AI-driven ERP (*)

ERP is traditionally responsible for preparing reports and enabling analysis of business results in real-time. However, today, companies increasingly need innovative technologies to identify trends, recommend actions, assess numerous interrelationships, and simplify dynamic operations [1]. Integrating artificial intelligence with ERP systems promises to fulfill the mentioned needs.

Artificial intelligence (AI), a term coined by scientist John McCarthy in 1955, has undergone a massive transformation in the last decade. Defining artificial intelligence is challenging because of two main questions: (1) what constitutes human intelligence and (2) whether aspects of human intellect can be amenable to computer simulation. OECD [2] defines an “AI system as a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations, or decisions influencing real or virtual environments. AI systems are designed to operate with varying levels of autonomy”.

Artificial intelligence has significantly changed how software works and functions within the enterprise. Today’s ERP software is more complicated than earlier versions, even from just five years ago [3]. Due to the increasing integration of AI solutions within the architecture of these systems, modern ERP systems are often called intelligent ERP systems.

Figure 1 shows a model of an AI-driven ERP, i.e., an ERP system with integrated elements of artificial intelligence. The model is based on Gartner’s composable ERP model [4] with packaged business capabilities (PBC) as basic software components. PBS elements encapsulate and reflect a well-defined business capability recognizable to the business user and packaged for programmatic access [4]. As applications begin to break down into components and define boundaries, it is essential to consider what will be internal and what will be external and use APIs to connect parts (for example, PBCs) to other types of IT components (for example, Micro Services, Apps, Macro Services, etc.). The enhanced model includes an intelligent AI-driven PBC that enables the integration of AI into the ERP architecture and an (intelligent) data AI-driven PBC that prepares and transforms data for intelligent PBC.

So that the integration of AI with the ERP system guarantees additional value for the company, it is necessary to establish a new business model and the fundamental principles of interaction of AI with the work of individuals, organizational functions, and business processes [5].

Figure 1. The architecture of AI-driven ERP

Some of the advantages of AI-driven ERP are:

  • Improving the decision-making process. One of the primary advantages of an ERP system is the ability to enhance and direct workflows and define anything from results to strategy [6]. AI can further enhance these capabilities by processing larger amounts of data than it was previously possible.
  • Easy integration of multiple departments and easier management of all parts of the company. AI-driven ERP systems can combine data from numerous departments into a single database and handle vast amounts of data.
  • Development of modern ways of managing human resources and changing how the organization manages it [7].
  • New business models bring benefits such as cost reduction, improved service quality, increased coordination and efficiency, and delivery efficiency [8].
  • Transformation and initiation of the next phase of digital transition in industries that have grown exponentially over the last decade [9].

Despite numerous advantages, integrating artificial intelligence with ERP systems also brings challenges, such as:

  • The lack of objectivity and caution in artificial intelligence reflects the biases of the people who design it [10].
  • Investment in data storage, computing power, and other digital assets is insufficient. The results of AI algorithms are sometimes less specific than initially expected, so many companies realize that human interpretation, thinking, and action are still needed to achieve tangible and valuable results [11].
  • Scaling of AI services beyond validated and proven solutions to larger groups of clients using AI business models and demonstrated solutions [12]. To make that possible, a much better understanding of AI-based business model innovation principles is necessary, where AI capabilities are integrated into business operations such as value creation, delivery, and data collection to ensure scalable growth.
  • Although Internet platforms make it easier for companies to take advantage of intelligent technologies, they are not ready-made industrial solutions. Working with a supplier who knows the nuances of business and the semantics of ERP information is crucial. The domain experience of the ERP supplier (e.g., for particular industrial needs or functional areas) is essential for developing intelligent systems for efficient and rapid process optimization [13].
  • Due to the resources required to work with such technologies, implementation is complicated. Companies need a lot of time to prepare for their introduction and training on how and for what purpose to use these tools [6].
  • Data security. Concern for data privacy is paramount when developing a machine-learning model that uses sensitive data [14]. A considerable privacy barrier exists in the AI privacy field due to differences in understanding of privacy between businesses and individuals.

As an integrated part of the ERP system, AI can change the entire nature of daily business to raise the company’s overall efficiency, competitiveness, and sustainability. Therefore, reducing the operational costs of integrating AI and ERP systems is the main prerequisite for developing a more massive application of such solutions in companies.

(*) The text has been adapted from the paper:

Gašpar, D., Ćorić, I., Mabić, M. (2023). Composable ERP – New Generation of Intelligent ERP. In: Ademović, N., Kevrić, J., Akšamija, Z. (eds) Advanced Technologies, Systems, and Applications VIII. IAT 2023. Lecture Notes in Networks and Systems, vol 644. Springer, Cham. https://doi.org/10.1007/978-3-031-43056-5_26

References:

[1] Appelbaum, D, Kogan, A., Vasarđhelyi, M., Yan, Z.: Impact of business analytics and enterprise systems on managerial accounting, International Journal of Accounting Information Systems, 25, (2017).

[2] OECD: Recommendation of the council on artificial intelligence, OECD/LEGAL/0449, (2019).

[3] Yathiraju, N.: Investigating the use of an Artificial Intelligence Model in an ERP Cloud-Based System, International Journal of Electrical, Electronics and Computers, 7(2), (2022).

[4] Free, D.: The Composable Enterprise, https://bian.org/wp-content/uploads/2021/12/BIAN-The-Composable-Enterprise-2021-12-08.pdf, last accessed 2022/12/09, (2021).

[5] Iansiti, M., Lakhani, K. R.: Competing in the age of AI: How machine intelligence changes the rules of business, Harvard Business Review, 98,(2020).

[6] Viorel-Costin, B.: The Current Opportunities Offered by AI And RPA Near to the ERP Solution – Proposed Economic Models and Processes, Inside Production Area. A Case Study, Annals of the „Constantin Brâncuşi” University of Târgu Jiu, Economy Series, (1), (2022).

[7] Dhamija, P., Bag, S.: Role of artificial intelligence in operations environment: A review and bibliometric analysis, The TQM Journal, 32(4), (2020).

[8] Davenport, T. H., Ronanki, R.: Artificial intelligence for the real world, Harvard Business Review, 96(1), (2018).

[9] Salam, M.A.: Analyzing manufacturing strategies and Industry 4.0 supplier performance relationships from a resource-based perspective, Benchmarking: An International Journal, 28(5), (2021).

[10] Hammond, K.: 5 Unexpected sources of bias in artificial intelligence, https://techcrunch.com/2016/12/10/5-unexpected-sources-of-bias-in-artificial-intelligence/, last accessed 2022/12/20, (2016).

[11] Tarafdar, M., Beath, C. M., Ross, J. W.: Using AI to enhance business operations, MIT Sloan Management Review, 60(4), (2019).

[12] Boström, T., Parida, V., Lahti, T., Wincent, J.: AI-enabled business-model innovation and transformation in industrial ecosystems: A framework, model and outline for further research, Journal of Business Research, 127, (2021).

[13] Mahmood, F., Khan, A-Z., Bokhari, R.H.: ERP issues and challenges: A research synthesis, Kybernetes, 49(3), (2019).

[14] Gaur, M.: Privacy preserving machine learning challenges and solution approach for training data in ERP systems, International Journal of Computer Engineering and Technology, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3679275, last accessed 2022/12/20, (2020).