Ericsson calls for unified AI architectures to scale autonomous networks

Anand Palkhiwala, Head of Cloud Software & Service, Americas at Ericsson, argues that CSPs must move beyond isolated AI deployments and adopt a single, integrated architecture to scale AI and network autonomy. Key requirements include a unified data foundation, a common agentic AI platform, pragmatic use-case prioritisation and automated conflict management. He also emphasises autonomous domains, standards-based APIs and closed-loop assurance, enabling operators to deliver differentiated services, reduce operational costs and convert business intent into measurable, assured outcomes.