Deutsche Telekom Replaced Phone Menus With AI Agents
The carrier now routes customer tasks to automated systems in up to 60 languages, with a live translation tool planned.
Updated on Sept. 24, 2026 in Artificial Intelligence
Live Poll
Do you prefer resolving customer service issues with AI agents rather than human representatives?
Deutsche Telekom has shifted customer support operations across 10 European countries to AI agents capable of handling tasks like router reboots and billing. These automated systems currently manage 20% to 40% of incoming calls shortly after implementation.
Why it matters
The transition replaces traditional phone menus with 24-hour service in 40 to 60 languages, aiming to eliminate hold times while improving call classification accuracy. This deployment indicates a broader industry trend of integrating generative systems into legacy telecommunications infrastructure.
The company achieves integration by adding caching layers and adapters to older systems, while using computer use to navigate interfaces that lack APIs. These agents are currently managing a high volume of tasks, supporting between 40 and 60 languages.
The players
Deutsche Telekom
A European telecommunications provider operating in 10 countries with a focus on integrating AI into consumer-facing support stacks.
Kartik Sheth
An executive who detailed the company's automation roadmap at the HumanX industry conference.
Lovable
A development platform used by Deutsche Telekom to build over 2,000 internal applications.
Decagon
A startup partner specialized in AI-driven customer service solutions.
Listen Labs
A startup collaborator involved in the company's voice and language processing initiatives.
The details
Deutsche Telekom employs AI agents to automate service requests by interfacing directly with existing backend infrastructure. When legacy systems lack formal Application Programming Interfaces (APIs — software intermediaries that allow two applications to talk to each other), the firm utilizes computer use to navigate the software visually. The architecture is supported by more than 2,000 internal applications developed on the Lovable platform, allowing for rapid scaling of service capabilities.
Timeline
September 18, 2026: Kartik Sheth presented on automation strategies at the HumanX conference in Amsterdam.
Calendar year 2026: The company plans to release a real-time call translation assistant.
Next one to two years: The share of AI-handled interactions is expected to continue its upward trajectory.
The Tech Race
This deployment follows a pattern set by the 2026 HumanX conference roadmap on enterprise AI agents, which emphasizes the transition from pilot programs to large-scale, multi-language support. Deutsche Telekom is currently scaling these agents against a target of increasing automated interaction shares monthly.
Users in the company's 10 European markets can expect reduced wait times and 24-hour availability for routine tasks like router reboots. The company requires explicit user consent before any call transcription or translation features are activated.
The takeaway
The firm is transitioning toward an automated service model that aims to scale AI-handled interactions continuously over the next two years. Interested observers should monitor the official release of the company's real-time call translation assistant slated for late 2026.
Further reading
For more developments on the intersection of language models and service automation, explore Artificial Intelligence.
Live Poll
Do you prefer resolving customer service issues with AI agents rather than human representatives?
