Deploying generative AI in Telecoms

Telecoms.com periodically invites expert third parties to share their views on the industry’s most pressing issues. In this piece Yannick Martel, Capgemini’s offer lead for data & AI in the telecom industry, takes a look at the GenAI trend from a telecoms perspective.

Generative AI’s recent advancements have unlocked opportunities for all industries. While areas of interest will differ by sector and organization, all businesses should investigate what they can stand to gain from implementing generative AI in their operations. The telecommunications industry is no exception.

However, before determining where generative AI can deliver strategic and tangible value for telco operators, it’s crucial to understand the broader industry landscape and the current challenges.

Understanding the telecoms landscape

The telecoms industry is experiencing a period of stagnancy in revenues. Despite this, operators continue to prioritize innovation, and invest in network technologies, such as 5G and fiber. However, due to the high cost of those investments, shareholders are asking for visibility on value recovery, while in parallel, customers are seeking enhanced service quality and convenience. To reconcile these demands, and to deliver on efficiency without compromising service quality, communication service providers (CSPs) must leverage automation.

When growth subsides, investment decisions become more complex. They demand greater strategic nuance and a clear focus on delivering tangible value. Operators are prioritizing investments and innovation that reduce costs, increase revenue from existing services and/or create new revenue streams. Artificial intelligence, particularly generative AI, presents an opportunity for delivering benefits of automation beyond traditional rules-and-flow-based techniques. This technology can improve service quality and increase efficiency, resulting in that all important cost reduction.

Creating a generative AI strategy

Generative AI is not a solve-all solution for telecoms operators, but when strategically applied, it can deliver tangible value. The good news is that generative AI is already part of the boardroom conversation. In fact, 96% of telco executives view generative AI as an important topic of discussion, and 60% of telecoms leaders are strong advocates. Additionally, nearly half (47%) of telcos have started exploring the potential of generative AI initiatives.

So, where can telcos look to strategically implement generative AI to realize value within their operations?

Experiment with live prototyping: Despite concerns about maturity, and the risks associated with generative AI, many CSPs have already started experimenting. In reality, those who haven’t yet started assessing generative AI’s potential are falling behind. That’s not to say that telcos should be approaching this technology with reckless abandonment, but rather, that they should take small incremental steps and test along the way. For instance, they could investigate its use across low-risk applications that enhance employee capabilities before implementing client-facing automation. CSPs should be seizing opportunities to implement small, quick changes to gain efficiencies and seek to familiarize their organization on such advancements as the technology matures. Indeed, in the first initiatives, it does not matter so much where generative AI is applied, as long as it helps generate some visible value while providing the opportunity for learning across the organization.  I know of one large US telco that is taking this approach and is investing in internal pilots, then scaling them across their organization to prepare for delivering innovative consumer-facing services at scale.

Efficiency in software engineering: CSPs often have large legacy software applications requiring large maintenance and modernization efforts. These applications consist of millions of lines of code, making software development inefficient and time intensive. This issue is compounded when talent leaves, and the legacy knowledge leaves with them. At the same time, CSPs need these software applications to evolve in line with changing business conditions. In this scenario generative AI functions act as an additional team member, assisting software engineers in understanding legacy software, facilitating conversion, testing, and ultimately reducing cost and risk associated with software maintenance. The transformation of software engineering through generative AI has quickly become an objective for many of our CIO clients.

Efficiency in business operations: Similar to many organizations, telcos develop a substantial amount of content. Whether it’s product briefs, marketing campaigns, job descriptions, generative AI can act as a technical writer efficiently creating content, ready for human review. For the employees that would previously have had to agonize over tone, length, and shape of messaging, generative AI can free up their time, allowing them to focus their attention on the intent and the overall message. For instance, we are currently scaling an application to generate all kinds of external facing documents with the appropriate tone, bringing efficiencies to one of our large clients.

Customer satisfaction: Generative AI is not yet able to interact on its own with customers without adequate safeguards. However, it can already improve customer service by automating simple functions, such as helping orient customers and directing them to a specialized group. One area where generative AI is showing considerable promise both currently and further down the line is with augmenting customer service agents. Right now, it can be used to create summaries of conversations, enabling customer service employees to have a greater understanding and awareness of a specific customer’s situation. In the future, customer service departments will be able to put a human avatar in front of the customer that possesses a comprehensive understanding of each customer’s history. It will be integrated with all customer data sources, and be able to enter and confirm transactions, all with the necessary safeguards and human oversight. Many of our customers are experimenting within what we term “safe ground” such as call summarization, while preparing customer-facing pilots. Of course, here it's important that the adequate preparation is taken, as true transformation will take considerable amounts of time and effort.

Beyond highly visible areas, such as developer copilots, and truly cognitive chatbots and voice-bots, generative AI will transform many processes and provide efficiency across organizations. . A lot of efficiency gains are easy, but those who win the race and gain a competitive advantage will be those who will industrialize and systematize tactical optimizations while driving large-scale transformations with a clear direction and purpose. To prepare for those critical transformations, CSPs should start building today, raising awareness at the executive-level and within teams. Only then will they be able to prepare for a future of enhanced automation, personalized customer interactions, and improved employee support.

Yannick is global offer lead for Data & AI in the Telecom industry at Capgemini. He has 30 years of experience working in the Telecom and Banking industries, helping organizations actually implementing their digital and data transformation with AI and Generative AI, improving operational efficiency, reducing their footprint and better serving their customers.