From Wall Street trading to the structured complexities of warehouse logistics, artificial intelligence has continually demonstrated its superiority over human capabilities in various sectors. The latest research led by Texas McCombs ventures into a realm previously unexplored, where AI competes against humans in high-level business strategy-making. Harsh Ketkar, an assistant professor of management, has found that AI not only matches but often exceeds the capabilities of humans in enhancing the speed, quality, and scope of strategic analysis. In tests involving human creators and evaluators of business strategies, conducted alongside Felipe Csaszar from the University of Michigan and Hyunjin Kim from INSEAD, AI showcased its prowess, equally challenging or surpassing its human competitors.

AI’s capabilities extend beyond drafting strategies; it is proficient in critiquing existing strategies and proposing viable alternatives. Ketkar highlights AI’s potential as a strategic sparring partner, capable of generating innovative ideas that human strategists may not easily envision. This can be particularly beneficial for executive teams, aiding them in overcoming the often invisible barriers that obstruct strategic insight, thus enhancing decision-making processes across the board.

This pioneering study was inspired by the financial trading sector, where algorithms now dictate 78% of trading decisions, a realm once dominated by human intuition and experience. This led to a curiosity about AI’s potential in strategic decision-making—a field that demands a deeper, more nuanced level of analysis and judgment, skills that are traditionally cultivated through extensive professional experience. How effectively could machines emulate or even improve upon these sophisticated tasks?

Ketkar and his team embarked on two experiments to explore this question. The first experiment involved collaborating with a European startup accelerator, examining business plans that were either accepted or rejected between 2021 and 2022. Using GPT 3.5, they created AI-generated versions of these plans, which 250 seasoned investors and managers evaluated. These evaluators rated the AI-generated plans higher across multiple criteria, highlighting a greater likelihood of these businesses being accepted into the accelerator and showing increased interest in initiating meetings and investing in these startups.

The second experiment reversed the roles: AI was used to assess business plans instead of creating them. The team partnered with a prestigious business school’s startup competition and selected 138 all-text business plans from previous competition cycles. Each plan was evaluated by GPT, using criteria that mirrored the real-world judging process. The results were astonishing, with AI’s evaluations correlating strongly with those of experienced human judges, often surpassing them in predicting the ultimate success of business plans.

These findings suggest a transformative potential for AI in strategic planning. Venture capitalists could use AI to streamline the evaluation of business proposals, potentially accelerating the decision-making process for startup funding. Similarly, accelerators could enhance their application review processes, and entrepreneurs could employ AI to refine and test their strategic plans, benefiting from its capacity to simulate complex scenarios and propose innovative solutions.

As AI becomes more integrated into strategic planning, it could give companies a competitive edge, enhancing their overall performance. This shift could lead to the development of new strategic frameworks, fundamentally altering the practice and perception of strategy-making. The advent of AI in strategic roles suggests a future where strategic decision-making is about human insight and the collaborative interplay between human creativity and machine intelligence. This dynamic could redefine the landscape of business strategy.

More information: Felipe Csaszar et al, Artificial Intelligence and Strategic Decision-Making: Evidence from Entrepreneurs and Investors, Strategy Science. DOI: 10.1287/stsc.2024.0190

Journal information: Strategy Science Provided by University of Texas at Austin