Verizon Advocates for Customer-Centric Network Automation at Ericsson Summit

At the Ericsson OSS/BSS Summit 2026, Praveen Atreya, VP of Network Systems at Verizon, highlighted a crucial disconnect in current network automation strategies: the gap between aggregate network performance metrics and actual customer experience. Atreya pointed out that network data, while often showing healthy overall infrastructure, can frequently obscure problems affecting individual users, such as latency issues. This means that traditional network monitoring and automation, even with AI overlays, may fail to address the real-world impact on end-users. This perspective is vital for practitioners because it forces a re-evaluation of how network automation success is measured. It's no longer sufficient to automate for efficiency or stability from a purely infrastructure-centric view. Instead, the focus must shift to how automation directly improves the customer's perceived quality of service. For DevOps and cloud engineers, this implies a need for more sophisticated observability and feedback loops that can correlate network events with user experience data, driving automation actions based on direct customer impact. This trend aligns with the broader industry movement towards intent-based networking and autonomous operations, where the network is expected to self-optimize to meet defined business outcomes, which increasingly include customer satisfaction. The integration of AI is not just about automating tasks but about enabling the network to understand, reason, and act on complex, often subtle, signals related to user experience. Atreya's proposed three-stage path—starting with closed-loop automation for repetitive tasks, moving to AI-assisted recommendations, and finally achieving cognitive autonomy where agents sense, reason, and act—provides a practical roadmap for this evolution. In practice, this means network engineers and architects should prioritize the development of automation systems that ingest and analyze user experience data alongside traditional network telemetry. This could involve integrating customer support tickets, application performance monitoring (APM) tools, and even direct user feedback into automation workflows. The trade-off might be increased complexity in data ingestion and correlation, but the benefit is a network that is truly customer-aware and self-optimizing. Practitioners should investigate AI/ML platforms that can handle diverse data sources and enable the creation of intelligent agents capable of making decisions based on holistic customer-centric views, moving beyond simple threshold-based automation.