Login
Register
AI voice momentum builds as telcos explore business models
Recent announcements by Telefónica and Mavenir, and earlier in the summer by Jio, point to growing interest from telcos in using AI voice to provide new services that tie more closely to their network capabilities.
Mavenir, which is integrating voice intelligence into its mobile network infrastructure in partnership with speech AI platform Sanas, described AI voice in a statement as a means to “transform voice from a connectivity service into a monetizable, AI-powered experience.”
However, telcos, which face significant competition in the AI voice market from hyperscalers and CpaaS providers, are making no revenue claims, even though the initial handful of Voice AI developments are exploring new ways to deliver outcomes to customers.
In the enterprise space, for example, AI voice can be an opportunity to prevent churn and revenue loss by offering enterprises alternatives to legacy voice services, or it can be part of a broader trusted or sovereign integrated AI service portfolio. For consumers, the emphasis is on persuading them to switch from over-the-top apps to network-integrated AI-driven services that depend on voice. All while creating the possibility to tie AI voice to network capabilities such as identity, security and local presence, and to benefit from analysing voice as structured data.
When it comes to the scale of voice AI ambitions Jio is currently far in the lead. The company, which claims to carry 20 billion minutes of voice a day for more than half a billion people, plans to make a native AI voice service called Jio Call Agent available to all its customers before the end of this year. Its aim is to move customers away from apps and onto network-integrated services:
“No app to download. No number to add. Available to every Jio customer. In every Indian language,” according to statements made at a general meeting by Reliance Industries in June.
Planned services include call transcription for up to 10 unique speakers in their own language and the sharing of summaries and action items, as well as the use of natural language to order services such as food or a taxi. This fits with Jio’s wider AI strategy, which includes turning its MyJio app into “a personalised Jio AI advisor” that customers can operate directly through speech rather than through app navigation.
Deutsche Telekom is also betting on network-based AI functions that will peel customers away from apps with its Magenta AI Call Assistant, announced as a world first in March this year. “We remove barriers — no apps, no special devices, no technical complexity,” stated the company at the time. Initial services will also include live translation, call summaries, and answering questions.
But it is still early days for telcos when it comes to integrating AI voice in the network for consumers, with Deutsche Telekom and Jio still to roll out mass-market services.
When it comes to the enterprise space, Telefónica, for example, is targeting existing medium and large enterprise customers with its offer of voice AI functionalities such as intelligent transcription, automated call summaries and virtual assistants without them having to change software platforms.
“The solution has been conceived as a cross-service AI layer that can be applied across Telefónica’s entire enterprise voice portfolio, both fixed and mobile, and across traditional as well as cloud-based services, including Centrex IP and Centrex 365,” according to a company spokesperson.
Trust is another selling point of AI voice, according to the Spanish operator, which says it is “seeing growing demand for AI solutions that combine innovation with high standards of protection, trust and governance.”
Orange Business is also emphasizing trust and sovereignty. In March it announced it is combining voice and AI in a range of managed, integrated enterprise services that include branded calling; deepfake detection across audio, video, images and documents; customer care automation for its enterprise customers (Orange Business said it managed 80 million AI conversations in 2025); and agentic telephony.
More recently, Orange has made voice AI available as an API to external developers, explaining that although voice drives 70% of customer service interactions, only 25% of these calls are used to generate valuable analytics.
“Until now reserved for operators’ internal use, voice is about to become a fully programmable building block. With Voice API, Orange puts the power of voice in developers’ hands to build unprecedented use cases, from analytics to voice AI,” according to thr company in an announcement.
Commenting on Linkedin, Sebastian Barros remarks that Orange’s Voice API offer “becomes more interesting when voice is combined with Orange’s other APIs,” to provide branded calling, number verification and SIM swap.
And whether telcos use AI voice to address the consumer or enterprise market, analytics promises to be one of the prizes, as Andrew Collinson, Managing Director, Connective Insight points out: “Where you build a transformational value [with voice AI] is when you can interrogate the information and find out what's really going on. That's where the leading players in the field are going: Turning calls into data that you can interrogate.”
