-
Telcos are lagging in AI adoption in part because they’re risk averse
-
AI in the core largely remains in the exploration stage
-
To really find AI success, operators need to build AIOps teams
Telcos may be the tortoise to the enterprise hare when it comes to adopting artificial intelligence (AI), thanks in part to their risk averse nature. They’ve only just started to scratch the surface with adoption of generative AI tools that can improve the customer experience and network operations. So, how long until AI crawls into the core? Well, let’s just say don’t hold your breath.
“AI for telecommunications is I think still emerging,” Alex Quach, VP and GM of Intel’s Wireline and Core Networks Division, told Silverlinings. “For AI in the core networks, today most of the use cases that have been explored have to do with network health. So, anomaly detection, to some extent power management as well…There’s also some thinking around AI for network slicing, especially for the 5G standalone core to be able to offer the right SLAs for customers.”
Quach said a lot of those kinds of use cases don’t involve training as much as inference, meaning they could be handled by hardware like Intel’s Xeon processors (though, of course he would say that).
All that said, Quach noted much of what’s being done with AI in the core today is still just exploration – not implementation.
“I think we’re still in fairly early days of AI adoption in the core network,” he said.
Microsoft VP Telco and Media Silvia Candiani similarly indicated telco AI is still a nascent space. That’s in part because the “network is much more manual to a certain extent versus the IT world.”
“What we saw in automating data centers in the IT world is just now happening in the network side,” she explained. “Of course, networks are very important, they are critical infrastructure, so we need to make sure that we have security and the network reliability at the center. But there are a number of tools that can help simplify and automate the operation of the network itself. So, I think that’s going to be very common.”
A recent McKinsey survey showed that despite telcos' hesitation, AI is already making an impact. Approximately 40% of the 130 operators who responded to the survey indicated they have achieve 1-5% cost savings in their network thanks to AI. Another 15% said they achieved 6-20% cost savings in the same domain. Cost savings on the customer service front were even more advanced.
No brains without muscle
As we’ve mentioned before, the kind of AI technology needed to make telecom networks fully autonomous is still under development. That means that people still need to be in the loop.
John English, director of Service Provider Marketing and Business Development at NetScout, told Silverlinings at an MWC networking party this week that no operator will be truly successful with AI unless they have an AIOps team.
Sure, he said, everyone has DevOps and DevSecOps teams, but AIOps teams remain scarce. English's assertion was backed up by McKinsey's survey which found "only about one-third of telco leaders said they have a capability-building plan for employees on gen AI or are investing in change management efforts—two core building blocks for building a culture of innovation and the test-and-learn mindset" that is needed to succeed with AI at scale.
That's not great news considering there's the potential for telcos to reap nearly $100 million in incremental value from AI, according to McKinsey.
So put that on your to-do list, folks: build an expert AIOps team. Don’t say we didn’t warn you.
Editor Elizabeth Coyne contributed to this story
Follow our coverage of Mobile World Congress 2024 in our dedicated channel here.