In this paper, we introduce a novel, profit-driven classification approach for churn prevention by framing the task of targeting customers for a retention campaign as a regret minimization problem within a predict-and-optimize framework. This is the first churn prevention model to utilize this approach. Our main objective is to leverage individual customer lifetime values (CLVs) to ensure that only the most valuable customers are targeted. In contrast, many profit-driven strategies focus on churn probabilities while considering average CLVs, often resulting in significant information loss due to data aggregation. Our proposed model aligns with the principles of the predict-and-optimize framework and can be efficiently solved using stochastic gradient descent methods. Results from 13 churn prediction datasets, sourced from an investment company, underscore the effectiveness of our approach, which achieves the highest average performance in terms of profit compared to other well-established strategies.