Software Networks built by combining Software Defined Networks (SDN), Network Function Virtualization (NFV) and Cloud principles call for agile and dynamic automation of management operations to ensure continuous provisioning and deployment of networked resources and services. In this context, efficient Service Level Agreements (SLA) management and anticipation of Service Level Objectives (SLO) breaches become essential to fulfill established service contracts with clients. In this paper, we design and specify a framework for cognitive SLA enforcement (using Artificial Neural Network learning) for networking services involving VNFs (Virtualized Network Functions) and SDN controllers. A proof of concept, a testbed description and an extensive evaluation assess the performance of the proposed framework.