By Jürgen Hatheier, Ciena’s International Chief Technology Officer

AI was all the buzz at this year’s Mobile World Congress show in Barcelona, with dozens of speakers, demonstrations and tradeshow floor chatter exploring how the technology will impact just about every company in the telecoms space. There was perhaps a sprinkling of overhype, but the attention on AI is hardly unjustified. The foundations are being laid now for AI to provide transformative change for telecom networks, and consequently, business models alike.

From my perspective, there are three lenses through which to view AI’s role in telecoms right now. The first is all about reducing OPEX, the second is embedding AI into networks to make them better, smarter and drive network efficiencies, and the third is AI as a new revenue opportunity for telcos.

First Lense; AIOps in Telecom: Efficiency and Innovation with AI

Using AI to help make operations more efficient – also known as AIOps – was the most prominent use case spoken about on the booths of Barcelona’s Fira. This includes using AI and machine learning (ML) to solve key operational challenges, as well as common use cases for generative AI technologies and large language models (LLMs) which have become rooted in consumer consciousness thanks to ChatGPT and similar tools. 

AI and ML use in AIOps is not new of course. Both technologies have been commonplace in inventory, network orchestration and assurance for more than half a decade. However, a telco’s strategy must revolve around applying the right AI technology to the right use case, including traditional unsupervised, supervised, and reinforcement learning, as well as Generative AI.

Using LLMs specifically for customer care and marketing materials has been widely spoken of. AT&T and Vodafone are among those that are using AI chatbots to help customers already. Others are exploring opportunities designed to boost productivity and proficiency when working with product documentation, and things like database reconciliation. Organisations often have several databases that contain partial information and contact details that can be consolidated to create one “source of truth” and remove duplication.

More use cases under investigation include leveraging LLMs for code generation and to provide a natural language interface for optimising service lifecycle automation, as well as for real-time data access to visualise network telemetry.

Data remains a key ingredient for every LLM and as we see specialised applications beyond the likes of ChatGPT or Claude (which primarily feed on public information). Combining industry-specific data into training sets for LLMs is crucial. We see this happening with the likes of the AI telco alliance which includes operators from EMEA and APAC

Second Lense; AI-Powered Network Optimization and Configuration for Telecoms

The second lens through which to view AI’s impact on telecoms is how it can improve network performance. AI holds the promise to improve network visibility, efficiency and planning.

Providers are now looking to automate networks using real-time performance metrics, making them adaptable to today’s increasingly dynamic service requirements.

Network management, control and planning has been made software-centric and is being enhanced with automation technologies, allowing operators to detect abnormalities and diagnose them. Large-scale networks are being made more efficient, with AI helping optimise networks for the lowest absolute cost, number of wavelengths, and crucially, for the lowest energy consumption.

AI can also thrive in conducting management and maintenance operations. Self-healing networks are the next step in intelligent networking, enabling the network to completely repair (and potentially even reconstruct) itself or reroute in a matter of seconds, should a failure occur. I believe self-healing networks will become the de-facto standard in the not-too-distant future. Using real-time data analysis, AI will compress decision-making timelines by orders of magnitude, minimizing or even eliminating disruptions from damaged cables or attempted network intrusions to save service providers significant downtime and revenue losses.

Beyond healing, AI will also enable to network to configure itself, based on intent, without complex configuration of every box. This will simplify operations, allowing highly skilled resources to focus on developing new revenue-generating services and customer offerings.

The end goal of improving network performance and efficiency would be for an operator to be able to cater for a major sporting event or festival without requiring a lot of manual effort, or even wasted effort. AI can help with network bandwidth management and provide deep insights in real-time. This helps service providers properly allocate bandwidth to meet demand, getting resources into the ‘Goldilocks’ zone where not too little or too much is assigned. Should an unexpected spike in demand occur thanks to a viral moment or similar, AI can also reallocate resources and pivot network management to deal with the changing situation.

The network needs to be capable of adapting and that is where AI plays an important role.

Third Lense; Beyond Connectivity: Telcos Harness AI for Growth

The last lense through which to view AI’s impact on telcos, is how it can be harnessed as a revenue-generating opportunity. This is the area where there needs to be more focus, rather than purely cost down, AI needs to fuel revenues up.

The first and perhaps most straightforward way to do this is by utilising AI to aid in cross-selling services more effectively, identifying and exploiting sales opportunities for new and existing customers.

Beyond that is where things get more exciting, with telcos being able to position themselves akin to the hyperscalers as innovative, forward-thinking AI companies that customers go to for advice – and services. This can create an entirely new product stream, with LLMs built on proprietary data being one piece of the puzzle, alongside dedicated network slices or private networks laced with AI provision for optimised and efficient performance.  

And while the above scenario is an ambitious one to achieve for any telco, there are various models in between, where partnerships and coopetition with hyperscalers at the edge become reality. Fundamental challenges like different commercial models need to be overcome with an appetite to increase the risk taking on the side of the telcos.

AI has of course been around for years, but the recent surge in popularity from consumerised services like ChatGPT and Midjourney has increased usage and heightened demand for even more. That means both individual and enterprise customers are more aware of AI-branded tools and are associating them with innovation. The door is open for telcos to associate themselves with that innovation too.

Opportunity requires strategy

AI was touted in nearly every booth at this year’s Mobile World Congress show. The potential of the transformative AI opportunity is overwhelming and requires a complete review of companies’ purpose, position, and desired outcomes. In essence, how do you want to participate in the new ecosystem?

Telcos are working on absorbing artificial intelligence technology into their strategies so they can truly benefit and use AI to shape the networks and customer services of tomorrow. Many are trying to rationalise what their strategy is and what the vendor community is doing. There is a role for GenAI, and “traditional” AI, but the right technology for the right use case is needed, with careful consideration for power consumption and data privacy.

Underneath the hype, there is a genuine transformative change to be had for telco business models. It is an absolute requirement to adapt not only the organisational strategy and optimise operations, but also the commercial models to truly reap the rewards of the AI opportunity.

It’s time to seize the AI opportunity with boldness to truly benefit from the latest digital revolution.

Image courtesy of Matressa on Pixabay