Getting closer to the customer with GenAI

While much of the focus for traditional forms of AI and machine learning sits within network and network operations most of the excitement about generative AI is in its transformative potential in customer and market-facing functions.

TM Forum has identified seven families of use cases spanning the entire breadth of the CSP organization, two of which sit within what could be seen as a larger category that spans customer experience, sales and marketing. In its research with CSPs, TM Forum has identified this as the biggest overarching category where operators will focus much of their early investments over the next one to two years.

Source: TM Forum

Use cases for GenAI sit across the entire customer services and marketing organizations offering benefits to the call center organization, the teams developing new digital tools for customers, to sales and market and, most importantly to the customers themselves. Deployed sensibly and responsibly, Gen AI-enabled use cases can help deliver a better customer experience, more loyal customers, efficiency gains, and net new revenues.

Delivering an enhanced chatbot experience is a priority for many CSPs. Today’s chatbot services can fall short of what customer expect because of their limited capabilities. Using GenAI, CSPs can transform the chatbot experience by evolving from a traditional AI and rules-based system which provides a limited set of customer resolutions to one that provides answers to a much wider range of queries and requests. It can do this by training millions of customer interactions on a large language model (LLM). GenAI chatbots use unstructured rather than structured data to understand what customers want and how it can help.

The transition from rules-based to GenAI-based chatbots will be gradual and a function of how quickly CSPs can train and fine-tune LLMs and use other techniques such as retrieval augmented generation and prompt engineering to deliver more accurate results than are possible today. For the time being, CSPs must use these generic LLMs and train them using their own internal data. However, various CSP initiatives are underway to build LLMs using only CSP data and the knowledge that sits between the customer and the operator. These promise much higher levels of accuracy.

In the short-to-medium – and even when generating results that are not accurate enough to be relied upon to make decisions – LLMs and GenAI can be a time-saving and performance-enhancing tool for customer service agents. With faster and better access to information about customers – their demographics, their spending and usage habits and their previous interactions – customer service agents can resolve issues more quickly, build better customer relationships and take a more proactive role to giving customers what they want. This includes cross-selling, upselling and converting early renewals.

Moving from customer service functions to sales and marketing, these same customer insights can have a transformational impact in terms of how CSPs personalize communications with their customers. CSPs have long aspired to use customer data in the same way as digital-native B2C companies to make personalized recommendations based on previous purchases and interactions.

CSPs may possess a wealth of customer but this often sits in different islands across the CSP organization. Bringing this data together to create strong customer insights and then leveraging such insights for personalized marketing campaigns has been a challenge. GenAI can now be used to do some of this heavy lifting, bringing together both the structured data that sits in different IT systems and unstructured data derived from real and online conversations. Using GenAI in combination with digital twin technologies can deliver even greater value, enabling CSPs to predict outcomes and optimize processes. It can help CSPs to deliver more tailored marketing through more effective data consolidation and analysis that helps them to anticipate customer needs.

GenAI also gives account managers the ability to send personalized messages to customers and prospects. It can also help sales teams identify the best prospects and create scripts for conversations with them.

While 2023 was the year of aggressive experimentation with GenAI, 2024 and 2025 will see operators take the first use cases into production. In doing so CSPs have the opportunity to bring their customers closer and bridge the customer experience gap with firms that were born digitally native.