Embodiments of the invention are directed to systems, methods, and computer program products for utilizing machine learning to calculate product replacement cycle times and predict resource transfers associated with said product replacements. As such, the system allows for receipt of resource transfer datasets which are processed along with historical datasets via a machine learning engine. The system may then identify data trends and generate predictions of future resource transfers associated with product or service lifecycles. Furthermore, the system may also identify potential merchants or other third-party systems which may be associated with the predicted future transfers.