On 29 October TelecomTV’s Unthinkable Lab takes up AI Voice for telcos. In the preview, Andrew Collinson of Connective Insight, who researched and wrote CPaaSAA’s March AI Voice research, names the piece he is still testing: the runtime, the environment in which a live call is managed and orchestrated. His March hypothesis, he says, “could be completely out of date by now.” Dean Bubley of Disruptive Analysis argues that a service provider is a good nexus for that orchestration, because it can anchor the business’s phone number.
CPaaSAA put that question to the room at CASA26. Sam Wilson, CEO of 8×8, told the room that within a few years no call will reach an enterprise without passing an AI agent first. If that is right, the question that matters is who controls the software that manages the call. So the room was asked.
Of 42 votes, 48% expected Intelligent Engagement platforms to run the AI Voice conversation by 2030. A third, 33%, favoured telco-integrated models. Hyperscalers took 14% and enterprise-controlled models 5%. Participants see the runtime mainly as a contest between platforms and operators.
Operators hold a third of the votes and a sliver of today’s deployments
Set that against the estimates from CPaaSAA’s AI Voice research, which describes four architectures for running the orchestration layer. Intelligent Engagement platforms are estimated at 40 to 60% of enterprise deployments today. Hyperscaler-led deployments are about 20 to 35%. Enterprise-controlled deployments are about 10 to 20%. Telco-integrated deployments are below 5%.
The gap between the vote and the deployments is not a contradiction. Each architecture wins for a different reason. Platforms win on governance and supplier neutrality, at the cost of some simplicity. Hyperscalers win on speed and architectural simplicity, at the cost of control. Enterprises that treat AI Voice as strategically differentiating build their own and accept the engineering overhead. Operators win where sovereignty, trust and resilience come first, and where AI should be in every call by default, at the cost of flexibility.
The operator products are in the call path now
That is why the third matters. Operators entered this market with real products during the year. Deutsche Telekom’s Magenta AI Call Assistant is network-native and needs no application or particular device. KPN launched Mobiel AI as an addition to a mobile subscription, deliberately as assistance rather than autonomy. Radisys reports demand from tier-one operators across several markets and is deployed with Rakuten Mobile. Ralph Page of Radisys told the room AI Voice had “gone mental”, and that conversations with operators had moved from technical teams to CxOs and RFPs.
Telco-originated AI Voice is generally embedded inside existing services, so it can be distributed fast once built. That is an advantage platforms do not have, and it is why a low share of deployments today does not tell you much about 2030.
Positions are being set before the market settles
The runtime is one of the clearest places where intelligence revenue is beginning to appear, and it is being contested by platforms, hyperscalers, AI-native specialists and potentially Meta. If operators wait, much of the runtime market may form around other players. If a messaging app gets there first, the conversation moves onto its platform, and that looks more like the Absorbed Engagement scenario than the Platform-Led future most participants voted for.
The vote says who people think will win. The deployments say who is in the call path today. The work for the next twelve months is to close the distance between the two, from whichever side you stand on.
Researched and written by Andrew Collinson of Connective Insight, commissioned by CPaaSAA.
I've spent my career in IT and telecoms building bridges, from the early days of cloud communications to today's reset around APIs, CPaaS and AI.
As Founding Partner of The Next Cloud and CPaaSAA, I help telcos, CPaaS players and AI-native scaleups work out where they fit in the AI era, through strategy, ecosystem building and events.
My view: the value is moving from platforms to outcomes, and real innovation comes from ecosystems, not incumbents acting alone.
Based in Amsterdam, usually somewhere else.
Comments are closed