<![CDATA[Accurate cash flow forecasting is critical for corporate decision-making, influencing investment strategies, dividend policies, and debt covenant compliance. Traditional statistical models such as ARIMA and accrual-based regressions struggle to capture nonlinear financial dynamics. At the same time, machine learning approaches, including LSTM and GRU, offer improved accuracy but face limitations in interpretability and governance. Against this backdrop, generative artificial intelligence (GenAI) has emerged as a transformative technology, capable of integrating structured and unstructured financial data, generating synthetic time series, and outperforming analysts in earnings prediction. This study conducts a systematic literature review of 55 peer-reviewed studies published between 2018 and 2025, employing PRISMA guidelines and bibliometric analysis. The findings reveal three significant insights: first, empirical research on GenAI in cash flow forecasting remains scarce compared to earnings and liquidity studies; second, evidence suggests significant potential for GenAI to enhance forecast accuracy and scenario-based analysis; and third, regulatory frameworks such as the EU AI Act and ESMA guidance emphasize the need for transparency, accountability, and auditability in AI-driven financial forecasting. The contribution of this study lies in synthesizing the current state of research, identifying a critical gap in the application of GenAI to cash flow forecasting in public companies, and proposing a research agenda that integrates methodological innovation with governance imperatives.]]>
