AI in Telecoms - Bounded Rationality or Wider Ecosystem Game Changer?

Subscribe To Download This Insight

4Q 2018 | IN-5293

ABI Research’s recent study of how Artificial Intelligence (AI) is being implemented in the telecoms industry underscores the fact that the ownership and technology architecture of the AI ecosystem is largely fragmented (PT-2184). With almost half of all commercial activities we tracked falling under customer management, much of the action in AI is around intelligent user interfaces, sales, and marketing. These findings highlight how AI is unfolding in the market; namely, on a per use case, per domain basis—a natural development given the monolithic environments in telecoms. The benefits that come from a narrow AI adoption notwithstanding, a harmonized AI platform may be required to enable a wide-ranging and scalable business case (AN-2557).

Registered users can unlock up to five pieces of premium content each month.

Log in or register to unlock this Insight.

 

Artificial Intelligence in Telecoms

NEWS


ABI Research’s recent study of how Artificial Intelligence (AI) is being implemented in the telecoms industry underscores the fact that the ownership and technology architecture of the AI ecosystem is largely fragmented (PT-2184). With almost half of all commercial activities we tracked falling under customer management, much of the action in AI is around intelligent user interfaces, sales, and marketing. These findings highlight how AI is unfolding in the market; namely, on a per use case, per domain basis—a natural development given the monolithic environments in telecoms. The benefits that come from a narrow AI adoption notwithstanding, a harmonized AI platform may be required to enable a wide-ranging and scalable business case (AN-2557).

Narrow AI may not be conducive to the creation of a new ecosystem—meaning an interconnected environment with uniform business connections and centralized sponsorship underpinned by standardized data formats and AI telco standards. AI opens new opportunities from operational efficiency and top-line growth perspectives but on the condition that the industry steps out of the “bounded rationality” phenomenon highlighted in our research.

  Machine Learning Use Cases  

Dual Artificial Intelligence Strategy 

IMPACT


At present, implementation of AI is characterized by an evolutionary singular strategy—one tweaked and fine-tuned in line with specific use cases typically associated with their own solutions partners, dedicated departments, or road maps. For example, customer services reduce human involvement by using AI in chatbots. For Mobile Service Providers (MSPs), it would be unwise to design an AI strategy based on current pain points, and, at the same time, it would be equally ill-judged to plan a strategy with a 10-year vision. Striking the right balance between the two is becoming increasingly important, but this remains a huge challenge since generic AI platforms targeting the telecoms’ vertical remain immature or even nonexistent. AT&T’s Acumos project—now open source in the Linux Foundation—is perhaps the only platform which can be used in several operator business domains, and it requires significant development, buy-in from several AI practitioners, and vendor involvement to become common practice.

Technology providers, on the other hand, must enable an adoption road map wherein specific actions taken now would yield short-term desired outcomes that also have an effect that radiates for years to come toward a wider AI stratagem. Ericsson, for example, is one vendor that has announced a road map to set up a global AI accelerator in the United States, Sweden, and India, with 300 AI experts to drive AI adoption in telecoms. The Swedish vendor also uses narrow AI use cases in several different product lines, including Operation Support Systems (OSS) and cellular domains. Huawei has also announced Noah’s Ark—a full stack AI technology developed by its Research and Development (R&D) center—that is available to its clients via Huawei’s Infrastructure as a Service (IaaS) and Platform as a Service (PaaS) services. Huawei’s AI and associated chipsets are powered by Da Vinci, an architecture that divides the computational processing into several domains capable of performing cloud-based training and inference.

Currently, MSPs and their partners are seeking to understand where it makes sense to use AI, how to understand the business value, and how to manage it on a per domain basis—bounded rationality in a nutshell. A digitally empowered telco, however, mandates varying degrees of integration, interoperability, data sharing, and openness—a feat that may run counter to the narrow, use-case-specific AI diffusion of today. That said, harmonization of a narrow approach and one that promotes a holistic strategy is certainly not easy and something beyond the power of any MSP—even Tier 1 heavyweights. A case in point is Amazon, Facebook, and Google, who have all designed far-reaching AI strategies. However, they did not do this by taking specific narrow domains and somehow transforming them. The industry at large must overcome divisions on two strands to fully capitalize on the benefits of AI: (1) the digital realm in the form of OSS/Business Support System (BSS) assets, networks, business modeling algorithms, and anything in between; and (2) the strategic realm in the form of new ethical frameworks, governance systems, and the necessary human capital (the most valued asset of all) that is fundamental to productivity.

Artificial Intelligence and Human Capital

RECOMMENDATIONS


The advent of AI in telecoms poses a fundamental question: how can adopters, be it MSPs or vendors, change the structure of current systems and processes to produce more of what is desirable and less of that which is undesirable? An effective strategy would be to look for leverage points—places in the “old” operations where a small change can lead to a large shift in behavior. In addition, the most profound aspect of AI and business intelligence at a broad level is, ultimately, that an organization’s only sustainable competitive advantage lies in its workforce. More specifically, what its employees know and how they apply AI input to business requirements is a key consideration that warrants a “continue to grow over time” mind-set. Solution providers should foster this mind-set by introducing reskilling programs, embedded in a culture of lifelong learning from the outset. Naturally, this change will not happen overnight, even in the most progressive of organizations. But it is vital that vendors plant the seeds of AI today to create value for their clients tomorrow.

MSPs, vendors, and technology multinationals must tackle multiple AI facets—one of which is organizational. Workforce mind-set, which is the common set of shared beliefs and values underpinning employee behavior, is one of the key defining attributes of a company and the hardest to influence because people are resistant to change. Therefore, MSPs and vendors must embrace a long-lasting organizational change if a top-down AI strategy with a global organizational reach is to succeed. One way of addressing this is by easing the workforce into a hybrid mode of operation that incorporates AI expertise in stages, starting with (small) activities that are best suited for AI. Another way of calming uneasiness is by introducing reskilling programs. AT&T Workforce 2020 is an example of a far-reaching retraining program aimed to engender a culture of perpetual learning. Further, market players that eventually aim for a global AI strategy should seek adequate sponsorship, unwavering commitment, and direct governance from one or more C-level executives.

Complexity may well be the hallmark of the industry for the next decade or two. Virtualization, edge computing, 5G, and the Internet of Things (IoT) will create and will require networks that cannot be managed with today’s processes. The demand therefore will be for levels of human capital that embrace radical technologies such as AI to manage that complexity and drive sustained growth. MSPs and vendors who get this right, the market entities that understand how to mobilize and apply the human capital, and the partners that produce or facilitate it will be the big winners. When assessing AI as a technology, decision makers should consider the implications of embracing it but should not believe that AI by itself is the sole answer to commercial success. For vendors and MSPs to succeed, they must create the right culture and environment to integrate AI into their DNA, thereby creating a workable “human plus AI” model.

Services

Companies Mentioned