Human-AI symbiosis: The best approach for AI implementation in business decision-making in complex systems

In today's business landscape, there is significant discourse surrounding the role of Artificial Intelligence (AI) in various aspects of business operations. Decision-making, in particular, is a crucial component of every business-related activity. As businesses expanded and generated massive amounts of data, it became clear that humans alone could no longer make consistently accurate decisions. Moreover, it is demonstrated that humans often rely on heuristics and cognitive biases in their decision-making, leading to suboptimal outcomes. Given today's business environment's complexity, instability, and interconnected nature, businesses possess all the characteristics of complex systems. With the aid of AI, decision-making can be significantly enhanced. Various subfields of AI, such as artificial neural networks, fuzzy logic networks, and agents, have been developed in recent years, playing a pivotal role in enabling AI-driven decision-making. Findings through using purposeful and complex systems suggest that although AI subfields in decision-making can make sound decisions, they exhibit deficiencies in complex systems where human interaction and interconnectedness across different organizational levels are present. Currently, AI technology is not equipped to address these challenges. As a result, the decision-making process should not be entirely delegated to machines and AI. This discussion gives rise to the duality of augmentation and automation. Decision-making can be categorized into three levels: operational, tactical, and strategic, ranging from structured to unstructured decisions. The analysis reveals that AI performs admirably as an assistant or replacement tool at the operational level. However, as moving towards tactical and strategic decisions, although its augmentation abilities remain somewhat consistent, its capabilities for replacement and automation diminish significantly. Consequently, AI is believed to lack the ability to automate strategic and unstructured business decisions completely