Agentic AI & Large Language Models for Enterprise Decision Intelligence and Business Process Automation

Abstract

The use of agentic artificial intelligence and large language models is increasingly becoming a vital tool in decision-making within organizations and business process automation. The present study focuses on understanding how agentic artificial intelligence and language models would assist organizations in engaging in reasoning, knowledge retrieval, workflow management, and decision-making. Interpretivist philosophy and an inductive research design were utilized in this secondary research study using a qualitative research methodology. Thematic analysis was adopted to identify the technological capabilities, enterprise usage, advantages, implementation challenges, and governance requirements. Market reports indicate that enterprises have started adopting agentic AI agents rapidly in fields like customer service, operations, marketing, logistics, and strategy. The Salesforce market data indicates that average activated agents increased from five to thirteen between February 2025 and April 2026. The performance of Agentic Work Unit increased by 15 % compound growth every month during the time frame. In terms of access to AI solutions that have been authorized, according to Deloitte, 60 % of people have access to these solutions. Deep business transformation is achieved in only 34 % of the organizations. The research shows that Agentic AI performs at the highest level in terms of intelligent workflow execution, contextual reasoning, enterprise knowledge retrieval, and repetitive task automation. Nevertheless, there are significant weaknesses in the case of hallucinations, data integrity, cybersecurity, explainability, authorization, and governance. Mature governance is only present in one out of five firms in relation to autonomous AI agents.