- Telcos want a bigger role in the AI value chain than just carrying traffic, and Personal AI thinks customer “memory” could be the answer
- The company believes operators could turn subscriber context into a secure intelligence layer that powers more personalized AI experiences
- But analysts warn consumers may not pay for something they think is already happening
Telecom operators are sitting with a do or die question: how do we make ourselves an indispensable part of the AI value chain? Sure, they sell the connectivity, but they’re unlikely to reap more ARPU just because AI traffic is flowing over their networks. That is, unless they can think of a way to add value for both customers and the apps they use.
Suman Kanuganti, founder of Personal AI, thinks he has the answer. Telcos, he suggested, can become the intelligence layer that feeds personalized AI experiences.
When Kanuganti talks about this concept, he uses the word memory – not in the DRAM or HBM sense, but as in human memory. People mentally store information about their likes, dislikes, movements and habits, and those of their friends and family. So too can telcos create a profile for each of their customers that can become a valuable asset in the AI era.
“Models are going to become a commodity,” he argued. “And if you think about not tying into any single ecosystem, then knowing who you are should reside with the telco carriers.” This kind of arrangement would create a highly secure, private and portable asset that “is necessary for the consumers to create value for themselves.”
Put another way, Kanuganti is suggesting telcos put customers in control of context that they can then share with apps and models to get a better experience. Rather than just supplying connectivity, they’d also be supplying intelligence.
Asked about the data privacy and security angle, Kanuganti noted telcos are already trusted stewards of subscriber information and subject to laws governing how data can and can't be used and shared. Plus, his vision calls for customers to hold the ultimate control over which apps and services can tap their personal "memory" profile.
Could an AI intelligence layer be a money maker?
The obvious question that follows is ‘how do telcos monetize this?’ Kanuganti suggested they could charge 15-20% more in ARPU for access to intelligence with their core subscription. And they could also charge apps on the other end for access to the data or expose certain intelligence-related APIs to developers for a fee.
Fierce ran the whole idea by Mobile Experts Founder and Chief Analyst Joe Madden. “This idea sounds like ‘cookies on steroids,’” he said.
Though Madden conceded that context would likely result in better AI support for each customer and thus provide some value, he said he has “serious doubts” about whether large numbers of customers would be willing to pay for something like this. After all, they’re already getting targeted advertising based on the profiles social media companies and tech giants like Google have built for them.
“Consumers are likely to expect this service to be free, as with advertising, and would be confused about paying for something that they perceive is already happening to them,” Madden said.
Madden also brought up the trust angle: Today, most people believe generative AI models aren’t “tainted” with personal data advertisers have pushed into the system. “But if the telco starts to collect money from the AI players, and the search results change as a result, I wonder whether people will start to distrust the AI results, thinking that they have become biased,” he said.
Where do things go from here?
Kanuganti said discussions with “C-level” executives at telcos around the globe – “big ones” – about this kind of offering are already underway.
While the idea is hitting, he said it’s also challenging because moving forward requires education and derisking. The next milestone, he added, will be one telco going live.
“That is the milestone, because that derisks the move for all other telcos,” Kanuganti said. “A lot of telcos are fast followers and not necessarily the first movers. So, my job is to figure out the first movers who are working with us to take a calculated risk and go in with a very methodical approach because we all understand that this isn’t another product bundling…this is fundamentally introducing a brand new primitive on the network.”