Due to the development of IMT-2020 networks, a large number of new applications and services have emerged, leading to network traffic’s explosive growth and much more complicated networks, which brings great difficulty to network design optimization. Many services are changing rapidly, and this requires the support of diverse service quality. Therefore, network operators need to dynamically design and adjust their networks based on the traffic distribution and the traffic increment trend, in order to guarantee the QoS. On the other hand, AI can be used for prediction and decision-making, which can make network design more intelligent and automated. This Recommendation focuses on using AI/ML to improve network design mechanisms, which deals with how to integrate AI/ML in order to optimize the design of network capacity, network topology, and routing for satisfying all the demands’ SLAs in a cost-effective way, instead of only guaranteeing the SLAs without taking into account the overall cost.