Operators are making real progress with AI-driven automation, but most gains remain confined to specific domains. Scaling that intelligence across the network is the next challenge.
About a decade ago, the idea of running a business on a single-server architecture became obsolete with the rise of cloud-based, distributed infrastructure. Today, enterprise IoT connectivity is apprOperators are making real progress with AI-driven automation, but most gains remain confined to specific domains. Scaling that intelligence across the network is the next challenge.
For many telecom network leaders, cost efficiency has defined the agenda in recent years. It is no surprise, then, that agentic technologies are gaining traction across network operations.
But as competition intensifies in increasingly commoditized markets, the priority is shifting. Customer experience is no longer a downstream outcome of network performance; it is now a primary driver of network modernization.
Autonomous networks are central to this shift. They are not just about managing the growing complexity of distributed, software-defined infrastructure or delivering services at lower cost. More importantly, they enable operators to actively deliver and sustain customer experience, anticipating issues and resolving them before they affect the end user.
The business case is established but expectations are shifting
The economic case for autonomous networks is now well understood. Accenture’s latest research reveals that operators expect autonomous networks deliver operating expenditure reductions of up to 30% and capital efficiency gains of around 17%, to revenue uplift in the range of 11% as networks become more responsive and dynamic.
At the same time, investment is shifting. Spending on AI, particularly generative and agentic AI, is increasing as operators build the intelligence layer required for more autonomous operations. But expectations are evolving. The focus is moving beyond isolated use cases to looking holistically at how intelligence is created, connected and sustained across the network to deliver consistent outcomes, particularly on customer experience.
Progress is tangible but constrained by design
There is no shortage of progress being made by the telcos. Deutsche Telekom’s work with Google Cloud on AI-driven RAN optimisation shows how far closed-loop automation has advanced within specific domains. Its “RAN Guardian” agent is designed to analyse network behaviour in real time, detect performance issues and trigger corrective actions to improve reliability and customer experience, marking a clear step toward autonomous, self-healing networks.
AT&T is also extending AI beyond isolated use cases into a broader operational model, enabling teams to deploy agents across network engineering, customer operations and internal workflows. Operators such as Telstra are investing at the platform level, building unified data and AI ecosystems to embed intelligence across workflows and enable more coordinated optimisation.
These examples demonstrate real progress. But they also highlight a structural limitation: each delivers value within defined domains. Extending that intelligence across the full network and coordinating decisions end to end, remains significantly more complex.
According to Accenture research, only 22% of operators expect to reach advanced levels of autonomy by 2030, with even fewer anticipating fully autonomous networks in that timeframe. For many, that trajectory is intentional.
Networks still must be modernised while remaining resilient, secure and continuously available. Investment cycles, legacy architectures and operational risk all shape how quickly change can happen. In that context, non-linear progress is not a failure. It is a feature of transformation at scale.
The constraint is less about technology and more about cohesion
The barriers to scaling autonomous networks are well understood. But their combined impact is often underestimated. Legacy BSS and OSS environments continue to fragment data and limit interoperability, making it difficult for AI systems to operate across the network. Talent gaps in AI and automation engineering slow down design and deployment. At the same time, transformation efforts are often spread across organisational silos, with uneven levels of ownership and alignment. Together, these issues create a cohesion problem. Not a technology gap, but a system-level one.
From isolated intelligence to AI-native network platforms
Scaling autonomous networks requires more than embedding AI into existing processes. It requires a platform that is explicitly designed to deliver customer experience outcomes, not just improve network efficiency.
At its core, this means connecting how the network operates directly to how customers experience it and enabling the system to continuously optimise that experience in real time. This model operates across three layers:
- An outside-in view of network experience: Operators need continuous visibility into what customers are experiencing, not just what internal systems report. This requires combining independent, real-world signals such as crowdsourced data, benchmarking and real-time incident detection, to create an accurate, external view of service quality. Without this, AI systems are acting on partial information and cannot reliably prioritise what matters most.
- A cognitive layer that turns data into decisions: This layer integrates external experience data with internal network data to understand why issues occur, predict what will happen next and determine the best course of action. It moves operations from reactive troubleshooting to predictive, experience-driven decision-making.
- An execution layer powered by autonomous agents: Here, insights are translated into real-time, policy-driven action. Agent-based systems can resolve issues, optimise performance and maintain service quality without waiting for manual intervention, creating a continuous, closed-loop system across detection, decision and execution. This is what fundamentally changes with autonomous networks: the ability to act continuously on the network so that customer experience is maintained, not just measured or repaired after failure.
Together, these layers transform the network from a reactive infrastructure into a predictive, experience-led system that can continuously adapt and self-optimise.
The question now is what comes next?
Autonomous networks are increasingly becoming the fulcrum of new value creation in telecom. Their real value lies not in automation alone, but in the data intelligence that connects the network with the broader enterprise.
Successful operators will be those who can translate enterprise-wide insights into intelligent actions reducing costs, accelerating new services and continuously improving customer experience.
The views expressed in this article belong solely to the author and do not represent The Fast Mode. While information provided in this post is obtained from sources believed by The Fast Mode to be reliable, The Fast Mode is not liable for any losses or damages arising from any information limitations, changes, inaccuracies, misrepresentations, omissions or errors contained therein. The heading is for ease of reference and shall not be deemed to influence the information presented.
Jacopo Sebastiani is a Senior Managing Director and the Global Communications & Media Industry Lead at Accenture. With more than 28 years of experience, he has led complex business and technology transformations for some of Europe’s largest telecommunications and media companies. Jacopo combines deep industry expertise with a track record of driving growth, innovation, and large-scale reinvention for global clients.
Rebellions’ Sunghyun Park on Redefining AI Inference with High-Efficiency AI Infrastructure
The Fast Mode
9658 likes
The Fast Mode
@TheFastModeTWEETS28.4K
FOLLOWING3479
FOLLOWERS13.2K
