Smartification in Accounting, Economics, and Financial Management

Developing a Bankruptcy Prediction Model for Banks and Financial Institutions Using Recurrent Neural Networks and Financial Soundness Indicators

Alireza Najafi Manesh ^* , Sara Sadat Kamali Nia

This study aimed to develop a conceptual-analytical model for predicting bankruptcy in banks and financial institutions using recurrent neural networks and financial soundness indicators from the perspective of banking, financial, and fintech experts. This study employed a qualitative exploratory design. The participants consisted of 22 banking experts, risk managers, financial analysts, fintech specialists, university faculty members, and supervisory experts in Tehran, selected through purposive and snowball sampling. Data were collected using semi-structured interviews and continued until theoretical saturation was achieved. The data were analyzed through thematic analysis using NVivo software. To enhance the trustworthiness of the findings, member checking, peer coding, constant comparison, analytical memoing, and an audit trail were used. Data analysis resulted in five main categories: fundamental indicators of financial soundness, quality and integration of banking data, the capability of recurrent neural networks in identifying hidden temporal patterns, governance and explainability of prediction models, and implementation requirements for an early-warning system in the banking sector. Participants emphasized that capital adequacy, asset quality, liquidity, profitability, maturity mismatch, non-performing loans, loan concentration, and cash-flow volatility should be analyzed as time-series indicators. LSTM and GRU models were considered suitable because of their ability to learn temporal dependencies and detect gradual financial deterioration before formal failure. The findings indicate that bank bankruptcy prediction cannot rely solely on static financial ratios; rather, it requires the integration of financial soundness indicators, sequential data, deep learning algorithms, human supervisory judgment, and model governance mechanisms. The proposed model may provide a foundation for developing early-warning systems, improving banking risk management, and strengthening financial stability.

The Role of Explainable Artificial Intelligence in Enhancing Financial Managers’ Trust in Algorithmic Outputs and Investment Decisions

Mehdi Taheri Nia , Ebrahim Royani , Seyed Reza Moradian ^*

The Role of Explainable Artificial Intelligence in Enhancing Financial Managers’ Trust in Algorithmic Outputs and Investment Decisions

Designing an Ethical Framework for the Use of Artificial Intelligence in Accounting and Financial Management with Emphasis on Transparency, Responsibility, and Algorithmic Bias

Akbar Sobhani ^*

This study aimed to design an ethical framework for the responsible use of artificial intelligence in accounting and financial management, focusing on algorithmic transparency, professional accountability, and algorithmic bias control. This qualitative study was conducted using thematic analysis. The research population consisted of financial managers, senior accountants, internal auditors, financial technology specialists, and artificial intelligence consultants working in financial organizations and companies located in Tehran. Participants were selected through purposive sampling based on their direct or managerial experience with artificial intelligence systems, financial data analytics, digital accounting, or algorithm-based financial decision-making. Data were collected through semi-structured interviews, and sampling continued until theoretical saturation was achieved. A total of 21 interviews were conducted; theoretical saturation was reached at the eighteenth interview, followed by three additional interviews to ensure data adequacy. After transcription, the data were analyzed using NVivo software through open, axial, and selective coding. The findings revealed that the ethical framework for using artificial intelligence in accounting and financial management consists of five main categories: algorithmic transparency and explainability, professional responsibility and accountability, data fairness and algorithmic bias control, security and confidentiality of financial data, and institutionalization of ethical culture and human oversight. Participants emphasized that the value of artificial intelligence in finance is not limited to prediction accuracy or processing speed; rather, its professional legitimacy depends on explainability, decision traceability, identification of final responsibility, evaluation of data quality, and prevention of the reproduction of historical bias in algorithmic models. The findings suggest that the ethical use of artificial intelligence in accounting and financial management requires a transition from a purely technological perspective to a governance-oriented, human-centered, and accountable approach. The proposed framework can help organizations institutionalize transparency, responsibility, fairness, data security, and human oversight throughout the life cycle of financial AI systems.

Analyzing the Effect of Financial Process Intelligence on Reducing Operational Risk and Improving the Efficiency of Internal Control Systems

Simin Karimi Rad , Mohammad Mehdi Ebrahimi Nia ^*

This study aimed to explain how the intelligence of financial processes contributes to reducing operational risk and improving the efficiency of internal control systems in organizations located in Tehran. This study adopted a qualitative research design using thematic analysis. The participants consisted of 18 financial managers, internal auditors, risk managers, financial IT managers, and accounting information system specialists working in companies and financial institutions in Tehran. Participants were selected through purposive sampling, and recruitment continued until theoretical saturation was achieved. Data were collected exclusively through semi-structured interviews. All interviews were transcribed, organized, coded, and analyzed using NVivo software. The analysis followed an inductive thematic procedure, including familiarization with the data, initial coding, theme development, theme review, theme definition, and final reporting. The findings revealed five major mechanisms through which financial process intelligence reduces operational risk and strengthens internal control efficiency: standardization and automation of financial operations, continuous monitoring and anomaly detection, reinforcement of segregation of duties and traceability, improvement of data quality and system integration, and transformation of internal auditing toward analytical and preventive assurance. Participants emphasized that intelligent financial systems reduce human error, late detection of discrepancies, reliance on manual controls, and unauthorized manipulation of financial data. However, they also noted that weak data governance, insufficient digital competence, and excessive reliance on algorithms may create new operational risks. The study concludes that financial process intelligence can sustainably reduce operational risk and improve internal control efficiency when it is not treated merely as a technical automation project. Instead, it should be implemented as an integrated governance mechanism that combines reliable data, human accountability, continuous monitoring, traceable controls, algorithmic validation, and periodic assurance.

Presenting an Intelligent Model for Detecting Real and Accrual-Based Earnings Management Using Data Mining and Analysis of Managers’ Behavioral Patterns

Seyed Mohammad Mahmoudi ^* , Ali Reza Majd , Hossein Zavarei

This study aimed to present an intelligent model for detecting real and accrual-based earnings management by integrating financial indicators, data mining techniques, and behavioral patterns of financial managers. This qualitative exploratory study was conducted in Tehran. The participants included 22 financial managers, independent auditors, internal auditors, audit committee members, financial analysts, and financial data-mining specialists selected through purposive and snowball sampling. Data were collected exclusively through semi-structured interviews, and the interview process continued until theoretical saturation was achieved. After transcription, the data were analyzed through open, axial, and selective coding. NVivo software was used to organize interview data, classify codes, develop categories, compare themes, and visualize conceptual relationships among the extracted components. Data analysis led to the identification of five main categories: financial and operational indicators of earnings management detection, behavioral and motivational patterns of managers, data-mining infrastructure and intelligent model architecture, corporate governance and internal control mechanisms, and explainability and trustworthiness of algorithmic outputs. The participants emphasized that accrual-based earnings management is mainly reflected in discretionary accruals, accounting estimates, provisions, and revenue-recognition timing, whereas real earnings management appears through abnormal operational decisions such as discretionary expense reduction, unusual sales discounts, overproduction, and manipulation of operating cash flows. The findings also indicated that integrating financial data, managerial behavioral signals, decision-making patterns, and governance indicators can enhance the detection accuracy and practical usefulness of intelligent models. The proposed model suggests that intelligent detection of earnings management is effective when it moves beyond traditional financial statement analysis and incorporates behavioral, governance, and explainability dimensions. The model can support auditors, regulators, audit committees, and financial analysts in the early identification of earnings manipulation risks and in improving financial reporting quality.

The Impact of Intelligent Financial Technology Adoption on the Financial Performance of Small and Medium-Sized Enterprises: The Mediating Role of Financial Innovation

Fatemeh Sadat Ahmadi Nasab ^*

This study aimed to explain how the adoption of intelligent financial technologies affects the financial performance of small and medium-sized enterprises, with particular attention to the mediating role of financial innovation in Tehran-based SMEs. This qualitative study was conducted using an interpretive phenomenological design. The participants included 18 chief executive officers, financial managers, senior accountants, and digital transformation managers working in small and medium-sized enterprises located in Tehran. Participants were selected through purposive sampling. Data were collected through semi-structured interviews and continued until theoretical saturation was achieved. The interviews were transcribed verbatim and analyzed using thematic analysis with the assistance of NVivo software. Trustworthiness was enhanced through member checking, peer debriefing, recoding, prolonged engagement with the data, and maintaining an audit trail. Four main categories emerged from the data: strategic and technological readiness for intelligent FinTech adoption, digital reconfiguration of financial processes, financial innovation as a mediating mechanism, and performance outcomes and adoption-related risks. The findings indicated that the adoption of intelligent financial technologies does not automatically improve financial performance. Rather, its financial value is realized when firms transform technological adoption into financial innovation, including data-driven financing, digital payment mechanisms, real-time cash-flow management, alternative credit assessment, automated financial reporting, and predictive financial decision-making. Participants emphasized reduced financing costs, faster receivables collection, improved financial transparency, enhanced liquidity control, and greater financial flexibility as the most important performance-related outcomes. The results suggest that financial innovation functions as the key connecting mechanism between intelligent FinTech adoption and SME financial performance. Therefore, merely implementing FinTech tools is insufficient. SMEs need to develop organizational, data-related, human, and strategic capabilities that enable them to convert technological adoption into measurable financial value.

Designing an Intelligent Model for Detecting Financial Anomalies in Organizational Transactions Using Unsupervised Learning Algorithms

Mahdieh Sadat Karimi Pour ^*

This study aimed to design an intelligent model for detecting financial anomalies in organizational transactions by emphasizing unsupervised learning algorithms, data requirements, expert validation, and financial governance mechanisms. This study employed a qualitative, applied-developmental design. The participants included 18 experts in finance, internal auditing, risk management, data analytics, and financial information technology from large organizations located in Tehran. Participants were selected through purposive and snowball sampling. Data were collected only through semi-structured interviews, and the interview process continued until theoretical saturation was achieved. Thematic analysis was used to analyze the data through open, axial, and selective coding with the support of NVivo software. The credibility and dependability of the findings were strengthened through member checking, peer review, audit trail documentation, and constant comparison of codes and categories. The analysis of interviews resulted in five main categories: data architecture and transaction integration, algorithmic capabilities of unsupervised learning, interpretability and human validation, risk governance and internal control, and organizational implementation requirements. Participants emphasized that an intelligent anomaly detection model can be effective only when financial, operational, and behavioral data are integrated, cleaned, traceable, and analytically usable. They also stated that algorithms such as clustering, DBSCAN, Local Outlier Factor, Isolation Forest, and autoencoders can identify unusual transaction patterns without relying on labeled fraud data. However, algorithmic outputs should be interpreted, prioritized, and validated by financial experts, internal auditors, and risk management units. The proposed model indicates that intelligent detection of financial anomalies in organizational transactions is not merely a technical issue; rather, it requires alignment among data quality, algorithmic design, accounting knowledge, internal control, data governance, and organizational accountability. Unsupervised learning can enhance early detection of errors, misuse, control circumvention, and atypical transaction patterns, provided that it is implemented alongside human, ethical, and organizational oversight mechanisms.

Explaining the Effect of Social Media Sentiment Analysis on Investment Decisions and Stock Market Volatility through an Artificial Intelligence Approach

Hossein Nouraei Tabar , Ladan Sedighi ^* , Mohammad Erfan Sadeghi

This study aimed to explain how social media sentiment analysis influences investment decisions, investor behavior, and stock market volatility, with particular emphasis on the capabilities and challenges of artificial intelligence in interpreting market-related emotional data. This study employed a qualitative exploratory design. The participants consisted of 24 active investors, capital market analysts, portfolio managers, financial technology specialists, and artificial intelligence experts based in Tehran. Participants were selected through purposive and snowball sampling. Data were collected exclusively through semi-structured interviews, and sampling continued until theoretical saturation was achieved. The interviews were transcribed verbatim and analyzed using NVivo software based on the six-phase thematic analysis procedure. To enhance trustworthiness, member checking, peer review, analytic memoing, audit trail documentation, and constant comparison of codes were applied. Data analysis led to the identification of five main categories: “emotionalization of investor attention,” “network effects of rumor, excitement, and herd behavior,” “amplification or moderation of market volatility through intelligent algorithms,” “trust, bias, and manipulation challenges in sentiment data,” and “the need for analytical governance and digital financial literacy.” Participants emphasized that social media sentiment, particularly during periods of uncertainty, can reshape risk perception, accelerate buy and sell decisions, increase trading volume, and intensify short-term market volatility. At the same time, AI-based sentiment analysis, when properly validated, may serve as a complementary decision-support and early-warning tool. The findings indicate that social media sentiment analysis should not be regarded as a substitute for fundamental or technical analysis, but rather as an additional analytical layer for understanding market behavior. Its effective application requires the integration of artificial intelligence models with human judgment, data-bias control, algorithmic transparency, and the development of investors’ digital financial literacy

Smartification in Accounting, Economics, and Financial Management

About the Journal

Current Issue

Articles

Developing a Bankruptcy Prediction Model for Banks and Financial Institutions Using Recurrent Neural Networks and Financial Soundness Indicators

The Role of Explainable Artificial Intelligence in Enhancing Financial Managers’ Trust in Algorithmic Outputs and Investment Decisions

Designing an Ethical Framework for the Use of Artificial Intelligence in Accounting and Financial Management with Emphasis on Transparency, Responsibility, and Algorithmic Bias

Analyzing the Effect of Financial Process Intelligence on Reducing Operational Risk and Improving the Efficiency of Internal Control Systems

Presenting an Intelligent Model for Detecting Real and Accrual-Based Earnings Management Using Data Mining and Analysis of Managers’ Behavioral Patterns

The Impact of Intelligent Financial Technology Adoption on the Financial Performance of Small and Medium-Sized Enterprises: The Mediating Role of Financial Innovation

Language

Issues

Indexing & Abstracting

Make a Submission

Editorial Board

Journal Statistics

Keywords

Developing a Bankruptcy Prediction Model for Banks and Financial Institutions Using Recurrent Neural Networks and Financial Soundness Indicators

The Role of Explainable Artificial Intelligence in Enhancing Financial Managers’ Trust in Algorithmic Outputs and Investment Decisions

Designing an Ethical Framework for the Use of Artificial Intelligence in Accounting and Financial Management with Emphasis on Transparency, Responsibility, and Algorithmic Bias

Analyzing the Effect of Financial Process Intelligence on Reducing Operational Risk and Improving the Efficiency of Internal Control Systems

Presenting an Intelligent Model for Detecting Real and Accrual-Based Earnings Management Using Data Mining and Analysis of Managers’ Behavioral Patterns

The Impact of Intelligent Financial Technology Adoption on the Financial Performance of Small and Medium-Sized Enterprises: The Mediating Role of Financial Innovation

Designing an Intelligent Model for Detecting Financial Anomalies in Organizational Transactions Using Unsupervised Learning Algorithms

Explaining the Effect of Social Media Sentiment Analysis on Investment Decisions and Stock Market Volatility through an Artificial Intelligence Approach