Gen AI Use Cases in Telecoms: Where is the Value?

Download Listen

Exploring Gen AI use cases has come into focus for many telecom operators amidst the hype around the technology. STL Partners has published extensive research on the topic, including our recent report ‘Gen AI: where should telcos start?’. Here is a quick summary of our findings.

Gen AI has gone a long way in the last 12 months

ChatGPT recently had its anniversary on the market: it was launched on 30 November 2022. A lot has happened since then.

The term ‘generative AI’, or in short ‘Gen AI’, gained huge popularity, and the awareness of AI grew significantly, so much that the word ‘hallucinate’ became the word of the year for the Cambridge Dictionary. Importantly, the adoption of ChatGPT and other similar tools that followed shot up: in November 2023, ChatGPT was believed to have been used by over 180 million people, with over 100 million active weekly users.

This democratisation of creative AI applications meant that many employees became enthusiastic about exploring how to implement these tools in their work lives. This created a pull for companies to figure out what the use cases of Gen AI for their business might be, and even more generally, what their Gen AI strategies are.

Key Gen AI use cases in telecom

While the most common use of Gen AI in its first year has been creating text and images, its capabilities go a long way beyond examples like writing a few paragraphs on a topic or proposing a list of interesting titles for a new marketing campaign.

STL has identified almost 100 use cases for gen AI in telcos, clustered into seven categories:

1. Content creation

Gen AI can be used to analyse, summarise, translate or restyle existing text, or to create new text, images and videos, improving efficiencies in several roles from sales and marketing to human resources and legal departments.

2. Human-machine interactions

Gen AI can help understand human input via natural language and generate human-like responses fit for the context. Improved chatbots are an example.

3. Human-human interactions

A good example here is using Gen AI in business meetings to create automatic summaries and logs.

4. Knowledge management

Gen AI can analyse data sets, fill in gaps, and find trends in large data sets- all valuable use cases across the organisation.

5. Process Improvements

There is a range of use cases here, aimed at improving operations. Two interesting examples are code creation and process documentation.

6. Data management

Gen AI can play a role in managing and analysing large datasets in telecoms, for example by augmenting incomplete data sets and improving telco knowledge bases and catalogues.

7. Intelligence improvements

Gen AI can help detect algorithm anomalies and improve predictions.

While the underlying technology is still in its infancy, there is limited differentiation among use cases when it comes to maturity. But when it comes to how valuable these cases are for telcos, the variance is significant.

Potential use cases of Gen AI in telecoms: availability and usefulness

 

Source: STL Partners

Use cases like content creation: image and text, are valuable for marketing and sales. Tools are already available, and people have embraced this use case at a personal level. Of course, what Gen AI creates should be taken as a draft, which then needs to be reviewed and improved by people. But while this can be useful, this use case is not the most valuable to telcos.

An example of a much more valuable use case for telcos is the use of large language models (LLMs) to write short pieces of code. Traditionally, telcos have found it difficult to compete with hyperscalers in attracting software developer talent. Now Gen AI will make it easier for telcos to fill in some gaps and let the available talent focus on developments that LLMs cannot deliver. Over time, more employees will be able to use Gen AI to help them with low-code activities. This use case is in its infancy, but it has the potential to deliver a lot of value to telcos.

How telecom operators can embrace Gen AI

Potential use cases are important to know and classify, but how can telcos make a proper step into utilising Gen AI beyond trials?

Indeed, telcos are not new to AI. Applications of AI around improving energy efficiency and network performance are already in deployment around the world. But these are very technical deployments. Gen AI is a whole new beast, providing a chance to use AI in less technical operations, by people in less technical roles within a company.

To utilise the potential of Gen AI, telcos need to develop knowledge and acceptance of Gen AI across the organisation. Fortunately, telcos can take advantage of the fact that people are excited and curious about Gen AI, so not only there will be little or no resistance or fear of using AI, but there will be cohorts within the workforce who will be very willing to develop Gen AI use cases. But companies need to think of a variety of deployment scenarios and decide which LLMs to use and whether they need to be proprietary or public.

One thing to bear in mind is the cost of using Gen AI. The human capital that needs to be deployed in developing use cases and processes around them is significant. OpenAI, the creator of ChatGPT, engaged thousands of people to label data. Over time, Gen AI will be able to save time and cost, and create new revenue streams, but first telcos have to invest in it, especially if they need to create and train proprietary models.

Telcos should stay open-minded about Gen AI

While Gen AI is still making its first baby steps, telcos should stay well-informed about its potential and make decisions on where and how to invest early, before their competitors beat them to it and build an advantage.

STL’s research covers Gen AI, analytics and automation from different angles and helps telcos build an understanding of how to best harness the transformative potential of Gen AI, from enhancing customer interaction to supporting innovation.