For SMEs, financial distress can emerge even when sales and accounting profits remain positive because the timing and uncertainty of cash movements determine whether obligations can actually be funded. This study investigates artificial intelligence-based cash-flow forecasting as a mechanism for identifying liquidity turning points before conventional distress indicators deteriorate. The proposed approach reconstructs enterprise cash positions from recurring and irregular receipts, payroll, supplier payments, taxation, debt service, capital expenditure, seasonality, and customer-payment uncertainty. Probabilistic forecasting generates alternative future cash trajectories rather than a single deterministic estimate, enabling calculation of liquidity runway, minimum projected cash balance, shortfall probability, cash-flow-at-risk, and time-to-liquidity breach. These forward measures are used to distinguish temporary cash compression from persistent trajectories associated with emerging financial distress. Scenario simulations examine how delayed receivables, revenue shocks, cost increases, or financing constraints alter the timing and severity of liquidity deterioration. Management responses can consequently be evaluated according to their capacity to extend liquidity runway and reduce shortfall exposure. The framework positions AI forecasting not simply as a prediction tool, but as a forward-looking financial resilience mechanism connecting uncertainty, distress lead time, and liquidity planning in SMEs.