AI-Driven Antifragile Knowledge Management System: Transforming ERP Simulated Disruptions into Learning Opportunities

Abstract

Purpose: Enterprise Resource Planning (ERP) systems, such as SAP (Systems, Applications, and Products in Data Processing), are critical to modern enterprises, enabling the integration of core business functions and the management of essential data. Ensuring their availability, reliability, and adaptability is paramount, as disruptions can result in significant operational and financial consequences. Traditional Knowledge Management (KM) approaches emphasize the preservation of ERP-related knowledge but often lack responsiveness to emergent risks. This study introduces a novel framework grounded in the concept of antifragility—where systems grow stronger under stress—by simulating disruptions to enable continuous knowledge evolution and system adaptation. Methodology: A mixed-methods research design combines simulation-based inquiry with Design Science Research (DSR) to investigate antifragile KM within ERP environments. Artificial Intelligence (AI) tools are integrated into the KM system to analyse ERP failures, generate runbooks, and proactively manage recovery knowledge. Controlled simulations of kernel upgrades and failure scenarios—modelled on ITIL 4 incident typologies—serve as structured stressors to expose vulnerabilities. Lightweight LLMs, Retrieval-Augmented Generation (RAG) pipelines, and semantic search tools are employed to codify procedural knowledge and enhance the responsiveness of ERP operations. Findings: The results demonstrate that embedding antifragile principles into ERP KM improves organizational learning, responsiveness, and recovery capabilities. Transitioning from static knowledge repositories to dynamic, AI-enabled systems allows for autonomous decision-making, decentralized knowledge flow, and adaptive documentation. Each disruption becomes a learning event, reinforcing the resilience and self-improvement of the ERP knowledge ecosystem. Implications: Empirical insights suggest that AI-driven antifragile KM transforms ERP disruptions into opportunities for growth, rather than threats to stability. The proposed framework supports the development of systems that not only recover from failure but also become progressively more robust and adaptive through structured experimentation and continuous learning.