Artificial intelligence and strategic decision-making: Evidence from entrepreneurs and investors

Felipe A. Csaszar; Harsh Ketkar; Hyunjin Kim

doi:10.1287/stsc.2024.0190

Citation: Csaszar, F. A., Ketkar, H., and Kim, H. (2024). Artificial intelligence and strategic decision-making: Evidence from entrepreneurs and investors. Strategy Science 9(4) 322–345. doi:10.1287/stsc.2024.0190

Cite

BibTeX file

Paper highlights

Strategic decision-making requires generating alternatives and evaluating them. This paper tests an early large language model on both tasks in two entrepreneurship settings. The model produced business plans that investors rated at least as favorably as entrepreneur-written plans, and its evaluations of other plans aligned with experienced investors’ judgments.

The broader argument is that AI can expand the search, representation, and aggregation behind strategy. It can generate many alternatives, apply several frameworks to the same problem, act as a critic, or simulate a set of viewpoints. Whether any firm gains an advantage depends on how widely those capabilities diffuse and how well the firm incorporates them into its decision process.

Study design

In the generation experiment, GPT-3.5 completed ten accelerator applications from their problem descriptions. A preregistered sample of 250 experienced investment evaluators blindly compared the AI- and entrepreneur-written versions across 2,500 evaluations.
In the evaluation study, the model scored 138 business plans from a startup competition using the same rubric as 137 venture-capital and angel investors who had supplied 541 evaluations.

Results at a glance

AI-generated plans scored 0.14 standard deviations higher and were five percentage points more likely to receive an acceptance recommendation. The advantage was concentrated among applications the accelerator had rejected.
AI scores correlated 0.52 with the average investor scores and explained about one-quarter of their variation.
Agreement between the AI and the investor panel was greater than agreement among individual investors, though neither benchmark establishes that the underlying judgments were correct.

Why it matters

The paper tests both generation of strategic alternatives and evaluation of those alternatives.
AI’s generation advantage appeared mainly among rejected ventures, suggesting that an LLM raised the floor of weaker applications more than the ceiling of strong ones.
The 0.52 evaluation correlation measures agreement with experienced investors, not objective truth.
Strategic impact depends on complements such as proprietary information, workflow design, verification, and the organization’s ability to act.

How to use this paper

Cite this for

Empirical tests of LLMs generating and evaluating entrepreneurial strategies.
Evidence that GPT-3.5-generated accelerator applications were rated at least as favorably as entrepreneur-written applications in the study setting.
A framework connecting AI to strategic search, representation, and aggregation.

Useful for teaching

The difference between generating strategic alternatives and evaluating them.
Why agreement with experienced investors is useful evidence but not the same as objective truth.
How AI’s strategic value depends on complements such as proprietary data, workflow design, and verification.

Careful claim

In the accelerator and startup-competition settings studied, GPT-3.5 performed comparably to humans on bounded generation and evaluation tasks; the result does not show that AI can make all strategic decisions or that investor consensus is ground truth.

Abstract

This paper explores how artificial intelligence (AI) may impact the strategic decision-making (SDM) process in firms. We illustrate how AI could augment existing SDM tools and provide empirical evidence from a leading accelerator program and a startup competition that current Large Language Models (LLMs) can generate and evaluate strategies at a level comparable to entrepreneurs and investors. We then examine implications for key cognitive processes underlying SDM—search, representation, and aggregation.

Our analysis suggests AI has the potential to enhance the speed, quality, and scale of strategic analysis, while also enabling new approaches like virtual strategy simulations. However, the ultimate impact on firm performance will depend on competitive dynamics as AI capabilities progress. We propose a framework connecting AI use in SDM to firm outcomes and discuss how AI may reshape sources of competitive advantage. We conclude by considering how AI could both support and challenge core tenets of the theory-based view of strategy. Overall, our work maps out an emerging research frontier at the intersection of AI and strategy.