions – a skillset vital for responsible citizenship and effective leadership. The emphasis on phronesis therefore positions education as a process of character formation, shaping individuals not just as repositories of information, but as thoughtful, ethical agents capable of navigating the complexities of life.

AI-driven simulations, structured around the SECI model of knowledge creation – socialization, externalization, combination, and internalization – offer students uniquely challenging learning environments. These aren’t simple exercises in recall, but immersive scenarios designed to confront learners with ethically ambiguous situations requiring nuanced judgment. By actively engaging with these simulations, students move beyond theoretical understanding to apply knowledge in dynamic, real-world contexts. The SECI framework ensures that learning isn’t merely individual, but a collaborative process where insights are shared, refined, and ultimately integrated into personal frameworks for responsible decision-making. This approach prioritizes the development of practical wisdom by demanding that learners articulate their reasoning, consider diverse perspectives, and justify their choices within complex, simulated realities.

The integration of AI-driven learning experiences is designed to cultivate not just knowledgeable individuals, but also critical thinkers capable of responsible action. Studies reveal this immersive approach actively promotes the development of sound judgment, enabling students to navigate complex ethical dilemmas and make well-considered decisions. Ultimately, this process prepares them for leadership roles demanding both competence and integrity; however, research consistently demonstrates that artificial intelligence serves as a powerful tool, but not a replacement for dedicated human mentorship in nurturing these vital competencies and fostering the development of practical wisdom, or $phronesis$ .

The study highlights a fundamental distinction between the capabilities of artificial intelligence and human cognition within experiential learning environments. While AI excels at processing and combining explicit knowledge – readily codified rules and data within the business simulation – the facilitation of tacit knowledge remains firmly within the human domain. This echoes Edsger W. Dijkstra’s assertion: “It’s not enough to show that something works; you must show why it works.” The simulation demonstrably works due to AI’s computational power, but the why – the nuanced understanding, the development of practical wisdom, and the socialisation crucial for converting information into actionable insight – necessitates human guidance. The instructor’s role, therefore, transcends mere technical support; it embodies the transmission of contextual understanding and the fostering of collaborative learning, elements AI cannot yet replicate.

The Path Forward

The observed partitioning of labor – artificial intelligence excelling at the manipulation of codified knowledge, while human guidance remains essential for the cultivation of experiential understanding – is not merely a practical observation. It hints at a fundamental asymmetry. The simulation games, however elegantly programmed, are ultimately bounded systems. True ‘wisdom’, if such a concept holds merit, resides not in the processing of data, but in the extrapolation of principles to novel, unforeseen circumstances. The current work, therefore, does not represent a solution, but rather a precise delineation of the problem.

Future inquiry should move beyond assessing whether artificial intelligence can assist in experiential learning, and focus instead on defining the limits of its contribution. Can algorithms be constructed that reliably identify and signal opportunities for tacit knowledge transfer? Or is such discernment inherently reliant on shared context, intuition, and the nuanced reading of human behavior – qualities that, at present, appear non-computable? The pursuit of ‘artificial wisdom’ may prove a category error, a misapplication of logical principles to a domain governed by different laws.

The ultimate metric of success will not be the efficiency of knowledge transfer, but the demonstrable improvement in decision-making under genuine uncertainty. Until that standard is met, the human instructor remains not a replaceable component, but the indispensable fulcrum upon which practical learning balances.

Original article: https://arxiv.org/pdf/2602.20633.pdf

Contact the author: https://www.linkedin.com/in/avetisyan/