The latest trend for AI development in telecoms: AI platforms that focus on generic use rather than individual use cases, but widespread commercial deployment is still years away, says a new report from ABI Research.
The use of AI and machine learning in the telecom domain is now an established practice – chatbots, customer services, and network management are areas where AI is part of everyday operations. But despite the widespread use of AI and machine learning across many separated telecoms domains, AI platforms are now the top concern for leading mobile service providers globally, including AT&T, Deutsche Telecom, Telefonica, and Vodafone.
“The promise of these platforms is great, but there is still heavy development to be done and many mobile service providers are still not sure where to place their bets,” said Dimitris Mavrakis, Research Director at ABI Research. “Initiatives like AT&T’s open sourced Project Acumos illustrate that open source will be the most suitable development environment for these telecom AI platforms, potentially meaning trouble for established AI vendors, such as Amazon, Google, and IBM.”
The use of these generic AI platforms is a natural fit for telecoms use cases, where mobile service providers can utilize the vast amounts of data they collect to improve their operations, enhance customer experience, and even provide value to business partners. For example, BT has saved tens of millions of pounds by combining human input, data collected from its network, and its own developed AI algorithm, to be able to predict the implication of any business decision on its network and financial position, Mavrakis says.
“Most mobile service providers are currently assessing AI platforms and whether they should invest in a centralized approach or continue with the use case specific AI applications they currently operate,” said Mavrakis. “Given that AI will be a critical technology for the future of any business, the right decision will translate to a serious competitive advantage and financial benefits. Tier-1 mobile service providers should embrace open source and get involved in the development of AI, arguably the most important technology in the 5G era.”