- Personal AI founder Suman Kanuganti predicted AI token plans are coming to a telco near you
- Consumers may be confused at first, but Kanuganti said the adoption path will be similar to what the industry saw with data plans
- Analysts, however, are skeptical metered token plans are the answer for AI monetization
Before everything went unlimited, the communications world was governed by metered plans.
People of a certain age will remember waiting for nights and weekends to make calls in order to conserve precious “daytime” minutes. Others might recall straining up against the limits of a text message allotment or running out of data at just the wrong moment.
Now, it seems, AI tokens could be the next communications medium telcos sell by the bundle.
Suman Kanuganti, founder of Personal AI, told Fierce that telcos are uniquely positioned to monetize AI by leveraging network data and offering users personalized AI agents that are embedded in the network itself.
This solves two problems, he argued. First, it provides a compelling use case for consumers by giving them a personal assistant that is instinctually in tune with the rhythms of their lives.
For instance, if a consumer gets a call from an unknown number, but the agent realizes the number is for a landscaping company or school that the customer has taken several calls from before, it can advise to answer the phone. The AI could also provide information about the last time the user spoke with a given caller and context about what they discussed.
Second, in a world where AI regulation is still playing catch-up, it can help alleviate concerns around privacy and security because the AI would be governed by existing telco regulations like the FCC’s CPNI rules, Kanuganti said.
But beyond use cases and regulations, there’s one other key question: how do telcos sell it?
Kanuganti said token packages akin to old school voice, text and data plans are the answer. He added that he expects the first version of a telco token plan will be available in the U.S. in the next six to nine months.
How will consumers react?
Kanuganti prediction is notable given Personal AI’s publicized engagements with U.S. telcos. It has previously disclosed work with the likes of AT&T, T-Mobile, Verizon and Comcast in different areas.
Getting consumers to buy in, though, could be tricky. Even Kanuganti admitted the token plan adoption curve could span three to six years as consumers grapple with what exactly they’re buying. But Kanuganti noted the industry has been here before.
With the advent of data plans, he said it was initially hard for consumers to understand what exactly they could do with 1 gig or 5 gigs or 10. But once the applications caught up – think things like Netflix – it was a lot easier for telcos to say ‘you can stream this many movies with this many gigs.’ It will be a similar journey with token plans, he said.
“Initially it doesn’t make sense,” he said, noting only a small group of people will really grasp that they can handle 100 missed calls and screen 300 more with their AI token plan. “But I think the journey is the same. The journey is no different here” than with any of the other products telcos have put out there, Kanuganti said.
Can telcos win with tokens?
Analysts, however, are skeptical.
Recon Analytics Founder and Principal Analyst Roger Entner told Fierce he doesn’t think a straight “by the token” metering system would readily find consumer acceptance. He noted that Recon Analytics’ data shows more than a quarter of mobile customers are already using AI on a daily basis, and many AI companies are selling bucket plans for access.
“For a telco to be successful, they would have to offer a similar structure and then combine it with a good AI model,” he argued. “They are pushing into an already crowded market.”
Entner’s point about competition isn’t one to take lightly. After all, telcos have already been well burned by cloud providers eating their lunch before, even in the API arena, where operators have sought to make a monetization play.
But Kanuganti argued that while the pendulum swung away from telcos’ core talk and text services and in the cloud’s favor with the advent of data, AI is seeing it swing back.
“The fundamental shift is AI is not APIs anymore. It’s fundamentally like language. That means more power to voice, more power to text, more power to basic language,” he said.
Kanuganti also argued that the ecosystem – notably Nvidia – is now much more invested in telco success. “Yes, it is a challenge. We will have our doubts but given level of investment and focus and attention, I see this as an inevitable path,” he added.
Disruptive Analysis Founder Dean Bubley, though, was similarly skeptical that token plans are the way telcos finally post a win.
“There might be ways to sell or bundle access to AI assistants, but I think counting tokens is a particularly bad way to do it… Trying to sell AI applications on token numbers is like trying to sell content by the pixel. Nobody understands it.” he told Fierce. “Very similar queries may imply very different numbers of tokens. Different types of tokens relate to different tasks. Agents will consume tokens as well. Nobody knows how many they use, except developers accessing via APIs.”
Beyond the consumer understanding element, Bubley pointed out that there are huge shifts happening with enterprises internally around routing the right tasks to the right model and managing exploding token budgets. “Consumers (and telcos’ consumer offers) are unlikely to be able to keep up with that,” he said.
So, what’s the answer then? According to Bubley, telcos should focus on selling complete services or applications rather than taking a piecemeal token approach.