Abstract This study investigates the transformative impact of Generative Artificial Intelligence (GenAI) on traditional cost accounting frameworks. While classical costing methods such as Activity-Based Costing (ABC) and Standard Costing have historically relied on static data and retrospective analysis, GenAI offers a dynamic, predictive approach to cost management. This paper utilizes a mixed-methods approach to analyze how Large Language Models (LLMs) can reduce variance in cost estimation, automate cost allocation, and enhance decision-making agility in volatile economic environments. The findings suggest that GenAI integration significantly improves the accuracy of indirect cost allocation and predictive budgeting, providing a competitive edge for firms operating in high-inflation and complex supply chain scenarios. Furthermore, this study explores the emerging role of the "AI-augmented accountant" and the necessity of new governance frameworks to manage algorithmic biases and ensure transparency in automated financial reporting. Keywords:Generative Artificial Intelligence, Cost Accounting, Predictive Analytics, Management Accounting, Strategic Management Accounting, Digital Transformation, Algorithmic Governance, Explainable AI 1. Introduction The digital transformation of the accounting function has evolved from basic Robotic Process Automation (RPA)which largely focused on rule-based transaction processing to intelligent decision support systems powered by Generative AI. Traditional cost accounting systems, often deeply embedded in legacy Enterprise Resource Planning (ERP) architectures, frequently struggle with the "data latency" problem. In this conventional paradigm, costs are analyzed long after the reporting period ends, rendering the information obsolete for real-time strategic pivots. As global markets become increasingly characterized by Volatility, Uncertainty, Complexity, and Ambiguity (VUCA), the ability to transition from descriptive analytics (what happened) to prescriptive analytics (what should we do to manage costs) is becoming a primary driver of sustainable competitive advantage. Generative AI introduces the capability to synthesize vast arrays of unstructured data market trends, global supply chain disruptions, social sentiment, and historical variance reports to forecast future cost behaviors with unprecedented precision. This article posits that GenAI is not merely a tool for incremental efficiency but a fundamental restructuring agent for the management accounting profession, necessitating a move toward "Continuous Accounting." 2. Theoretical Framework and Literature Review Current literature on management accounting emphasizes the need for systems that provide real-time strategic alignment. Traditional costing models, particularly Activity-Based Costing (ABC), were designed for stable manufacturing environments that have largely been superseded by modern digital, service-oriented, and high-frequency workflows. 2.1 The Limitations of Current Costing Models Standard costing and ABC rely heavily on the assumption of stable activity drivers. In contemporary global supply chains, drivers are inherently unstable due to fluctuating commodity prices, geopolitical risks, and erratic consumer demand. Relying on static drivers often leads to: Under-costing of complexity: Hidden costs in logistics, compliance, and cybersecurity are frequently buried in general overhead, skewing the actual profitability of products. Delayed feedback loops: Standard variances are typically reviewed monthly or quarterly, providing no opportunity for immediate corrective action during market fluctuations. 2.2 Theoretical Lens: Dynamic Capabilities Theory To understand the necessity of this shift, we apply the Dynamic Capabilities Theory, which posits that a firm’s competitive advantage resides in its ability to integrate, build, and reconfigure internal and external competencies to address rapidly changing environments (Teece et al., 1997). GenAI acts as a catalyst for these dynamic capabilities within the accounting function, allowing firms to pivot cost structures in real-time, effectively treating cost accounting as a strategic lever rather than a compliance burden. 2.3 The Evolution to AI-Augmented Frameworks The transition is not merely about replacing manual spreadsheets with automated code. It is about a fundamental architectural shift. Where traditional systems required accountants to "map" data manually, GenAI models can infer relationships between data points that are not explicitly linked in the ERP. This capability, referred to as "Semantic Mapping"allows for a more organic, iterative process of cost discovery. By leveraging natural language processing, these systems can "read" procurement contracts, shipping invoices, and internal communications to identify cost drivers that traditional ERP logic would ignore. 2.4 Comparative Analysis: Traditional vs. GenAI-Augmented Costing Table 1 summarizes the paradigm shift from traditional methodologies to AI-augmented models, highlighting the transition from retrospective monitoring to proactive prediction. Table 1: Paradigm shift from traditional methodologies to AI-augmented models Feature Traditional Cost Accounting GenAI-Augmented Costing Data Basis Historical, Structured (ERP) Historical + Unstructured (Market/Sentiment/External) Allocation Logic Static Activity Drivers Dynamic, Machine-Learned Adaptive Drivers Forecasting Linear Extrapolation/Budgeting Non-linear Predictive Modeling Response Time Periodic/Lagging (Monthly/Quarterly) Near Real-time/Proactive (Continuous) Accounting Role Data Recorder/Compiler Strategic Advisor/AI Orchestrator 3. Methodology This research adopts a qualitative case study analysis of three multinational manufacturing firms that have piloted GenAI integration for overhead cost allocation. Data was collected through semi-structured interviews with CFOs, senior cost accountants, and IT system architects, supplemented by quantitative variance analysis reports from the pre- and post-AI implementation phases. The criteria for selection were: High Transaction Volume: Companies with complex product lines requiring granular overhead tracking. Operational Volatility: Firms exposed to significant supply chain fluctuations in the last 24 months. AI Maturity: Organizations that had already migrated basic data processes to cloud environments. The interviews utilized a thematic coding approach, focusing on the implementation process, organizational resistance, and technical hurdles associated with integrating LLMs into existing SAP/Oracle accounting environments. The analysis focuses specifically on how these firms moved from annual static budgeting to dynamic rolling forecasts. 4. Findings and Discussion 4.1 Reduction in Cost Variance Our empirical analysis indicates that firms leveraging LLMs for cost forecasting experienced a significant reduction in budget-to-actual variance. By analyzing external macroeconomic variables such as inflation indices and shipping lead times alongside internal production data, LLMs were able to adjust cost projections in real-time. Table 2 shows the comparative variance in direct material costs over a six-month period following AI implementation in the observed pilot programs. Table 2: Comparative variance in direct material costs over a six-month period Month Traditional Method Variance (%) AI-Augmented Method Variance (%) Month 1 8.4% 4.2% Month 2 7.9% 3.5% Month 3 9.1% 2.8% Month 4 7.5% 3.1% Month 5 8.2% 2.5% Month 6 8.6% 2.2% 4.2 Strategic Implications and The "AI-Augmented Accountant" The integration of GenAI facilitates three major strategic pillars: Granular Cost Driver Identification: AI identifies correlations between non-obvious activities and indirect cost accumulation, such as the relationship between shipping route weather patterns and warehouse utility costs. Predictive Budgeting: Shifting from "Fixed Annual Budgets" to "Adaptive Rolling Forecasts" allows companies to reallocate resources weekly rather than annually. Enhanced Governance: Automated audit trails generated by AI systems reduce manual error and the potential for manipulation, though this introduces a new need for "explainable AI" (XAI) to ensure audit compliance. The role of the accountant is shifting from "scorekeeper" to "orchestrator." In this new model, the accountant does not perform the allocation calculation themselves; rather, they design the prompts, audit the model's logic, and contextualize the output for strategic decision-making. This requires a new pedagogical approach focusing on data literacy, model governance, and logical reasoning over manual computational proficiency. 4.3 Organizational Barriers and Implementation Challenges A critical finding is that the success of GenAI implementation is rarely about the model's accuracy but rather the organization's AI readiness. Several barriers were identified during the study: The "Black Box" Anxiety: Senior stakeholders expressed reluctance to trust a model that could not clearly articulate the "why" behind a cost allocation. Data Silos: Many firms struggled to integrate unstructured external data (e.g., news feeds, social sentiment) with structured internal financial data. Cultural Inertia: The workforce perceived AI as a threat to job security. Successful firms mitigated this by repositioning AI as a tool for eliminating "drudgery" rather than "people." 5. Governance and Ethical Considerations As AI takes on a greater role in cost allocation, the risk of "black box" accounting emerges. Organizations must establish strict governance frameworks to monitor for algorithmic bias ensuring that cost-cutting algorithms do not unfairly penalize specific business units or lead to unethical labor practices. The "Explainab
Rossi et al. (2026) studied this question.