Telecom operators have started looking at artificial intelligence (AI), particularly generative AI (GenAI), as a core component of their business operations. Incorporating GenAI into telecom has the potential to revolutionise business operations, elevate customer experiences and enhance network management. GenAI can drive telcos towards more agile, data-driven solutions that can redefine the future of connectivity.

Telcos view the integration of GenAI as a two-step process:  first, embedding AI into the company’s operational activities, and second, using AI to create new services and enhance existing ones to provide greater value to their customers.

According to the latest report by Capgemini, three out of five telcos globally are exploring the potential of GenAI for autonomous networks. This will help them in numerous ways – by improving the reliability and resilience of the network, as well as handling traffic fluctuations and network anomalies in real time.

GenAI, particularly through chatbots, is widely being utilised by telecom companies to enhance customer experience, boost agent productivity and facilitate fully digitised transactions. According to McKinsey, telcos are upgrading their AI chatbots to improve agent support, expecting to achieve cost reductions of 15-20 per cent. Additionally, they are leveraging GenAI to summarise both voice and written customer interactions, with an anticipated cost reduction of up to 80 per cent.

GenAI is also being used to assist in complex event processing, which involves aggregating data and analysing cause-and-effect relationships between events in real time or near-real time. This method identifies patterns in event relationships and extracts insights, enabling users to make informed actions. The integration of large language models (LLMs) can assist field engineers in querying data and efficiently correlating issues with their potential causes.

Telcos can even utilise GenAI and machine learning algorithms to enhance resource utilisation and manage network traffic fluctuations in real time. These algorithms identify bottlenecks and optimise traffic routes by analysing extensive traffic data to dynamically allocate bandwidth, improving network efficiency while also lowering energy costs and carbon emissions. For instance, GenAI can predict the likelihood of a thunderstorm and estimate the chances of it impacting a telecom tower well in advance. This allows the telco to proactively deploy solutions much earlier than they could with traditional analysis, ensuring uninterrupted service delivery.

GenAI can also aid engineers stationed at network operations centres by analysing network data and providing actionable insights for decision-making. It can quickly convert complex data into clear, understandable text, thereby enhancing the efficiency and effectiveness of operators. The integration of GenAI can also expedite the preparation and despatch of documentation related to fraud, facilitating quicker legal interventions.

The use of GenAI enables telcos to quickly prototype and deploy new solutions, reducing development cycles from months to just days or hours. As telecom operators explore GenAI’s capabilities, they are discovering its potential to significantly enhance customer experiences by providing more personalised and responsive services, boosting customer loyalty and retention, and opening up new revenue streams.

However, the rapid adoption of GenAI also brings considerable risks, particularly in terms of technical debt. In the race to implement these advanced technologies, many companies overlook the long-term upkeep that these systems demand. Every new update or feature adds complexity, making future modifications more challenging and prone to errors. Striking a balance between embracing new possibilities and managing the ongoing need for upgrades is essential for telecom operators, moving forward.

The key to effectively utilising GenAI lies in a measured approach, where AI-driven solutions are thoroughly tested and validated incrementally. This ensures that they meet the necessary standards of reliability and accuracy, becoming an asset rather than a burden. Leveraging GenAI to automate tasks and free up human resources could be a critical solution to maximising its benefits.

There are several significant challenges that telcos may encounter when implementing GenAI within their operations. A major concern is that most advanced LLMs are publicly available, raising serious concerns about security risks when working with sensitive customer and network data. Training these models with telco-specific information requires utmost caution to prevent breaches in confidentiality. Additionally, the dynamic nature of telecom data, which is often real-time and constantly evolving, makes it difficult to apply conventional “fine-tuning” techniques that are typically used to adapt pre-trained models for domain-specific tasks. This means that the rapid changes in telco data are not always well suited to existing models that rely on static data sets.

Moreover, the accuracy of GenAI models is entirely dependent on the quality and relevance of the data they are trained on. Since these models lack inherent knowledge of the telecom industry and its complex processes, there is a risk of incorrect assumptions, which can lead to inaccurate predictions or decisions. This challenge is particularly pronounced in a sector where precise outcomes are critical for ensuring smooth network operations and customer satisfaction.

The way ahead

While GenAI offers transformative potential for the telecom industry, its benefits will only be realised if implemented securely and effectively. Telcos must navigate various challenges in adopting AI models, including establishing a closed-loop system where data not only provides insights but also helps continuously refine AI. Financial considerations are also crucial, requiring cost-effective strategies such as balancing between predictive AI and GenAI to avoid excessive expenditures. For large telecom operators with extensive data, defining and executing effective GenAI use cases is essential. Additionally, while GenAI can enhance customer and network services, it must comply with regulatory requirements and data governance standards. Ensuring security and scalability is vital in a sector where data protection is critical, and AI implementations must meet stringent security protocols.

Moving ahead, telecom providers that adopt GenAI early are more likely to experience faster growth and capture a more significant share of the nearly $100 billion  incremental value, as forecasted by McKinsey. This is on top of the projected $140 billion-$180 billion in productivity gains that GenAI could unlock in the telecom industry compared to what traditional AI offers. Moreover, telecom leaders must remain vigilant about the sustainability and ethical implications of AI.