A method, apparatus, system, and computer program, for operating a network planning tool. The method includes recognizing at least one predetermined criteria being satisfied or triggered, and learning at least one network attribute, to be used in subsequent network planning. Various types of network attributes can be learned, such as, e.g., network element parameters, site parameters, link parameters, demand parameters, optical parameters, general parameters, and networking parameters. Example attributes may specify, e.g., a maximum number of add/drop channels, at least one alarm threshold, a maximum number of sites per network ring, a maximum number of add/drop sites per ring, and a maximum ring circumference. Others can specify a maximum light path distance, a routing tendency, at least one grooming node, or the like.