Agentic Artificial Intelligence in Auditing: A Systematic Literature Review

Background: The advent of agentic artificial intelligence (AI) is a paradigm shift in auditing practices, from the traditional rule-based automation towards autonomous, goal-oriented machines that have the capacity to make decisions without human interaction. Agentic AI systems use large language models (LLMs) as well as multiple agents to perform complex audit procedures with very little human interaction. Objective: Within a literature review in a systematic way, an analysis of the current situation for the application of agentic AI in auditing will be accomplished, the relevant frameworks and architectures, advantages and challenges will be assessed and some recommendations for research and practice will be introduced. Methods: Using the criteria recommended by PRISMA 2020, we performed a systematic search of academic databases (Web of Science, Scopus, IEEE Xplore and ACM Digital Library) and grey literature from the period January 2020 to February 2026. Studies were evaluated using formative inclusion and exclusion rules and finally 47 studies were included in a synthesis. Results: The applications uncovered during the review were among five major areas: (1) financial statement audit (2) internal audit and control testing (3) compliance and regulatory audit (4) fraud detection and risk assessment (5) audit planning and resource assignment. The most important agentic AI frameworks include CrewAI, LangGraph, AutoGen, and audit-focused systems. Major benefits are increased efficiency (40-60% reduction of time), increased accuracy (15-25% reduction of error), continuous monitoring capability and scalability. Critical challenges include limitations of transparency and explainability, regulatory uncertainty, data security, data ownership and accountability, model drift, and data integration. Conclusions: Agentic AI shows transformative potential for auditing in terms of facilitating change from periodic to continuous auditing and reactive to proactive risk management. Successful adoption requires governance frameworks, transparent audit trails, specialized auditor know-how, and regulatory standards.