How AI Large Language Models Can Change the Way We Strategise

LLMs have notable limitations when it comes to strategy formulation, but deliberately imperfect prompts can challenge conventional thinking.

There is a growing perception of an imminent occupational apocalypse as different aspects of managerial roles seem to be inevitably threatened by Large Language Models (LLMs) like ChatGPT.

LLMs are indeed replacing managerial tasks by speeding up routine activities – from formatting and data cleaning to streamlining information-seeking and aggregation processes and providing comprehensive summaries within specific domains. There are also numerous business-related use cases for LLMs, including generating ideas and marketing content, conducting traditional strategic analysis (such as SWOT and Porter’s Five Forces) and making forecasts.

An essential component of receiving high-quality responses from LLMs lies in writing appropriate prompts. For those who already possess a clear objective and seek guidance on how to achieve it, meticulously engineered prompts are of utmost importance.

However, a strategist does not solely desire a singular path to reach a goal; they rely on diverse perspectives. As much as strategy formulation is about asking the right questions, it is also about understanding what can be learned from posing imperfect yet thoughtfully crafted questions.

Does this mean that we should let LLMs take care of formulating strategies in organisations? To answer this question, we conducted an experiment comparing research findings presented at a recent strategic management conference with ChatGPT’s answers to similar research questions. Our main conclusion is that although LLMs offer numerous advantages in accelerating processes, there are still important limitations to their use in organisations, especially in the realm of strategy formation.

Successful strategies are not generic

LLMs can quickly provide a synthesis statement in response to specific prompts. However, such statements are publicly accessible, and any benefits derived from these ideas are likely to diminish rapidly. After all, LLMs develop synthesis responses based on an aggregate common-sense understanding and can only provide what appears to be best practice.

More importantly, depending on how prompts are formulated, LLMs may give the same generalised answer to many different questions. For instance, ChatGPT’s main response to our various research questions were often generic and aligned with “participative management”, recommending transparent communication, co-creation along with cross-functional collaboration, fostering a learning and innovation culture and employee alignment with organisational goals. Evidently, relying on generalised answers and conventional wisdom to formulate strategies is inadvisable. The more generic a strategy is, the less advantage it can generate.

Strategies result from synthesising various perspectives

Strategy is not the synthesis of all publicly available data in response to a prompt, nor is it a generic prescription. Organisations are unique and require their own distinct strategies. If all competitors pursued the same strategy, it would eliminate any differentiation advantage in competitive environments. Moreover, every organisation has different circumstances and understandings of the world, necessitating the development of a unique strategy.

Strategists must be able to look at the same reality and develop alternative perspectives. This enables them to gain a richer understanding of the context they are operating in and formulate a singular strategy that offers the best dynamic fit between the organisation and the external business environment. In contrast, human beings struggle to think outside the box. LLMs can help us challenge and transcend those limitations when it comes to strategy development.

Managers can develop strategies that best align with their organisation by using LLMs to first understand common practices and then explore alternative, more specific actions by using less precise and imperfect prompts. These questions can range from more open-ended to closed form, or highly specific to more general in nature. Framing the core strategic issues through multiple deliberately imperfect prompts, which are diverse and conflicting, allows us to perceive and challenge our cognitive limitations. A crucial aspect for strategists is to craft these diverse prompts in a way that elicits revealing answers from LLMs.

LLMs open a window to the world beyond the organisation

If organisations were run by a single strategist who had full authority and control over implementation, the use of LLMs could be beneficial. However, strategies are products of complex social processes in organisations that involve affective reactions, power relations and hierarchical structures with varying control over processes.

In such a social system, one way an organisation can achieve coherence is through a “dominant logic” – a shared understanding of the world among members that facilitates communication, decision-making and eventually organisational action. This shared understanding is reflected in structures, processes and culture, encompassing assumptions about the competitive landscape, best practices, status hierarchies and more.

However, it often becomes restrictive when such shared beliefs become too strong and homogeneous among organisational members. This can lead to groupthink biases, false assumptions and other misconceptions arising from selective information gathering and misinterpretation. A prominent example is Nokia’s continued commitment to mobile phones as products rather than platforms, which led to their competitive demise in this market.

LLMs can act as a check on the homogenisation process and a constant reminder of what is happening outside of the organisation. Are all competitors developing the same technology? Are consumers really interested in what the organisation is offering? Are there alternative ways of operating? While the answers provided by LLMs are likely to be far from perfect, and in some cases ludicrous, they serve as a reminder of the importance of the world beyond the organisation.

Use LLMs as jesters, not oracles

There is no doubt that LLMs are impacting the world of knowledge workers, introducing a form of hyper-efficiency and serving as all-knowing oracles that describe the past, present and future of a given business environment. While those who know how to use LLMs can reap the most benefits, the efficiency boost LLMs provide is not exclusive or unattainable. It is available to all, and, over time, it will become familiar to all.

LLMs are sources of diverse information, knowledge and perspectives. Asking LLMs the “wrong” questions can diversify and broaden a strategist’s thinking. Free from social constraints, LLMs can also mercilessly challenge organisational beliefs. Those who master the art of using them to disrupt the status quo, almost in a jester-like fashion, are likely to develop successful strategies that provide a competitive advantage to their organisations.