Telecommunications operators have long prioritised expanding radio access and boosting network speeds, leaving the mobile core as an afterthought. That calculus is shifting as artificial intelligence opens fresh avenues for network optimisation and new business models. Dejan Leskaroski, Chief Revenue Officer at Mavenir, contends that the core's historical neglect masks its critical role in everything from subscriber management and authentication to policy enforcement, billing and security.
According to Leskaroski, "Core has definitely been overlooked in the industry. Once you have radio access, everything from mobility management and subscriber authentication to policy enforcement, billing and security happens in the core." As automation, AI and cloud technologies reshape network operations, operators possess a chance to extract greater value from infrastructure already in place. The challenge lies in whether they can adapt quickly enough to seize that opportunity.
Learning from 5G's missed opportunities
The industry has encountered revenue promises before. When 5G emerged, network slicing, ultra-reliable low-latency communications and enterprise applications were touted as generators of fresh commercial opportunities. Converting those technical capabilities into actual revenue proved far more difficult than anticipated.
Leskaroski observes that "Many of the promises around revenue generation and monetisation in 5G, particularly 5G Standalone networks, were tied to specific 5G capabilities. There was considerable dependency on the core, the radio access network and devices." Even when operators deployed the necessary technology, they faced obstacles: the need for compatible handsets, sufficient 5G standalone uptake and alignment across multiple vendors. These dependencies complicated service launches beyond initial expectations.
AI presents a potential escape from some of these constraints. Rather than awaiting the next network generation or waiting for industry standards to mature, operators could deploy services functioning across current networks. Leskaroski notes that "Take HD voice, voice translation, enhanced security or power savings. None of these is inherently a 5G or 6G concept. Moving towards what you might call an AI-native core, or an AI core by design, opens up opportunities to monetise use cases that aren't necessarily telecommunications-specific."
Making better use of the data operators already have
A substantial opportunity resides in network data that operators have collected for years. Leskaroski explains: "The data has always been there. Everything from subscriber awareness and location awareness to application usage has always passed through the core. Traditionally, we've largely been focused on getting that data from point A to point B." AI enables operators to process enormous data volumes and respond with greater speed, transforming data from a byproduct into a strategic asset.
Security represents another domain where operators could leverage the core more effectively. As AI-driven threats grow more sophisticated, services including deepfake voice detection, spam prevention and rogue agent identification gain urgency. Rather than outsourcing these functions entirely to external vendors, operators could potentially deliver them directly through their networks.
Leskaroski cautions against overstating AI's role in these improvements. He states: "Automation and service assurance existed long before AI. AI didn't introduce those concepts. What it has done is accelerate how quickly operators can perform certain tasks and realise the benefits."
Modernising the core is no small task
Core networks present formidable technical challenges. They typically comprise equipment and software from multiple vendors and support various generations of mobile technology. Leskaroski notes that "The core is different. It's typically a multi-vendor environment involving at least two or three vendors, sometimes more." Introducing new capabilities becomes particularly difficult when they must coexist with established billing systems, customer platforms and operational workflows.
Simply grafting AI onto existing networks will not resolve these underlying structural problems. Leskaroski emphasises: "You can't simply rip everything out and replace it overnight. There needs to be an evolution path that allows operators to modernise their networks in the right way, enabling automation and creating a path towards autonomous networks."
Could charging become the next bottleneck?
Beyond the core itself, Leskaroski identifies billing and charging systems as deserving greater scrutiny. Numerous operators continue operating legacy billing and charging platforms that have undergone extensive customisation over decades. While functional, these systems resist modification due to risk and complexity. Leskaroski warns: "Charging is the next frontier that needs to be tackled. Right now, the attitude towards charging and billing is often: 'Don't touch it. It works.'"
This hesitation could become problematic as operators experiment with novel service charging approaches, including token-based models tied to AI capabilities. Leskaroski concludes: "The danger is that we could fall into the same trap we experienced with 5G. If charging systems don't evolve alongside the network, operators may find themselves unable to monetise new services, new business models or token-based charging opportunities."
