The advent of Large Language Models (LLMs), such as OpenAI's ChatGPT, has revolutionized various fields through their ability to process, understand, and generate human-like text. These models are already being employed in a multitude of tasks, from customer service to content creation, and most certainly, we have only started to scratch the surface. However, the question arises: can LLMs engage in strategic thinking, a domain traditionally reserved for human experts? This article explores the capabilities and limitations of LLMs in the realm of strategy and expands on the rationale for a collaborative approach between human strategists and AI. 


    To understand whether LLMs can formulate a strategy, it will be useful to understand what strategy entails and how LLM models work.

     

    What does strategy entail? 

    Strategy involves long-term planning and decision-making to achieve specific goals. It requires analyzing complex variables, forecasting future trends, understanding competitive landscapes, and making decisions that balance short-term actions with long-term objectives.

    Strategic thinking is inherently dynamic, iterative, and adaptive, often necessitating a deep understanding of context, human behaviour, and nuanced judgment. The essence of strategy, as Dr. M. Porter pointed out, is about being unique. This uniqueness is derived from the understanding of context, a task that AI alone cannot accomplish. 

     

    How do LLM models work? 

    LLMs, such as ChatGPT, are trained on vast datasets. They learn to predict text by identifying and following patterns that appear most frequently in the data. This statistical approach means that the models are inherently biased towards producing the most common or frequent responses. 

    I ran a quick experiment to test this and asked ChatGPT and Copilot to tell me what strategy is, and here is what they produced: 

      • Strategy is a plan of action designed to achieve a long-term or overall aim…. (ChatGPT) 
      • Strategyis a deliberate plan or set of actions designed to achieve specific goals (Copilot) 

    These definitions certainly represent the most frequent concepts on the training dataset; however, they differ from the point of view offered by some of the most recognized experts in the strategy field.

    If strategy is about being unique, the most frequent answers on the training datasets won’t give us the tails of the distribution where uniqueness lies. In other words, we will get average rather than innovative or exceptional responses. 

    The sentence “the art of the possible,” attributed to Otto von Bismarck (19th Century), is commonly used in the political and consulting advisory sectors. This sentence suggests that rather than pursuing idealistic goals, focusing on achievable objectives and practical solutions is better. 

    Inspired by that quote and thinking of strategy as being unique, I’d argue that strategy is the art of the IMPOSSIBLE. It is about creating new ways to generate value that sometimes seems impossible to achieve. It is not about doing something average or good enough to compete. It’s about creating distinct value and amplifying it.

    Describing strategy as "the art of the impossible" underscores the essence of strategic thinking: transforming ambitious, seemingly unattainable goals into reality through creativity, innovation, resourcefulness, and determination. It's about envisioning and achieving outcomes that go beyond conventional expectations and limitations. 

    To illustrate the point, let me mention a couple of examples. When I was studying business many years ago, it was often pointed out that for an organization to succeed, it must choose between being the lowest-cost provider and being different; you cannot be both. The Blue Ocean approach to strategy showed us that, in fact, you can be actually both. 

     Business schools also usually stress that the only way to succeed is to “win,” meaning “by defeating the competition.” Yet nowadays, we see how many incredibly successful companies cooperate with direct competitors to generate more value.

    There are a multitude of examples of this, such as Apple, Google, and Microsoft, who are fierce competitors in various sectors (including operating systems), and yet they have collaborated to develop the FIDO Alliance, which is dedicated to creating open and free authentication standards to reduce reliance on passwords. In the airline industry: American Airlines, British Airways, and Qantas compete for passengers on many routes; however, they have formed an alliance called Oneworld, allowing them to coordinate schedules, share airport lounges and offer reciprocal frequent flyer benefits. In North America, BMW, General Motors, Honda, Hyundai, Kia, Mercedes-Benz, and Stellantis have joined forces to build a network of EV charging stations, and the list goes on; you get the point I’m trying to make.

    There are many ways to succeed; however, being unique instead of average is at the core of strategy. 

     

    Capabilities of LLMs 

    LLMs, including ChatGPT, exhibit remarkable capabilities in processing and generating human-like text based on vast amounts of data. They can provide insightful analyses, generate creative content, and even simulate decision-making processes to some extent. There are some specific strengths of LLMs that can be used in strategic contexts: 

    • Data Analysis: LLMs can quickly analyze large datasets, identify patterns, and extract relevant information. This capability is crucial for strategic planning, where data-driven insights are necessary for informed decision-making. 
    • Scenario Simulation: LLMs can simulate various scenarios and their potential outcomes. Users can explore multiple strategic options and their implications by inputting different variables. 
    • Trend Prediction: Leveraging historical data, LLMs can identify trends and forecast potential future developments. This predictive capability can aid in anticipating market shifts and preparing strategic responses. 
    • Knowledge Integration: LLMs can integrate vast amounts of information from diverse sources, providing a comprehensive view of complex issues. This holistic perspective is beneficial for strategic analysis. 

     

    Limitations of LLMs 

    Despite these capabilities, LLMs face significant limitations when it comes to true strategic thinking, which is the creation of distinct value. Once again, it's about striving for the tails of the distribution, not the median. 

    • Lack of Contextual Understanding: LLMs rely on patterns in the data they have been trained on. They lack a genuine understanding of context, which is crucial for nuanced strategic decision-making. Human strategists often rely on tacit knowledge, experience, and intuition, which LLMs cannot replicate. 
    • Adaptability and Creativity: While LLMs can generate creative content, their creativity is bounded by the data they have been exposed to. They may struggle with truly novel or out-of-the-box thinking that can redefine strategies. 
    Exploring the Potential of LLMs

     

    • Ethical and Moral Judgment: Strategic decisions often involve ethical considerations and moral judgments, which require a deep understanding of human values and societal norms. LLMs do not possess intrinsic values or ethics, potentially leading to decisions that may not align with human ethical standards. 
    • Dynamic Adaptation: Strategy requires continuous adaptation to changing circumstances. LLMs can process new data, but they lack the self-awareness and goal-oriented flexibility to dynamically adjust strategies in the way humans do. 

     

    LLM has a lot to offer if you know what the answer should be.  

    Exploiting the full potential of LLM for strategy planning requires great expertise. You need to have the expertise to discern what is valuable from what is not. 

    Asking an LLM to write a strategy for you will have limited value, and it could even be very risky. Expert human intervention is still crucial. 

    Human strategists bring a wealth of experience, intuition, and ethical considerations that LLMs cannot match. Below is a summary of key reasons why human strategy experts are indispensable: 

    Deep Contextual Understanding: Humans can comprehend and interpret complex contexts, including cultural, historical, and socio-political factors, that machines cannot fully grasp. This understanding is critical for making decisions that are not only effective but also appropriate and sensitive to specific contexts. 

    Ethical Decision-Making: Humans have the capacity to make decisions based on ethical principles and moral values. They can consider the broader implications of strategic choices on stakeholders, society, and the environment, ensuring that decisions align with ethical standards and societal norms. 

    Intuition and Experience: Experienced strategists rely on intuition and a deep well of experience to make quick, informed decisions in uncertain and rapidly changing environments. This human intuition is often derived from years of practice and cannot be replicated by algorithms. 

    Adaptive Creativity: Humans excel in adaptive and creative thinking, which is essential for developing innovative strategies that break from conventional patterns and create innovative value. They can envision and implement novel solutions that may not be evident from existing data alone. 

     

    Augmenting human strategic thinking with AI 

    Rather than viewing LLMs as autonomous strategists, it is more practical to see them as powerful tools that can augment human strategic thinking.  

    While Large Language Models like ChatGPT possess impressive capabilities in processing information and generating text, they fall short of replicating the complex, context-sensitive, and ethically nuanced nature of human strategic thinking. However, when used as tools to augment human intelligence, they can significantly enhance strategic planning and decision-making processes. 

    In my opinion, the near future of strategy likely lies in a collaborative approach, where human ingenuity and the computational power of LLMs are combined to navigate the complexities of the modern world. This synergy can lead to more informed, creative, and ethically sound strategies that leverage the strengths of both human and artificial intelligence. 

    At Stratford Group, we have a unique combination of technical and human expertise that can take your organization to the next level. If you have questions, let's talk! We’ll be happy to discuss how we can harness the power of technology and human ingenuity to deliver powerful strategies that will translate into extraordinary results for your organization. 


     

    About the Author

    HEADSHOT-Carlos Duran (low quality) Carlos Duran is a senior professional with over 25 years of experience in strategy formulation and execution. He helps organizations globally to design and implement strategies and business transformation initiatives across various industry sectors. His areas of expertise include digital transformation and the integration of ESG considerations into core business strategies.  Carlos also possesses in-depth experience in business process improvement, quality management systems, and project and business performance management. He holds an MBA from Newcastle University in the UK, a Master in Quality Management and Process Improvement from Cologne University of Applied Sciences, Germany, a Civil Engineering degree from UNT, Argentina, and specializations in Digital Strategy from MIT, USA and INSEAD, France. As a Senior Advisor at Stratford, Carlos helps organizations create and sustain long-lasting value. 

     

     

    You May Also Be Interested In: