From cloud to cognition: the rise of the AI-native telco

The next frontier of telco evolution is already emerging: the AI-native networks. Industry leaders from AWS, Dell Technologies, BT Group, Orange, and Rakuten Symphony recently came together to explore what it means to move beyond automation and truly embed intelligence into the telco fabric. Their discussion focused on AI enablement across the network lifecycle, the importance of clean data pipelines, and the operational shifts required to support AI-native infrastructure.

Speakers:

• Prof. Mohammad Zoualfaghari, Global Head of Emerging Technologies for Telecom, AWS
• Neil Bowden, AI Business Development Executive, Dell Technologies
• Nektaria Efthymiou, Network Platform and Security Research Director, BT Group
• Sana Ben Jemaa, Technical Lead of AI Empowered Networks Program, Orange
• Vivek Murthy, President, OSS Business Unit, Rakuten Symphony

Watch the full interview.

What does “AI-native” actually mean?

The panel agreed that AI-native is not about sprinkling AI on top of legacy networks. It’s a fundamental shift in how networks are designed, deployed, and operated, from static to dynamic, reactive to predictive, rule-based to self-learning. Unlike automation, which relies on predefined scripts, AI-native networks can adapt based on patterns, context, and continuous learning.

This transformation requires cloud-native infrastructure, real-time observability, and a reimagining of the operating model. It also means trusting the AI to act and designing safety, accountability, and transparency into every layer.

Making it real: the building blocks of an AI-native telco

The transition to AI-native is as much about operational mindset as it is about technology. For AI models to be effective, they need clean, high-resolution data that is collected continuously, labeled accurately, and shared across domains. Telcos must invest in data engineering, not just data science.

AI-native also means defining clear roles: what tasks are performed by AI, which are overseen by humans, and where joint decision-making is needed. As networks become more autonomous, these boundaries must be tested, codified, and continuously evaluated.

Another key shift is the need for platform thinking. Instead of deploying AI in pockets (for fault prediction or energy optimization) operators must think horizontally, enabling AI services that are reusable, composable, and scalable across the network stack.

Key takeaways

• AI-native is a mindset shift. It requires more than deploying models; it’s about designing for continuous learning, adaptation, and decision-making at scale.
  
  
• Data is foundational. Clean, real-time, and labeled data is essential for training reliable AI models and ensuring accurate outputs.
  
  
• Cloud-native is the prerequisite. Elastic compute, containerization, and API-based architectures lay the groundwork for scalable AI.
  
  
• Operational models must evolve. AI ownership, governance, and explainability are critical as decisions shift from human to machine.
  
  

• Platform thinking unlocks scale. Building AI services as reusable components accelerates adoption across domains and teams.

“The way our operational users interact with the systems is going to fundamentally change with GenAI. So having an NLP based system where they can actually, for example, tell us the sites which are facing call drops today and so on, feels like the most obvious thing to provide as an interface and a service.”

-Vivek Murthy, President, OSS Business Unit, Rakuten Symphony

Stay tuned for more insights from industry leaders on the future of telecom and technology. Follow us on LinkedIn for updates on upcoming discussions.