“Generative AI changed the pace and the mindset, and agentic AI is now shifting the objective from containing contacts to resolving them.”

Telcos have spent more than a decade trying to make conversational AI pay off in their contact centres. Intent-based bots disappointed, generative AI reset expectations, and agentic AI now promises to resolve customer requests rather than simply answer them. This Appledore Research note examines what that shift means for CSPs.

We trace the evolution from rule-based bots to autonomous agents and set out why chatbots that inform but cannot act still leave operators with high cost to serve and fragmented journeys. We assess the four foundations CSPs need in place to scale agentic AI across care and sales, and argue that they must be advanced together rather than as separate workstreams. We also weigh the open risks, from the maturity of security controls to the stubborn challenges of real-time voice.

The note closes with practical recommendations for operators. It is written for CSP customer experience, digital and IT leaders, and for vendors positioning conversational and agentic AI solutions to telcos.

Companies mentioned in this report: Telefónica and Vodafone.