Implementing autonomous networks offers an average CSP cost savings and annual revenue uplift of around US$794 million. However, challenges impede progress to realise this financial potential. This report addresses five of the main obstacles faced by CSPs and provides recommendations to help overcome them.

The long (and promising) road to network autonomy

CSP networks are growing in complexity. The amount of data they generate and require to function is increasing and network-tied services are growing more diverse (for example, private networks and network slicing). Steve Jarrett, Chief AI Officer at Orange, reveals that they are managing petabyte-scale data daily from network telemetry alone. Handling this complexity demands change. Traditional human-driven operational models fall short of managing this dynamic operating environment with the coexistence of legacy networks and next-generation technologies.

Taking AI and automation to the next level, to autonomous networks, requires a fundamental shift from executing simple rules-based tasks to orchestrating complex processes with minimal human oversight. This transition can create faster and error-free network operations. To achieve this, CSPs must implement automation across domains (access, transport, and core networks) and vertical stacks, and implement advanced intelligence for more agile learning and decision-making. Achieving autonomous networks is imperative to creating more effective business outcomes. This includes ensuring services never fail by uncovering hidden issues and pre-empting problems that would have not otherwise been anticipated, and reducing the dependency on humans to operate networks, enabling CSPs to develop more state-of-the-art services.

If you are not a subscriber, enter your details below to download an extract of the summary deck

The TM Forum’s Autonomous Network Maturity Model is a globally recognised framework used to evaluate CSPs’ advancement toward network autonomy (see the graph below for a detailed breakdown). Many CSPs have achieved Level 1 or 2 autonomy to date, with most network operations still facilitated by some degree of human intervention or oversight. CSPs want to pursue system-wide automation to reach Level 4 autonomy and beyond, but they face challenges.

TM Forum’s autonomous networks maturity model and example use cases

To move beyond Level 3 autonomy, CSPs must pivot from intra-domain automation to inter-domain automation. Intra-domain automation are siloed automation efforts focusing on the automation of isolated workflows or processes. In comparison, inter-domain automation focuses on the integration of automation across domains, teams, and processes. To achieve highly autonomous networks (Level 4 or 5), CSPs must also add intelligence into their automated processes. For instance, instead of using automation to do the same process as previously would have been done manually, adding intelligent decision-making systems means that the processes continuously change and adapt to deliver the most efficient and accurate outcomes. These shifts (intra- to inter-domain, simple automation to intelligent automation) can unlock significant opex and capex savings, and new monetisation opportunities through better allocation of human capital towards service innovation.

This report leverages key insights from CSP interviews and secondary research to set out the financial benefit of introducing automation and intelligence in the networks and identifies the conditions CSPs need to create to progress to greater autonomy.

Table of contents

• Executive summary
• Introduction
• The US$800 million opportunity: The estimated impact of network autonomy
• CSPs must overcome five barriers to achieve autonomous networks
  • Usefully interpreting industry frameworks and building a business case for nascent opportunities
  • Accessing the right data at the right time from multiple domains and systems
  • Ensuring people trust the systems to deliver high-quality outcomes
  • Overcoming the unique risks of AI-driven innovation in mission-critical environments like the network
  • Separating fact from fiction: Understand the role of generative AI in the network
• Conclusions and recommendations

Related research

• Hyperscalers in the telco vertical: Strategies and successes
• Telco Cloud Deployment Tracker Q2 2024: Is AWS winning in hyperscale telco cloud?
• Finding value from AI, analytics and automation (A3) in the telco network
• Advancing sustainability through network automation

Harine Tharmarajah

Harine Tharmarajah

Strategy Consultant

Harine is a Consultant at STL Partners, who joined after completing her undergraduate studies. She earned a First class honours degree in BSc Economics from University College London. Since joining STL Partners, Harine has worked with telecoms and technology companies on strategic engagements in network transformation, but also works within STL's Edge practice.