As dawn rises on the age of telco artificial intelligence (AI), Silverlinings reached out to Roman Ferrando, principal AI analyst at Appledore and a man who has a lot of experience in AI and machine learning, to talk about the growth of telco AI and what it means for the industry segment.
Ferrando, spoke with us on the heels of the launch of the Global Telco AI Alliance formed by SK Telcom, Deutsche Telekom, e& and Singtel, which he noted on LinkedIn that "the wheels of progress are turning in the right direction” for AI in the telco sphere.
While AI may seem like it could be a challenge for the slow-to-change telco industry, Ferrando told us instead, “the big problem we have is not really technical, it is the lack of realistic expectations...from the decision making level."
Lay of the land
With telco AI, Ferrando noted that the expectation is “this is going to save the world...we’re going to make millions...this is going to be fantastic.”
“I think there is a lack of leadership...when it comes to connecting the [dots] between the AI potential and the business reality,” he explained.
In the wider technology world, there are already plenty of examples of AI working in the wild. From facial recognition on your cellphone to AI customer service bots answering your anguished wails on corporate websites with a breezy “how can I help you today?”
Ferrando noted that there are a lot of AI apps and services concerned with predicting customer churn on the telco network already in place. Vendors supplying machine learning and AI churn reduction services include Akkio and HEAVY.AI among others.
To further develop the telco AI ecosystem though, “we need standard interfaces'' from vendors that customers can integrate with other tools, Ferrando said. That’s happening, but it's happening slowly, he added.
The situation is much like what happened with the cloud. Even now, Ferrando doesn’t think that many big operators will convert to the public cloud in the midst of this AI revolution. He notes that, technically, microservices are key in the telco world today, even if the hyperscalers don’t get a look in.
Of course, the hyperscalers do want in on the AI world.
Operationally, Ferrando thinks that it may take quite a bit longer before network operations are free of human hands and have moved to a telco AI environment for most operators. This is largely for two reasons, he said. One is that the network is a very, very complex environment and every single network is different. So, that’s not too helpful for a machine wanting to learn how to run operations.
Second, operators are “very conservative” and are “not happy with the idea that a machine is changing the network,” he said. “I think it will take some time before we see a significant reduction of...human intervention in the operation process.”
Until then, it sounds like this is the name of the game: "I'm sorry Dave. I'm afraid I can't do that."
Want to learn more about cloud-native 5G networks? Check out our Cloud-Native 5G Summit on demand now.