Narrow AI heals telco pains for now but a widespread harmonized approach needed soon

London, United Kingdom - 23 Oct 2018

Artificial Intelligence (AI) for intelligent user interfaces, sales and marketing remains the highest priority among mobile service providers, found ABI Research , a market-foresight firm providing strategic guidance on the most compelling transformative technologies. AI is being positioned as a key pillar in telco digitization, in parallel with customer-centric network modernization projects, virtualization, and cloud-native infrastructure.

ABI Research’s recent study of how AI is being implemented in the telecoms industry has concluded that the ownership and technology architecture of AI solutions is largely fragmented between different service provider departments. Customer services, network management, service assurance, and cybersecurity are popular areas for AI in telecoms, but most of these come with their own solutions partner, framework, and roadmaps. “At present, AI in telecoms has a narrow application in numerous use cases, particularly customer management and network acceleration. A holistic and general-purpose approach to AI may be required to enable Mobile Service Providers (MSPs) to become AI influencers rather than followers. This is fundamentally different from the siloed AI operations that predominate today,” said Don Alusha, Senior Analyst at ABI Research.

Several MSPs are aggressively adopting AI in many business areas. For example, AT&T and Telefónica are taking an all-encompassing approach to developing AI across many platforms, with the Acumos and LUCA projects, respectively. To become a powerhouse in AI, these two operators notwithstanding, telcos will have to overcome their divisions, start to digitalize their specialized and hardwired assets, and become strategic with new ethical frameworks and top-down global governance systems and strategies. This transition needs to start from the very top of the MSP organization for board or executive management buy-in.

The industry must coalesce around a unified vision for an AI ecosystem with clear commercial relationships and developer frameworks that span cloud (public or private) providers, vendors’ own hosting environments, and telco data centers and platforms. Specifically, numerous open source AI frameworks and platforms are already being used by telcos such as NTT DOCOMO and others, to build up their AI expertise internally. Open source AI frameworks and generic platforms may well become the norm in service providers’ operations, in turn calling for vendors to align their business models to account for this. The sheer multiplicity of open source platforms, however, may impede far-reaching AI standardization if no caution is not exercised.

“Work remains to be done on two fronts: one, standardize the data format to achieve data harmonization; and two, institute AI standards with particular focus in the network domain, where little attention has been paid thus far,” concluded Alusha. “Pushing for an AI global strategy beyond what is currently in place today is likely to have a stronger impact over time, particularly in the context of developing AI platforms with an eye to enabling massive scalability and agility of features.”

These findings are from ABI Research’s Artificial Intelligence and Machine Learning Use Cases in Telecoms report. This report is part of the company’s Telco Digitization research service, which includes research, data, and Executive Foresights .