Artificial İntelligence in Health Marketing Personalization: Integration with HR Systems to Predict Healthcare Staff Training Needs

The increasing demand for healthcare services, along with the rising complexity of the skill set of the workforce, has made it necessary to adopt data-based methods of human resource planning. This study develops an integrated predictive–prescriptive framework that systematically links healthcare service demand forecasting to the optimal allocation of staff training programs. This applied quantitative research was performed at Imam Khomeini Hospital with 12 months of anonymized operational data. In the first stage, multiple forecasting models (XG-Boost, Random Forest, LSTM, and ARIMA) were trained and evaluated using rolling-origin cross-validation. Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) were the basis for model selection, and performance was measured by comparison with a straightforward benchmark model. In the second stage, forecasted service demand was translated into time varying skill requirements through a service–skill matrix. A multi-objective mixed-integer quadratic programming model was then formulated to minimize overall skill gaps and total training costs subject to budgetary and capacity constraints. Scenario and sensitivity analyses were performed to see if the results would still be valid in the face of budget and training capacity constraints, changes in demand, and different levels of course effectiveness. The chosen forecasting models obtained very good accuracy levels (MAPE within 1. 6% to 2. 6%) and in all cases showed greater performance over the benchmark. According to the baseline scenario, there were 87. 8% and 80. 0% reductions in the gaps of two critical skills, respectively, while the levels of non-critical skills remained largely the same. A 20% budget reduction reduced the overall gap to 55.6%, whereas a 20% budget increase increased it to 92.2%. In a combined stress scenario (reduced budget and increased demand), the framework still closed approximately 70% of the initial gaps, indicating stable and interpretable behavior under adverse conditions. By integrating AI-based demand forecasting with a prescriptive training optimization model, the proposed framework enables proactive, evidence-based workforce planning in healthcare organizations and supports more targeted, cost-effective staff development.