Adaptive Agent-Based Business Automation Using Artificial Intelligence for Autonomous Task Coordination

Abstract

This study aims to examine the potential of multimodal artificial intelligence (AI) for adaptive business-process automation and autonomous task coordination. It focuses on how different forms of workflow information can support the automated understanding of business activities. The study examines multimodal artificial intelligence performance in standard operating procedure generation and evaluates AI capability for workflow demonstration segmentation as a foundation for autonomous task coordination. The findings indicate that incorporating richer workflow information improves SOP-generation performance across the evaluated AI models. Combining textual intent with visual keyframes and action traces provides a more comprehensive representation of workflow activities and supports more effective procedural understanding. The workflow segmentation findings further demonstrate differences in model capability for identifying and organizing individual business-process activities. The results indicate that multimodal information can enhance AI-supported workflow understanding and provide useful capabilities for intelligent business-process automation. The study demonstrates that multimodal AI capabilities can provide an important foundation for adaptive agent-based business automation and autonomous task coordination. Integrating workflow objectives, procedural information, visual context, and interaction evidence may support future systems capable of more informed task planning and coordination. However, the findings are based on benchmark outcomes and do not represent independent experimental results from a newly implemented agent framework.