Title: The potential of utilising large language models in empowering start-ups to master risk management

Authors: Pascal Renz; Johannes Martin; Michael Flad

Addresses: University of Applied Sciences Esslingen, Kanalstraße 33, 73728 Esslingen, Germany; University of Tübingen, Geschwister-Scholl-Platz, 72074 Tübingen, Germany ' University of Applied Sciences Esslingen, Kanalstraße 33, 73728 Esslingen, Germany ' University of Applied Sciences Esslingen, Kanalstraße 33, 73728 Esslingen, Germany

Abstract: Entrepreneurs often grapple with managing risks inherent in innovation due to limited resources and the unpredictability of new ventures. Traditional risk management approaches require specialised expertise, time and data - assets that start-ups lack. This paper explores how large language models (LLMs) can enhance risk management processes for start-ups. By examining the four key phases of risk management, we assess the potential benefits of integrating LLMs within this process. Based on a conceptual synthesis of literature, we illustrate how LLMs could help overcome cognitive biases, identify risk patterns, streamline decision-making and enhance the development of risk mitigation strategies. While current limitations are acknowledged, we argue that LLMs might become essential tools for start-ups, enabling them to better manage the risks associated with innovation. Furthermore, we expand the current knowledge by providing a second, advanced framework for automated risk management based on interconnected LLM-bots.

Keywords: risk management; artificial intelligence; large language models; LLMs.

DOI: 10.1504/IJMED.2026.154728

International Journal of Management and Enterprise Development, 2026 Vol.25 No.3, pp.257 - 278

Received: 28 Jan 2025
Accepted: 12 Jun 2025
Published online: 12 Jul 2026 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article