Author: Nor Adila Binti Zulkifli/Rohana Binti Mohd Noor
Affiliation: Universiti Teknologi MARA,Shah Alam/Puncak Alam
Email: This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Activity Based Costing (ABC) is a technique used to determine the absorption of indirect costs (overhead) in product or service costs. The introduction of ABC took place in the 1990s to cover the limitation from the use of traditional methods. The appearance of artificial intelligence (AI) in work environments has changed the paradigm of the ABC technique. Three themes have been established: The use of ABC method for overhead absorption, the benefits of adopting AI in ABC method and the challenges of using AI in ABC method.

Keywords: Artificial Intelligence and ABC Method

INTRODUCTION

The use of ABC for overhead absorption

In contrast to traditional cost accounting systems, the ABC technique first collects indirect costs for every organisational activity before allocating them to cost items, such as goods or services, corresponding to the activities that generated them. Activity analysis is the most essential element of the ABC approach and involves defining relevant metrics of activity outputs and resources (cost drivers), as well as impact on the cost incurred in manufacturing of the product or service.

The main objective of ABC technique is to reclassify the majority of indirect costs which include overhead cost that will result in a significant improvement in cost calculation accuracy. The implementation of ABC techniques in calculation of costs can be summarised as follows:

1. Identification of activities
2. Identification of the total cost for every activity (cost pool)
3. Identification of the driver for each activity cost (cost driver)
4. Calculation of the Overhead Absorption Rate (OAR) Budgeted Overhead/ Budgeted Activity Level
5. Absorption to product/service costs (OAR x Actual Activity Level)

The use of ABC for indirect costs absorption into product costs will ensure that cost allocated to each product is as accurate as possible. The accuracy of product cost will ensure the accuracy of selling price setting that in return maximises organization’s profit making. However, traditional ABC techniques are starting to demonstrate its limitations as product, services and project complexity rises in the age of digitisation. There are few examples of AI applications in ABC including Machine Learning, Natural Language Processing, IOT Integration and Anomaly Detection.

Benefits of adopting AI in ABC Technique

The complexity in the business world makes precise cost estimation and effective cost management extremely difficult. In a case study of infrastructure projects, they reported a 25% improvement in cost accuracy over traditional ABC methods. Their approach used neural networks to identify complex relationships between project features and costs, allowing for more detailed cost predictions. A deep neural network was trained on historical project data to forecast activity costs, considering various project characteristics and parameters (Judijanto, 2024).

Machine learning techniques, such as linear regression, and artificial neural networks, have been applied in various ways to estimate building costs, including for preliminary cost assessments (Aung et al., 2023). Building on this research, an AI-driven ABC framework has been proposed, which leverages deep learning techniques to analyse historical project data and provide more precise cost estimates. The predictive analytics enabled by AI use historical project data to increase the accuracy of the cost estimation process and lower the chance of a budget overrun (Aung et al., 2023). By examining historical cost data and detecting patterns, AI can predict future costs and uncover potential financial risks before they arise(Chen, 2025).

Artificial Intelligence (AI) is at the leading edge of innovation in several fields, including project management. The shortcomings of traditional ABC may be greatly reduced by AI's capacity. AI helps to analyse large data, identify intricate patterns, and generate precise forecasts (Judijanto, 2024). A convolutional neural network designed to identify hidden patterns and relationships in project cost data. Machine learning and artificial intelligence are valuable for analytics as they can extract meaningful insights from raw data and provide actionable advice and predictions. ML provides tools to extract valuable insights, recommendations, and predictions from large and complex datasets. AI, on the other hand, has a broader purpose: to replicate human intelligence or specific aspects of it, along with other cognitive functions (Kibria, 2018).

Absorption of overhead cost to cost items required data analysis and faster decision making. AI has been recognised for its ability to improve prediction accuracy and accelerate decision-making processes. (Judijanto, 2024). A crucial aspect of AI, machine learning enhances data analysis and supports better decision-making by providing insights on how to manage costs effectively. While human expertise is valuable for narrowing the focus to generate solutions and manage complex issues, it has limited capacity for discovering new answers and insights. The future of wireless networks will undoubtedly depend on AI (Kibria, 2018).

Traditionally, all data collected regarding the analysis of the indirect cost and driver of the costs mainly done by an accountant, AI-powered ABC systems automate the processes of data collection and classification, greatly minimising the time accountants spend on manual data entry and routine reconciliations. For instance, AI can automatically classify transactions and identify cost drivers using machine learning models, removing the need for human involvement in these time-intensive tasks (Chen, 2025). The AI-APBC model showed a 40% reduction in time for conducting complex cost analysis, compared to the traditional ABC technique (Judijanto, 2024).

Challenges of using AI in ABC Method

The use of AI in ABC Method has also raised a number of limitations and challenges which include ethical consideration. Among the considerations that need to be taken into account are obtaining informed consent from organisations providing the data, implementing data anonymization techniques to protect sensitive information, ensuring transparency in the AI model's decision-making process, and addressing potential biases in both the training data and model outcomes (Judijanto, 2024).

Moreover, implementing AI-driven ABC requires significant upfront investment (Chen, 2025).  This includes purchasing the necessary hardware, software, and possibly upgrading existing IT infrastructure to support AI technologies. Companies also need to consider the cost of hiring AI experts or working with external consultants to design, develop, and integrate the system. Employees will need training to understand how the new AI-powered system works and how to use it effectively. There may also be a period of resistance or adjustment to change, which requires time and resources to manage.

AI-driven ABC tools must integrate smoothly with the company’s existing ERP systems, financial software, and strategic data streams. Integrating AI-driven ABC systems can face challenges due to compatibility issues with existing legacy software, potentially leading to operational disruptions. To ensure efficient operation, smooth data flow between different departments and the AI algorithms is essential. The automation of activity tracking and cost categorisation processes can be intricate, requiring the system to accurately identify and classify activities for correct cost assignment (Chen, 2025).

Businesses must consider whether the long-term benefits of AI-driven ABC such as more accurate cost allocation and better decision-making justify the initial investment. The ROI might not be immediately apparent, so companies must be prepared for a potentially long-term investment before seeing measurable returns. AI systems need to be continuously updated and maintained to keep up with technological advances, changing business conditions, and evolving market trends. Companies must budget for ongoing maintenance and adaptation to ensure the system remains effective.

The ability to develop AI-driven ABC systems that are customised to specific industries is also a challenge. For example, different industries like manufacturing, health and retail will have different cost structures tailored to the business nature. Thus, several critical steps including data preprocessing, creating cost pools through clustering and regression algorithms, and allocating resources using decision tree models are required (Chen, 2025).