Modern networks are no longer defined solely by physical towers and fibre optic cables, but are increasingly shaped by invisible forces of software, algorithms and distributed intelligence. As this digital transformation unfolds, automation has become the foundation for managing complex, large-scale infrastructures with greater efficiency and minimal human intervention. For instance, in April 2024, the Centre for Development of Telematics (C-DOT) partnered with IIT Jodhpur for automated service management of 5G network and beyond, using an artificial intelligence (AI) project, reflecting a research focus on AI-driven operations. Further, in March 2025, C-DOT launched the Samarth incubation programme for telecom/ICT start-ups, explicitly aimed at fostering domestic innovation for advanced network services.
Evolutionary stages of autonomous networks
This shift is best understood through a layered model of network autonomy, where each level builds upon the previous one to progressively reduce human control, while increasing machine-driven intelligence.
The journey begins at Level 1, which introduces basic automation to assist human operators. This includes tools that can generate performance reports, send alerts when certain thresholds are crossed, or automate simple and repetitive tasks. However, these systems are reactive and rely heavily on predefined rules. Human intervention is still required for decision-making and handling complex scenarios.
At Level 2, the network begins to handle specific tasks autonomously, under certain conditions. Automation is applied to repetitive and well-defined processes, such as load balancing or basic fault management. The system can execute these tasks based on predefined rules or simple AI models. While this reduces the operational burden on human operators, significant oversight is still necessary, especially when dealing with unexpected situations or complex decision-making.
Level 3 marks a significant advancement in network autonomy. The system gains the ability to perceive its environment and make decisions based on real-time data. It can predict potential issues and take corrective actions within certain domains without human intervention. For instance, in a radio access network, the system might adjust parameters to optimise performance, based on current traffic conditions. While the network can handle more complex tasks autonomously, human oversight remains critical for cross-domain coordination and handling novel scenarios.
At Level 4, the network achieves a high degree of autonomy across multiple domains. It can perform end-to-end service life cycle management, including design, deployment, optimisation and healing, with minimal human intervention. The system employs advanced AI and machine learning (ML) algorithms to predict and proactively address issues before they impact service quality. Cross-domain orchestration allows the network to coordinate resources and services seamlessly, enhancing efficiency and resilience.
Finally, Level 5 represents the pinnacle of network autonomy. The network operates entirely independently, managing all aspects of its life cycle across all domains. It possesses self-learning capabilities, allowing it to adapt to new situations and optimise its performance continuously. Human involvement is limited to defining high-level policies and objectives. The network can also handle unforeseen events, make complex decisions and evolve its operations without human input, embodying the vision of a truly autonomous network.
Zero-touch networks (ZTNs) align with Level 5 autonomy. Moreover, ZTNs represent an evolution in network automation, aiming for complete automation of network services without human intervention. This approach leverages AI and ML to manage tasks such as provisioning, monitoring and maintenance. The ZTN and service management framework embodies this concept, focusing on self-management and self-healing capabilities to handle the increasing complexity of modern networks.
For example, Airtel’s new core network roll-out explicitly includes advanced automation. Airtel has committed to using an automation framework so that new services can be launched with zero touch and core functions can be efficiently managed through automated life cycle tools. The roll-out expressly entails advancing autonomous networks by utilising generative AI (GenAI) for service orchestration and assurance. In simple words, the AI-driven packet core will enable greater network agility and reliability, meeting surging data demand with minimal manual intervention.
Further, Vi has begun automating 4G/5G network roll-out with software tools. By integrating a multiprotocol label switching solution, Vi will bolster the reliability and robustness of its network infrastructure. This collaboration will also enable Vi to efficiently manage high volumes of data traffic, offering consumers and businesses an improved level of network performance and stability. With these advancements, consumers can enjoy faster, more reliable internet connections for high-quality video streaming, online gaming, video calls and everyday browsing.
Building blocks for autonomous networks
The core building blocks for autonomous networks include AI/ML, intent-based orchestration, closed-loop automation and virtualised orchestration platforms. AI and ML are used to analyse vast network data (traffic patterns, user demand, equipment status) for fault prediction, anomaly detection and automated decision-making. For example, Vi has deployed an augmentednetwork automation platform, which uses AI to simplify the management of its complex multivendor, multitechnology and multilayered network. This platform acts as a multivendor self-optimising network, empowering Vi to manage and automate its network independently.
Intent-based networking is another key element. Instead of manually configuring devices, operators can specify high-level business or service objectives (intents) that the system translates into detailed network actions. The intent-driven orchestration lets providers define high-level business and service intents, which are then automatically mapped to configurations; and continuous assurance ensures that the network continues to meet these intents. In practice, this means AI-enabled orchestration layers design service blueprints and dynamically adjust network slices to satisfy service level agreements without manual coding. AI-driven intent systems can even autogenerate and validate service blueprints.
Further, closed-loop automation and orchestration tie these technologies together. Closed-loop automation refers to feedback loops, where monitoring systems detect issues (such as congestion or failures) and trigger automated remedies. In practice, this involves integrated orchestration layers that continuously observe the network, identify deviations and reconfigure resources automatically (for example, rerouting traffic and adjusting capacity).
Edge computing plays a crucial role in autonomous networks by bringing computation and data storage closer to the data source. This proximity reduces latency and bandwidth usage, essential for applications requiring real-time processing, such as autonomous vehicles and industrial automation. In India, telecom operators are increasingly adopting edge computing to support 5G and internet of things (IoT) deployment demands. Jio, for example, has enabled edge computing on its cloud-native 5G network at more than 50 facilities. Airtel is also developing edge servers, in partnership with cloud providers. The company is deploying an edge computing platform across 120 network data centres in 20 cities in India. This platform is aimed at enabling large enterprises across multiple industries to accelerate innovative solutions that deliver new value to their clients and operations securely at the edge. Airtel believes that edge computing can be significantly enhanced with 5G, and in India, it has the potential to cross $1 trillion in value by 2035.
Challenges
Autonomous networks are not without their challenges. Financial constraints remain a major hurdle, with legacy debts and high costs incurred from spectrum auctions over the past decade weighing heavily on operators. Existing network components may not support automation, necessitating significant upgrades or replacements from enterprises.
Achieving full autonomy is also constrained by a lack of maturity of physical infrastructure components. Many existing microwave antennas, optical fibres and
other hardware are not yet capable of supporting seamless automation across all scenarios. Integrating AI-driven automation into these existing systems demands extensive customisation, which can cause operational disruptions, increase costs and prolong implementation timelines.
Moreover, from a security standpoint, autonomous networks handle massive volumes of real-time data and make independent decisions, expanding the potential surface for cyberattacks. The protection of sensitive user data, including location and usage patterns, demands robust privacy mechanisms such as encryption and anonymisation to retain users’ trust.
Outlook
Looking forward, India is proactively laying the foundation for 6G technology, envisioned not just as a faster network but as a fully intelligent, self-evolving digital ecosystem. To realise this vision, networks must evolve to become self-configuring, self-optimising, self-healing and self-protecting, essentially embodying the core principles of autonomous networks. The government’s support is evident, with over 111 research proposals sanctioned for 6G development, covering diverse areas from new air interfaces to innovative network architectures.
While the path ahead is challenging, at the same time, opportunities are immense. By continuing to invest in cloud-native core networks, edge computing, AI/ML capabilities and IoT platforms, Indian telecom operators are well positioned to ride over the next wave of digital growth.
