By Anthony Goonetilleke, Group President of Technology and Head of Strategy, Amdocs.  

Keeping up with the times is becoming even more challenging as the pace of technology accelerates to unprecedented levels. Those who fail to adapt risk becoming irrelevant in a fast-moving market. The exponential growth in generative AI (GenAI) over the past year highlights just how quickly technology is moving, redefining not only how we work, but increasingly, how we live.

One sector in which the growth of GenAI will be particularly impactful is telecoms. Connectivity is critical to almost every aspect of our lives – and communications service providers (CSPs) need to embrace its capabilities across their customer experiences, networks, operations, and IT systems.

Research by management consultancy firm Analysys Mason, commissioned by Amdocs, recently found that 90% of CSPs have a clear understanding of the profound impact GenAI will have on their business goals. However, just 22% have implemented it to date, with 32% currently running proof of concepts and 29% looking to explore it in the next 12 months – but at the current pace of change, the world could look very different in 12 months.

Generating benefits

I’m fortunate to meet C-suite executives from some of the largest telecoms and technology companies around the world, including those at the forefront of GenAI development efforts. My conversations with them often end with the same conclusion: the benefits of GenAI are clear and significant, but there are challenges to overcome its most significant potential benefits and deploy at scale.

Analysys Mason found that 84% of those working on data, AI, and IT for CSPs believe they will struggle to acquire the capabilities required to deploy GenAI successfully, and 81% consider integrating GenAI with current business processes the greatest challenge – not only from a technology perspective, but from an organizational change management and adoption perspective, which internally we refer to as Guided Disruption.

The telecom industry impacts most people’s daily lives, so the delivery of carrier-grade experiences, trusted AI, and high accuracy are fundamental to embracing and monetizing GenAI. Ensuring these challenging requirements are met demands the development of telco-specific GenAI capabilities built to meet the industry’s distinct demands and use cases. This will help to manage hallucinations, control emergent behavior, create interactions aligned with content guardrails, and reduce bias – all while ensuring observability and consistency, which is essential in large-scale and customer-facing implementations.

Taking a step back

Our research uncovered that CSP leaders consider alignment with compliance, privacy, and security regulations, the need to train LLMs with telco-specific data, and access to high-quality data to be some of the most significant hurdles to successful GenAI deployment. The importance of high-quality data is especially crucial as the results that GenAI tools provide can only be as accurate and insightful as the data they are built on.

Fortunately, telecoms companies don’t struggle with a data shortage. But they do need to prioritize data readiness, as the data they hold is often complex and difficult for GenAI tools to leverage. To overcome this, telco leaders must focus on creating a viable data ingestion strategy that involves a robust approach to governance, a well-defined architecture, and a plan for migration and consolidation.

Telco-targeted AI

Almost three-quarters (74%) of CSPs are exploring what it takes to build their own GenAI platform to reap the rewards of everything from simplified network operations to streamlined interactions with customers, according to the Analysys Mason research. However, this can be slow and costly. This is why 42% are relying on professional services providers to implement telecom-specific GenAI-driven use cases, which can help them strike a balance between innovation, responsibility, and cost-efficiency.

Leveraging GenAI copilots trained on telco data sets and taxonomy that use Retrieval Augmented Generation (RAG) and other testing and tuning methodologies, will drive the first phase of accuracy while heavily focusing on cost optimization as they scale up. These AI-driven assistants can help CSPs augment human capabilities and creativity with AI power. In practice, this will be used to create proposals or optimize pricing, generate insights based on what is known about a customer to provide upsell opportunities, offer a new subscription that might improve their experience, or identify network issues. In a contact center, telco staff could use these insights to improve customer experience and proactively manage customer interactions. Customers themselves can also be empowered with GenAI embedded into the billing process, which will reduce billing queries and misunderstandings. In fact, we already see early implementations of the technology making meaningful improvements in cost, speed, and even experience.

The best part is that we’ve only just scratched the surface of GenAI’s capabilities. We’ll be watching Meta, OpenAI, Microsoft, NVIDIA, and others with anticipation as they – and we at Amdocs – work to define what comes next in GenAI’s evolution, impact, and productionization.

Image by Gerd Altmann at Pixabay.