Lead propensity prediction is a data-driven method used to define the value of prospects, by assigning points to them based on their engagement with the business's digital channels, based on multiple key attributes correlating to their attraction to the proposed services or items. The resulting score is closely related to the financial worth of each lead and may be revealing its position in the buying cycle. The marketing teams can then focus on generated leads and prioritize the most prominent ones to improve the conversion rates, using the assigned score on the lead scoring step. The authors investigated using a combination of a data-driven approach and Artificial intelligence (AI) techniques for the lead-scoring process. The experimentation shows that the random forest (RF) is the most suitable model for this task with an accuracy score of 93.04% followed by the decision tree (DT) model of 91.47%. In contrast, when considering the training time, DT and logistic regression (LR) needed a shorter time to learn from the dataset while maintaining decent performances. In contrast, these models represent promising alternatives to the RF model especially in the case of a huge volume of transactions and prospects or in a big data context.