AI ALGORITHMS IN BUSINESS PROCESS AUTOMATION

Purpose:The aim of the study is to describe cognitive aspects of intelligent AI algorithms, defined in the context of their evolution, and outline AI functionalities necessary for a given organization to support business process automation.The aim of the research is also to describe process according to the systems analysis method in Graham's approach and the cybernetic systems analysis approach, both of which are elements of organisational preparation for implementing tools enabling business process automation.Design/methodology/approach: The methodology for defining the factors and the relationship between the factors determining process automation depending on the choice of AI algorithm.Findings: This paper presents ontological assumptions of business process algorithmization, describes algorithm types and defines new models for managing intelligent algorithms in the development of AI for business process automation.Research limitations/implications: The limitations of using AI algorithms in process automation include issues with data quality, process complexity, high implementation costs, and difficulties in adapting to changing conditions.Additionally, AI can be problematic due to the lack of decision transparency, dependence on technology, system security, and social resistance.Practical implications: Depending on the main reason for automating a given process (i.e.repetitive data, actions, or exceptions occurring within the process), one can apply deterministic, organized, or self-programming algorithms.Originality/value: A matrix of AI functionalities was developed in the context of the rationale for automation vs. the type of algorithm that implements the AI functionalities in automating data flows, activities and the occurrence of exceptions in the business process.