Enhancing small and medium enterprise performance through artificial intelligence integration in accounting

Jupić, Nedžad; Gadžo, Amra Conference Paper — Accepted Manuscript (Postprint) Enhancing small and medium enterprise performance through artificial intelligence integration in accounting Suggested Citation: Jupić, Nedžad; Gadžo, Amra (2025) : Enhancing small and medium enterprise performance through artificial intelligence integration in accounting, Scientific conference with international participation SMEPP 2025: "Artificial intelligence as a Driving Force for the Development of Underdeveloped Areas: Opportunities and Perspectives", Novi Pazar, Serbia, 3. jun. 2025, In: Univerzitet u Novom Pazaru (Ed.): SMEPP 2005: zbornik radova, ISBN 978-86-83074-11-2, Univerzitet u Novom Pazaru, Departman za ekonomske i računarske nauke, Novi Pazar, Srbija, pp. 33-46, https://publikacije.uninp.edu.rs/index.php/smepp2025/article/view/327 This Version is available at: https://hdl.handle.net/10419/324135 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ ENHANCING SMALL AND MEDIUM ENTERPRISE PERFORMANCE THROUGH ARTIFICIAL INTELLIGENCE INTEGRATION IN ACCOUNTING Nedžad Jupić Agency for Accounting and Business Consulting Cortex LLC Tuzla, Bosnia and Herzegovina nedzad[email protected] Amra Gadžo Faculty of Economics, University of Tuzla Tuzla, Bosnia and Herzegovina amra.ga[email protected] Abstract This paper examines the degree of implementation of artificial intelligence (AI) tools within accounting information systems in small and medium-sized enterprises (SMEs) in Bosnia and Herzegovina, with the objective of identifying the perceived benefits, key barriers, and their effects on business performance. The research was conducted on a sample of 99 enterprises, utilising a structured questionnaire. Findings reveal that only 21.21% of enterprises currently employ AI tools in their accounting processes, while a mere 20% of respondents make use of advanced document management systems (DMS) or cloud-based solutions—technologies that facilitate integration with AI. The research highlights a generally positive attitude among respondents regarding the impact of AI application on SME operations. The highest average rating was attributed to the statement that AI enhances the efficiency of accounting processes (4.22), followed by improved quality of financial reporting (4.02) and more effective managerial decision-making (3.93), indicating a recognised added value of AI tools in the areas of analytics and decision support. The main obstacles to AI adoption, as identified by respondents, include a lack of knowledge and expertise (3.80) and insufficient regulatory framework (3.73). In terms of financial readiness to invest in AI technologies, 44.40% of enterprises indicated willingness to invest up to approximately EUR 500 annually, 37.40% between EUR 500 and 1,000, and only 18.20% more than EUR 1,000. Overall, the findings suggest a significant, yet underutilised potential of AI tools in SME accounting, characterised by favourable user perceptions but constrained by educational, financial, and legislative limitations. Keywords: AI tools, accounting digitalisation, business enhancement, small and medium-sized enterprises. INTRODUCTION The application of artificial intelligence (AI) in business processes has significantly reshaped the way enterprises operate. In the field of accounting, AI tools enable the automation of routine tasks, enhance the accuracy of financial data, and facilitate more efficient and timely business decision-making. Furthermore, the digital transformation of tax authorities and efforts to establish real-time transaction monitoring systems—primarily aimed at reducing tax evasion—underscore the need for the adoption of modern digital solutions. Digitalisation is, in fact, a prerequisite for the automation of accounting processes and the integration of AI tools into accounting information systems. However, small and medium-sized enterprises (SMEs) often lag behind, firstly in terms of digital transformation, and subsequently in the adoption of artificial intelligence. Bosnia and Herzegovina, as a transitional economy, faces challenges in the digital transformation of its SME sector. SMEs in Bosnia and Herzegovina are characterised by limited resources, a shortage of qualified personnel, and often inadequate support in terms of infrastructure and legal frameworks. Accordingly, the aim of this paper is to analyse the extent to which AI tools are applied in the accounting practices of SMEs in Bosnia and Herzegovina, and to identify the perceived benefits and obstacles to their implementation. The paper is structured into three main sections. The first section provides an overview of previous research on the advantages and challenges of introducing AI tools in accounting, along with a review of statistical data concerning the level of digital transformation and the use of AI tools by SMEs in the Western Balkans. The second section outlines the research methodology, objectives, and data sources. The third section presents the research findings and offers a discussion of the results. LITERATURE REVIEW Artificial intelligence (AI) significantly enhances business operations by enabling more efficient data management and the improvement of business processes. However, the implementation of AI also brings challenges, including ethical concerns related to data privacy and trust (Stanisic et al., 2024). Moreover, small and mediumsized enterprises (SMEs) often lack the financial and human resources necessary to adequately assess and apply the potential of AI applications for their specific needs (Szedlak et al., 2021). According to the WB DESI 2022 report (p. 10), the adoption of digital technologies by SMEs in the Western Balkans remains substantially below the EU average (35% in the Western Balkans compared to the EU average of 55%). On average, only 7% of businesses in the Western Balkans used big data, 16% utilised cloud technologies, and merely 3% adopted AI in 2021. In comparison, these figures in the EU were 14%, 34%, and 8%, respectively. The same report also states that in 2021, 32% of SMEs in Bosnia and Herzegovina had at least a basic level of digital intensity, closely aligning with the Western Balkans average of 35%. Electronic information exchange was used by 26% of enterprises, which is slightly above the regional average of 24%. Social media platforms were not widely utilised by businesses in Bosnia and Herzegovina, significantly falling below the regional average. Advanced technologies were modestly employed: 7% of businesses used cloud solutions, 4% big data, and only 2% AI. Electronic invoicing was used by 19% of enterprises, a higher rate than the regional average. Regarding the use of information and communication technologies (ICT) for environmental sustainability, 54% of enterprises in Bosnia and Herzegovina employed ICT for environmental protection purposes, compared to the Western Balkans average of 56% (WB DESI, 2022, p. 26). Research by Rawashdeha, Bakhitb, and Abaalkhailb (2022) identified four key technological factors—compatibility, challenge readiness, efficiency improvement, and time saving—as significant in influencing the adoption of AI by SMEs through the automation of accounting processes. Accounting automation not only facilitates AI adoption but also independently increases interest in the technology. A firm's readiness to embrace new technologies and the alignment of AI with existing business practices are essential prerequisites for successful implementation. Furthermore, time savings and improved efficiency have a direct impact on the decision to adopt AI, suggesting that automation is not the sole determinant. AI simplifies processes, accelerates decision-making, and improves the accuracy of financial reporting (Panwar, 2023). According to research by Okeke, Bakare, and Achumie (2024), AIbased accounting tools are revolutionising routine tasks such as bookkeeping, invoicing, and cost tracking. By automating these functions, SMEs can reduce human errors, save time, and allocate resources more efficiently. This automation enhances efficiency, accuracy, and competitiveness while reducing staff workload (Khashimov & Khaydarova, 2022). AI systems are also capable of analysing large datasets to detect anomalies and patterns in financial data, thereby improving decision-making and financial sustainability (García-Vera, Juca-Maldonado & Torres-Gallegos, 2023). Iyelolu, Agu, Idemudia, and Ijomah (2024) highlight that AI technologies can substantially enhance operational efficiency, product development, customer engagement, and competitive advantage for SMEs. RESEARCH OBJECTIVES, SAMPLE, AND DATA SOURCES The subject of this research is the examination of the extent to which artificial intelligence (AI) tools are applied in the accounting practices of small and mediumsized enterprises (SMEs), users’ perceptions of their usefulness, obstacles encountered during implementation, and the analysis of the impact that AI integration has not only on the efficiency and quality of accounting processes but also on the overall business performance. The study sets out the following operational research objectives: • To determine the level of application of AI tools in accounting • To assess users’ perceptions of the contribution of AI tools to SME accounting • To identify barriers to the implementation of AI tools in accounting • To analyse the impact of AI tools on the overall business performance of SMEs The research population consisted of managers and accountants working in SMEs in Bosnia and Herzegovina. Data were collected through an online survey. The questionnaire included a combination of Likert scale questions, closed-ended questions with multiple-choice answers, and open-ended questions designed to provide qualitative insights. The questionnaire was distributed to 550 email addresses, and 99 responses were collected, resulting in a response rate of 18%. The data were analysed using descriptive statistics, correlation analysis, and regression analysis. To provide a better understanding of the research results, the key characteristics of the sample are first presented. Among the surveyed enterprises, 55.55% had been operating for more than 10 years; 17.17% had been in operation between 5 and 10 years; and 27.28% had been operating for up to 5 years. In terms of sectoral distribution, 66.67% of enterprises operate in the service sector, 15.15% in manufacturing, and 18.18% in trade. The structure of respondents was as follows: 36.36% accountants, 28.28% managers, and 35.36% employees working in accounting and finance departments. RESEARCH FINDINGS AND DISCUSSION In order to comprehensively explore the role of artificial intelligence in SME accounting practices, this chapter outlines the key operational objectives that guide the research. These objectives aim to provide structured insight into the extent, perception, challenges, and effects of AI tool implementation in the accounting function. THE EXTENT OF APPLICATION OF AI TOOLS IN ACCOUNTING When it comes to the use of artificial intelligence (AI) tools—such as those for text, image, and code generation (e.g. ChatGPT, Microsoft Copilot, Google Gemini), developer and data analysis tools (e.g. GitHub Copilot, DataRobot), accounting and finance tools (e.g. UiPath, Xero), and predictive analytics platforms (e.g. Power BI, Salesforce)—the findings indicate that 55.56% of respondents make use of at least one such tool, whereas 44.44% do not. However, the situation appears even less favourable with respect to the application of AI tools specifically in accounting. In response to the question of whether the enterprise uses AI in accounting processes— such as optical document recognition, automatic posting, predictive analytics in accounting, or chatbot support—only 21.21% of participants confirmed usage, while 78.79% reported no application of these tools. Nevertheless, 16.16% of enterprises indicated plans to introduce AI in their accounting functions, and 13.13% stated that they currently lack sufficient knowledge on the topic. The remaining respondents were undecided. In order to assess the prerequisites for the introduction of AI in accounting, a series of questions was posed to explore the current degree of digitalisation and automation in accounting processes. One of the key inquiries related to the format in which documentation is most frequently received. Results show that in 73.74% of cases, documents are received in paper format, as PDFs or images, while only 26.26% are received in structured digital formats such as Excel, XML, or CSV. These results provide insight into the potential for the adoption of AI in accounting. AI tools in accounting tend to operate most efficiently when working with structured digital data (e.g. Excel, XML, CSV). In contrast, paper documents, PDFs, and image files require additional conversion processes to become machine-readable. Such processes, including Optical Character Recognition (OCR), can be slow, less accurate, and place additional burdens on the system. It may therefore be concluded that enterprises need to undergo both technical and organisational transformation in order to effectively integrate AI tools into the accounting function. Beyond the format in which documents are received, another important dimension of digital readiness concerns the manner in which electronic documents are archived. Efficient archiving and data accessibility are essential foundations for automation and the use of advanced technologies such as AI software. To this end, the study also examined how enterprises manage the archiving of documents received via email. Responses to this question can serve as indicators of the presence—or absence—of structured digital document management systems (such as DMS or cloud-based solutions). The distribution of responses is presented in Image 1. Image 1. Method of Archiving Documentation Received via Email Source: Author’s own analysis The results indicate that a significant proportion of enterprises rely on basic document archiving methods, such as physical storage (37% plus 4% of those awaiting document receipt by post) and storage directly within email accounts or on local drives (39%). These practices hinder systematic data processing and retrieval. Conversely, only 20% of respondents utilise advanced document management systems (DMS) or cloud-based solutions, which facilitate better integration with AI tools. This situation highlights the need to improve digital infrastructure to enable full automation of accounting processes. To assess the automation of business transaction recording, we investigated whether the accounting software used by firms automatically suggests accounts for posting. The results are presented in Image 2. Image 2. The capability of accounting software to automatically suggest accounts for posting Source: Author’s own analysis Automatic suggestion of accounts for posting forms the foundation for faster and more accurate processing of accounting documentation. This reduces the possibility of human error and processing time, while enabling employees to focus on analytical and control tasks. With the same purpose of assessing the automation of accounting records, we also investigated the method of posting bank statements. The results are presented in Image 3. Image 3. Methods of posting bank statements Source: Author’s own analysis The majority of respondents (57%) manually transcribe data from bank statements, while only 12% utilise the option to import files that are automatically posted. An 38% 34% 28% When you enter invoice data, does your software automatically suggest accounts for posting? No Yes Unknown 57% 12% 6% 25% How do you currently post bank statements? We manually transcribe the data We import a file that is posted automatically The software automatically recognises and posts bank statements Unknown even smaller proportion, 6%, report that their software recognises and posts bank statements automatically, whereas 25% of respondents are either unsure or unfamiliar with the posting method. These findings indicate a high degree of manual data entry in a critical segment of the accounting process, suggesting that businesses still lack the adequate technical prerequisites for the implementation of AI tools in accounting procedures. Regarding the digital maturity of accounting—i.e., the extent to which digital technologies and AI are integrated into accounting processes—23.2% of respondents consider themselves to be in the initial phase (processes are predominantly manual and lack standardisation), 39.4% regard themselves as being in the second phase—digitally aware (recognition of the need for digitalisation and basic initiatives have been launched), 27.3% view themselves as being in the third phase—digitally active (digital tools introduced, partial automation), 8.1% identify with the fourth phase—digitally advanced (integrated systems and analytics, high efficiency), and only 2% consider themselves in the fifth phase—innovative (use of AI, predictive analytics, with strategic impact). It can be concluded that the current state reflects a low level of digitalisation and automation within accounting processes, which constitutes a significant barrier to the effective utilisation of AI technologies. However, the expressed intention of some businesses to adopt AI tools, alongside existing basic forms of digitalisation, indicates potential for future development. To increase the adoption of AI tools, a comprehensive technical, organisational, and educational transformation within enterprises is essential. CONTRIBUTION OF AI TOOLS IN ACCOUNTING The evaluation of respondents’ views on the contribution that AI technology can bring to accounting is presented in Table 1. The research results indicate that respondents recognise the positive potential of AI tools in improving accounting processes. The highest average ratings were awarded to the areas of automation of postings (3.77) and cost analysis (3.73), suggesting that these are perceived as the most suitable fields for the application of artificial intelligence. Table 1. Respondents’ Views on the Contribution of AI Tools in Accounting by Specific Segments DESCRIPTION AVERAGE STDEV Automation of postings 3,77 1,25 Preparation of accounting reports 3,65 1,21 Cost analysis 3,73 1,18 Management of receivables and payables 3,63 1,19 Cash flow forecasting 3,45 1,21 Error detection 3,62 1,17 Source: Author’s own analysis Respondents’ views highlight the recognised value of AI technology in accounting, particularly in automation and analytical tasks. Conversely, there remains scope for further awareness-raising and education regarding more advanced functionalities such as prediction and anomaly detection. The assessment of respondents’ opinions on what would facilitate their work over the next 12 months (in the short-term future) is presented in Image 4. Image 4. Respondents’ perceptions of the needs for facilitating accounting tasks by segment Source: Author’s own analysis The research results indicate that respondents perceive the greatest potential for improving their work through the reduction of manual data entry in accounting. No fewer than 67 respondents identified reduced manual input as the key factor that would most facilitate their daily tasks over the next twelve months. This is further complemented by the desire for faster processing of accounting documentation, specifically through the automation of posting entries. Respondents’ answers clearly highlight the need to enhance efficiency by automating fundamental accounting processes, particularly those involving manual data entry and processing. When asked whether they consider it important to have an AI tool in accounting that independently recognises and processes documents, 47.5% replied affirmatively, 48.5% considered it partially important, and 4% stated it was not important. Conversely, when questioned whether the transition to automatic posting would reduce their control over data, 26.3% responded affirmatively, 31.3% said no, and 42.4% were uncertain. These findings suggest a cautious attitude among respondents, underscoring the need for further education and demonstration of the security and transparency of AI tools in order to overcome concerns related to potential loss of data control. In conclusion, the majority of respondents recognise the significant potential of artificial intelligence application in accounting, particularly in the areas of posting automation and cost analysis. The results also reveal a strong demand for reducing manual data entry and accelerating document processing, thereby confirming a readiness to embrace automation of basic accounting tasks. However, there remains a degree of caution among respondents regarding full automation, largely due to concerns over loss of control of data. This highlights the necessity for comprehensive education on the capabilities, security, and benefits of AI tools in accounting processes, alongside awareness of associated risks such as security vulnerabilities and technological dependency. 67 55 41 28 010 20 30 40 50 60 70 80 Less manual data entry Faster processing of accounting documentation Automation of report preparation Automation of client communication (Chatbot) What would make your work easier over the next 12 months?