News —

In the paper “,” Csaszar and his coauthors explore how artificial intelligence may impact the strategic decision-making process of firms. Through empirical studies involving a leading accelerator program and a startup competition, they demonstrate that current large language models can generate and evaluate strategies at a level comparable to entrepreneurs and investors participating in these competitions.

The paper uncovers several factors that make AI tools particularly promising for strategic decision-making: their ability to process vast amounts of information quickly, their capacity to generate and evaluate numerous strategic alternatives, and their potential to enhance existing strategic decision-making frameworks.

Csaszar shared a few insights from the research in the following Q&A. 

What motivated you to explore the impact of AI on strategic decision-making processes?

My interest in AI and strategy dates back to my undergraduate years in computer science and my time as a startup founder.  When ChatGPT was released in late 2022, I saw an opportunity to test a controversial idea I had presented at a strategy conference — that AI might be capable of making strategic decisions. Several of my colleagues had dismissed this possibility, arguing that strategy required uniquely human traits like creativity, judgment, and nuance, but GPT's capabilities suggested this future might be closer than we thought.

Why did you choose to specifically focus on AI integration in the entrepreneurship and investment field?

We chose startup competitions as our research setting because they allow us to measure two fundamental aspects of strategic decision-making: strategy generation and strategy selection.  In this context, we could test AI's ability to generate business strategies from initial problem statements and compare its evaluation of business plans against the judgments of experienced venture capitalists. This provided a realistic experimental setting to assess AI's strategic capabilities.

What are the most significant barriers to successful integration of AI?

The integration of AI faces similar challenges to previous technological revolutions, like the internet. Just as with the internet's adoption, which took decades despite its clear potential, AI integration requires multiple pieces to fall into place: technological infrastructure, business process adaptation, workforce training, and regulatory frameworks.

Even if AI development stopped today, it would likely take businesses more than a decade to fully integrate current capabilities.

Based on your findings, what ethical considerations should companies keep in mind when implementing AI-driven decision-making?

The primary consideration should be ensuring AI doesn't harm people, which often means keeping humans in the loop until AI systems prove themselves significantly safer than human decision-makers.  

Beyond safety, companies should focus on how AI can help democratize access to services that are currently limited to a privileged few — for instance, using AI to provide personalized education that previously required expensive human tutors.

In terms of scalability and competitive advantage, how can AI-driven decision-making provide an edge to new companies and their investors?

During this period of rapid technological advancement, companies that can adapt quickly and learn continuously will gain the most advantage. The winners will be organizations that are agile enough to implement new ideas rapidly, learn from mistakes, and iterate quickly. Eventually, when technology stabilizes, competitive advantage will shift to those with complementary assets like proprietary data or strategic locations.