Bold reality check: true 5G performance doesn’t guarantee business continuity. A flawless network can still leave a critical customer factory offline, showing that the real value lies in how subscriber experience translates to business outcomes. This is the core tension explored here: network-centric metrics miss end-to-end customer impact, so AI-driven operations must pivot to a subscriber-focused view to truly deliver value.
Ruth Brown, a principal analyst at Heavy Reading (now part of Omdia), examines how to bridge this gap. The piece, published December 4, 2025, highlights that while 5G networks meet traditional performance SLAs, real-world disruptions often occur in the end-user experience and across the service chain. The challenge is clear: operators must move beyond latency, jitter, and throughput to incorporate customer experience, application performance, and service impacts into decision-making.
The push toward customer-centric AIOps is ambitious but essential. Many current AIOps implementations remain immature, and the shift to subscriber-centric data introduces hurdles like vast data volumes, new tool adoption, data integration, and achieving end-to-end visibility. Yet the data shows a strong consensus: 99% of service providers believe AIOps should be driven by subscriber-centric data. A majority (58%) agree it adds value, but isn’t always essential, with 16% somewhat agreeing it helps in some cases. This suggests operators are still balancing ideal objectives with practical constraints, often stitching together subscriber insights from multiple sources or enriching legacy data to achieve a usable view.
Key benefits of subscriber-centric data include direct ties to business value (improved retention, higher Net Promoter Score), earlier detection and automated mitigation (per-user QoS and network slices), and the ability to tailor service management to subscriber or SLA priorities (industrial automation vs. consumer video). In other words, it enables proactive, personalized optimization rather than generic network-centric adjustments.
However, moving to this model requires overcoming visibility barriers in the 5G user plane. End-to-end visibility from RAN to core remains a major or significant barrier for many operators (31% major, 36% barrier, 29% somewhat). A striking 78% report only partial visibility across their 5G RAN and core, while 95% cite a lack of tools to monitor, analyze, and act on data. Data volume compounds these challenges, not because of sheer quantity but due to the effort and cost to capture, process, and correlate it at speed. Advanced AI approaches, including agentic AI, demand broad data access to deliver a complete picture of customer experience.
To realize TM Forum Level 5 autonomy, operators must blend AI/ML with accelerated computing and end-to-end visibility. The path forward involves balancing the appetite for comprehensive insights with cost realities, and actively adopting assurance capabilities that support automation at scale. There is appetite among operators to explore agent-based AI, but success hinges on robust foundational tooling and data access.
In practice, the customer experience becomes the ultimate performance metric. Operators agree on subscriber-centric data as the driver for AIOps, but practical barriers persist: limited end-to-end visibility, gaps in new tooling, and the sheer scale of data. The recommended strategy is to pursue assurance solutions that leverage accelerated computing, deliver end-to-end monitoring, and integrate AI-driven assurance to support operations. When combined with network metrics and customer experience insights, AIOps can yield actionable, proactive recommendations that improve retention, satisfaction, and service quality across all subscribers.
What this means for the industry is clear: a transition from purely network-centric monitoring to a holistic, customer-centric AIOps approach is not just desirable—it’s necessary for translating network performance into meaningful business outcomes. The trend toward agent-based AI and deeper end-to-end visibility will shape the next wave of automation, but only if foundational data access and processing capabilities keep pace with ambition.
This analysis is sponsored by RADCOM and draws on the 2025 Heavy Reading (now part of Omdia) AIOps Network Operator Survey. For operators seeking a practical path to subscriber-centric AIOps, the key steps are to (1) invest in end-to-end visibility tools, (2) integrate AI assurance with accelerated compute resources, and (3) consolidate multiple data sources to generate a true, end-to-end view of customer experience. As networks become more complex, the ability to align technical metrics with business outcomes will determine who leads in customer satisfaction and operational efficiency.
Are there aspects of subscriber-centric AIOps you’d challenge or reinterpret? Do you believe automation should prioritize per-subscriber guarantees even when it costs more, or should it optimize for aggregate service levels first? Share your perspective in the comments.
