As CSPs deploy cloud-native 5G networks, they are increasingly able to make dedicated slices of connectivity available to specific customers or applications. This Catalyst will enhance existing standard APIs to help CSPs capture the initial slice profile and then use machine learning to capture all the information they need to manage the slice and perform dynamic orchestration. To that end, the project team will develop a proof-of-concept (PoC) that will map the customer’s intent to a slice. The PoC will define the APIs required to manage a slice across multiple networks, and the APIs required to enable a machine learning algorithm to simultaneously balance multiple intents (as defined by the TM Forum’s TR290A model). The proposed PoC will be built in a sandbox with API calls to network functions stubbed out. This Catalyst, now in its fourth iteration, builds on the previous phase, which developed a more comprehensive understanding of intent APIs to unlock the potential business benefits that can be generated using intent in autonomous networks, such as improved customer experiences and business outcomes. The goal of this phase of the Catalyst is to enable CSPs to provision, operate and maintain a slice using TM Forum’s standardized ‘intent grammar’ in an autonomous network environment. The project team plans to demonstrate the intent grammar-translation of an ‘expectation intent’ to a ‘system intent’ for provisioning a slice, while developing specifications that help maintain the slice's state and take actions autonomously when the intent or service level agreement is not met.