23079: GenAI-Empowered Problem-Solving: Innovating in Times of Agentic AI

This PDW provides an introduction to agentic AI, currently one of the most exciting areas of AI research. We introduce attendees to common frameworks like crewAI or AutoGen used to implement agentic AI systems. Agentic AI describes systems composed of “agents”—artificially intelligent entities that can interpret their surroundings and take actions to achieve specific objectives or tasks (Jiang et al., 2024). These systems represent a significant advancement, enabling the creation of artificial teams or organizations made up of individual AI instances (Park et al., 2023). Their applications span a wide range of domains, including the natural sciences, engineering, and the social sciences (Cheng et al., 2024). While multi-agent systems have been a focus of research in computer science for years, the advent of LLMs has expanded their capacity to address diverse challenges (Xi et al., 2023). These systems can function autonomously or integrate into human organizations, with real-world uses already evident in fields such as software development, sales and healthcare. After visiting the session, you will know what agentic AI refers to, how it works, what opportunities and challenges it presents to organizations, and what the wider implications of agentic AI for organizational decision-making and problem-solving are. We provide a broad panel discussion with experts from diverse contexts covering distinct levels of interaction (e.g., societal, organizational, innovation process) and areas of concern (e.g., scope, practical and theoretical implications).