Balancing fantasy with reality in Telco AI deployment

Generative AI (GenAI) is rapidly becoming a cornerstone in telecom developers' toolkits, optimizing network infrastructures, enhancing customer experience and spurring innovation. The estimated size of the global “GenAI in telecom” market reached $150 million in 2022 and is projected to reach nearly half a billion by 2032, with an annual growth rate of approximately 41%.

Yet, while telecoms enthusiastically adopt GenAI, it comes with the significant risk of technical debt. In a few short years, GenAI has evolved from a nascent technology to a commercialized and democratized tool, transforming operations and processes for many telcos. While GenAI's promises are intoxicating, telcos now find themselves in a position where the “fantasy” of GenAI must be balanced with reality. 

Pressing full steam ahead with GenAI without considering the operational and technical resources needed to sustain it reliably may result in a situation for operators where their reach exceeds their grasp. 

Dreaming big with GenAI

GenAI is transforming the telecom sector and reshaping how telecom companies think about customer interactions and service delivery. The technology's allure lies in its ability to rapidly prototype and deploy new solutions, turning what used to be months of development into days or hours. As telecom operators explore GenAI's potential, they’re finding that it can dramatically enhance customer experiences by offering more personalized and responsive services, increasing loyalty and retention while creating new revenue opportunities. 

Initial success stories have been compelling and triggered a wave of enthusiasm across the industry as players envision what their operations could achieve with such a powerful tool. One telco in Europe recently increased conversion rates for its marketing campaigns by 40%, using GenAI to personalize content. It took this operator just two weeks to create and deploy the model.  

As McKinsey notes, one of the most impressive aspects of this GenAI telco wave is how rapidly new models and services are deployed: “For an industry with a mixed track record for capitalizing on new technologies and legacy systems that slow innovation, these early results and deployment times illustrate the potentially transformative power of GenAI.”

However, the rapid adoption and integration of GenAI tools also expose telecom operations to significant risks, primarily technical debt. In the rush to implement these advanced technologies, many companies overlook the long-term maintenance these systems require. Each new update or addition increases complexity, making future changes more cumbersome and error-prone. 

The sparkle of GenAI's capabilities often obscures the reality of its demands, including the need for continuous updates and monitoring and the potential for increased operational costs over time. Balancing these aspects becomes crucial as telcos venture further into this innovative yet challenging terrain.

The hidden costs of AI-driven development

The concept of “technical debt” will be familiar to any software development team, but GenAI introduces new layers to this age-old challenge. While GenAI can accelerate development timelines, it often does so at the cost of increasing the complexity and maintenance requirements of the resulting systems. As telco developers adopt these tools, they find themselves navigating a landscape where AI-generated code can be opaque and not fully aligned with existing systems. 

This misalignment may lead to longer debugging and integration times, paradoxically increasing the time and resources needed for future modifications and updates. The essence of managing GenAI's impact lies not just in adopting the technology but in integrating it with a clear strategy for long-term system maintainability. While quick development gains might seem enticing, they can lead telecom operators to underestimate the ongoing support these AI-generated systems require. 

To avoid creating isolated silos of technology that can degrade overall system performance, each new function or feature added via GenAI needs to fit seamlessly within the broader architecture. This scenario often results in a hidden operational toll, where teams spend more time fixing issues and adapting the GenAI outputs to work harmoniously with traditional coding environments. Addressing these challenges early and continuously reassessing the integration of GenAI within telecom infrastructures is vital for companies that wish to truly benefit from the efficiencies promised by the technology. Documentation is also a critical component of the software development process, which is often overlooked when rapidly generated GenAI code is used. 

Finding the GenAI sweet spot

For telco leaders, GenAI's practical applications range from enhancing network efficiencies to refining customer engagement strategies. These leaders pinpoint areas where AI can automate routine processes and free up valuable human resources for more complex, value-adding tasks. For example, GenAI can streamline the deployment of network resources, optimize route planning for data traffic and predict equipment failures before they cause service disruptions. These applications not only boost operational efficiency but also significantly reduce downtime.

The humble business support system (BSS), which has become a key focus for mobile virtual network operators (MVNOs), is also a prime candidate for GenAI-based innovation. Operators rely on BSS platforms for everything from customer support and service interactions to billing and the deployment of new features. While traditional network operators tend to have cumbersome, legacy BSS systems that are slow to innovate, MVNOs are typically smaller and more agile and are looking to adopt their own BSS solutions to capitalize on the GenAI boom. 

GenAI can automate billing processes, personalize marketing efforts based on consumer behavior and predict and resolve billing disputes before they reach the customer. Imagine a service where mobile tariffs can be automatically downgraded or upgraded based on the needs of each subscriber — and where operators can predict when a subscriber might be likely to leave their network and offer fresh, personalized service bundles to retain them. This is what GenAI brings to the table. 

However, the risk of disruption and the creation of technical debt will be significant unless operators handle these integrations with precision. Effective use of GenAI in BSS requires a careful approach where AI solutions are incrementally tested and validated. This will ensure they meet the high standards of reliability and accuracy so they can become a benefit rather than a hindrance. Suppose personalization attempts go awry, or the overall system isn’t “in sync” due to siloed data or duplicated records. In that case, any attempt to engage or retain customers might fail or make things worse. 

Navigating the hype

Amid widespread enthusiasm for GenAI's potential to transform software development, telco leaders must approach its integration with a meticulous strategy for validation and risk assessment. It's crucial to critically evaluate the perceived benefits of GenAI's rapid deployment and high functionality. To avoid pitfalls, leaders should insist on transparency from vendors regarding the origin and capabilities of the AI tools they plan to employ and demand comprehensive demonstrations and pilot tests. 

Deployment references are also critical to ensure the software is more than "magical" — but that it's also proven and reliable. These steps will ensure that the GenAI solutions fit seamlessly within existing infrastructures and are stable, scalable and secure. By setting rigorous standards for adoption and detailed testing protocols, telecom operators can leverage GenAI's transformative potential while maintaining operational integrity and their customers' loyalty and trust.

Matt Halligan is the Chief Technology Officer at Optiva. He has over 30 years of experience in the telecoms industry. Prior to joining Optiva, he served as the CTO of Openwave Mobility.

Joy King leads Optiva’s go to market strategy. She has three decades of software industry experience. Previously, she led Product & GTM strategy, including Product Management & Marketing, for Vertica.