Artificial intelligence (AI) is a key enabler for future 6G networks. Currently, related architecture works propose AI-based applications and network services that are dedicated to specific tasks (e.g., improving the performance of RAN with AI). These proposed architectures offer a unique way to collect data, process it, and extract features from data for each AI-based application. However, this dedicated approach creates AI-silos that hinder the integration of AI in the networks. In other words, such AI-silos create a set of AI-models and data for AI-based applications that only work within a single dedicated task. This single-task approach limits the end-to-end integration of AI in the networks. In this work, we propose a network architecture to deploy AI-based applications, at different network domains, that prevents AI-silos by offering reusable data and models to ensure scalable deployments. We describe the architecture, provide workflows for the end-to-end management of AI-based applications, and show the viability of the architecture through multiple use cases.