How cellular operators can use generative AI

Generative Artificial Intelligence (or GenAI) is the use of machine learning models to produce various types of content (including text, images, programming code, audio, video, or other forms of media) in response to prompts and by applying rules based on the data that models have been trained on. It is a broad term to describe an AI system whose primary function is to generate content and similar outputs. This is different from other types of AI, like discriminative AI which focuses on tasks such as classifying or identifying content.

This isn’t a new concept, but it has garnered a lot of attention in the last year or so because of the mass adoption of models like ChatGPT, DALL-E, and others. The training data sets, and the number of parameters used in the training of such models has also increased enormously in the last couple of years. Moreover, until now companies have used AI to predict or identify patterns, but generative AI takes this concept a step further by generating outputs in formats that are easier for users to interact with, including text, video, and images.

Emerging use cases are diverse, but share a common thread of content generation

Generative AI can be deployed to support several enterprise functions, with the most prominent being marketing and sales, customer service operations, IT processes (such as application development), and product R&D. Generative AI also has potential in the context of HR, legal, supply chain, risk, and compliance areas and together these contexts are next tier of priority use cases for early adopters.

Figure 1, below, includes a summary of early-adopting functional areas and the use cases that Transforma Insights has identified as typically being leveraged in those functional areas.

Figure 1: Early adopting functional areas and associated generative AI use cases [Source: Transforma Insights, 2024]

The use cases listed above are currently being explored by companies across various industries at different rates. In pharmaceuticals, for example, generative AI has a strong potential to transform drug discovery processes by accelerating the development of novel molecules, disease target identification, and prediction of clinical trial outcomes. An initial focus in the IT sector has been to use generative AI to support the migration of programming code from legacy languages and environments to more modern implementations. Also in the IT sector, generative AI can be potentially used to support the semiconductor chip design process. Consumer Packaged Goods companies, meanwhile, have typically focussed on deploying generative AI to enhance marketing, and multiple industries have experimented with generative AI to support customer service.

Use of Generative AI in the telecommunications sector

The mobile telecommunications sector can benefit significantly from generative AI, and in many contexts. Those that have been explored so far include multiple use cases in customer service contexts, and also in sales and marketing, IT systems administration and knowledge management. Some of the leading use cases are discussed in more detail below.

The first and most obvious way in which cellular operators can leverage generative AI is for supporting customer service, either using on-line or voice-based chatbots. For many years, cellular operators have sought ways to automate customer service and so reduce costs and the advent of generative AI will allow for significantly more sophisticated solutions.

Generative AI can also play a role in customer service centres, supporting customer service agents. There are two main ways in which such support can be provided. It can take the form of prompted responses either to real-time customer enquiries or to written correspondence, which customer service agents can review and tailor before responding to a customer enquiry. Alternatively, generative AI can support the navigation of in-house knowledge to help customer service agents to respond to more unusual situations quickly and efficiently.

Still within the domain of customer service, generative AI can be used to help monitor customer service performance. Today customer service agents and customers both are often asked to review engagements after completion, providing ratings according to whether the issue motivating the customer enquiry was resolved efficiently, or not. Generative AI techniques can be used to automate this process, providing unbiased and consistent results across all enquiries received (compared to the typically biased, inconsistent, and partial feedback provided by human participants).

There is significant potential to use generative AI to support general sales and marketing of telecommunications services, including the generation of new messages targeted at specific segments and also identification of new segments.

Some leading cellular operators have used generative AI to help develop initial responses to requests for proposal (RFPs) received from enterprise clients. In this context, generative AI would typically be used to create an initial draft response, which bid managers or sales personnel can further refine before sending to a potential customer. This approach can save significant human resource by helping to automate many of the more basic and mundane aspects of responding to an RFP.

In technical support contexts, generative AI can be used to help migrate legacy software applications to new environments and also to test new software applications and support data translation between environments.

In a wider business administration context, generative AI can be used to support human resources and recruitment processes and even to help analyse legal documents and contracts.

In the widest sense, generative AI can be used to support enhanced knowledge management and training within a cellular operator. Typically, generative AI techniques can help newly hired staff relatively quickly to reach levels of efficiency more usually associated with knowledgeable staff members who have been in place for years.

These are still early days

It’s only a little over a year since generative AI hit the headlines with the release of ChatGPT based on the GPT-3 Large Language Model and a rumoured USD10billion investment from Microsoft. Since then, the technology has grabbed management attention and diffused across industries and functions at breakneck speed. We’re still very much at an early stage of adoption though and it’s clear that in future the technology will be leveraged far more widely than it is already today.

Clearly, generative AI as a technology has huge potential to reduce costs and make key processes within adopting companies more efficient. But, for now, generative AI deployments typically focus on significantly increasing the efficiency of things that could already be done, rather than introducing new and truly transformative propositions.