Deutsche Telekom Replaces Phone Menus with AI Agents Across Europe

Deutsche Telekom is phasing out automated phone menus in favor of AI agents for customer service across its 10 European markets. The system, which operates in 40 to 60 languages and requires no hold time, is already live in some countries and aims to handle an increasing share of routine interactions.

AI Agents Replace Phone Menus Across Europe

Deutsche Telekom is replacing its automated phone menus with AI agents for customer service interactions, according to claims reported by The Next Web. The company, which operates in 10 European countries, aims to streamline customer experience by using AI to handle calls directly.

The AI agents are designed to answer calls 24 hours a day in 40 to 60 languages, eliminating hold times entirely. This shift is already live in some markets, with others expected to follow within months.

AI Adoption Grows, But Scaling Remains a Challenge

When Deutsche Telekom introduces AI for a new use case, the agents initially handle around 20% of calls. This share typically scales to 40% within months, though progress slows significantly beyond that point. According to The Next Web, moving from 40% to 60% adoption is far more challenging, and reaching 80% is described as "really hard."

The AI agents manage routine tasks such as checking or rebooting routers and explaining billing inquiries. The monthly share of customer interactions handled entirely by AI grows by 1-2 percentage points.

Partnerships and Infrastructure Drive AI Deployment

Deutsche Telekom’s AI agents operate across diverse tech stacks and regulatory environments in countries like Poland, Greece, and Germany. To navigate these complexities, the company has built over 2,000 apps on its Lovable platform, reducing reliance on third-party vendors.

The company collaborates with AI startups, including Decagon and Listen Labs, to develop and refine its AI tools. Deutsche Telekom provides infrastructure, co-sells solutions to small businesses and consumers, and invests in its partners.

Real-Time Translation Tool Coming This Year

Deutsche Telekom announced plans for a call assistant capable of real-time translation, unveiled earlier this year in Barcelona. The feature is expected to launch before the end of this calendar year.

The tool will translate calls in real time without requiring special hardware, provided at least one caller is on Deutsche Telekom’s network. Both parties must consent to the translation and transcription, with the company requesting permission from the non-subscriber before proceeding.

Context: AI Deployment in Fragmented Markets

Deutsche Telekom operates in a fragmented regulatory environment across 10 European countries, each with unique tech stacks and compliance requirements. This complexity has influenced its approach to AI deployment, prioritizing flexibility and control over its tooling.

The company’s collaboration with startups like Decagon and Listen Labs reflects a broader strategy to leverage external expertise while maintaining ownership of its AI infrastructure.

Booking.com’s experience, as reported earlier, highlights the risks of building AI tools in-house when frontier models can quickly replicate or surpass custom solutions. Deutsche Telekom’s use of the Lovable platform may help mitigate this risk by balancing internal development with external partnerships.

What this means

LazyFounders analysis — our interpretation, not reported fact.

Deutsche Telekom’s shift to AI agents for customer service reflects a broader trend of enterprises replacing legacy systems with AI-driven automation. For founders, this is a case study in balancing scalability with precision—while AI can handle a significant share of interactions, the last 20-40% of use cases remain challenging and require human oversight.

The company’s partnerships with startups like Decagon and Listen Labs also highlight an opportunity for AI infrastructure providers to collaborate with telecom giants. However, the challenges of integrating AI across fragmented tech stacks and regulations in multiple countries underscore the complexity of deploying AI at scale in highly regulated industries.

Key takeaways

• Deutsche Telekom is phasing out automated phone menus in favor of AI agents across its European markets.
• AI agents currently handle 20-40% of calls for new use cases, with steady monthly growth in fully automated interactions.
• The system supports 40-60 languages, operates without hold times, and manages routine customer service tasks.
• A real-time call translation assistant is set to launch this year, requiring consent from both parties.
• Deutsche Telekom avoids vendor lock-in by building its own tooling and partnering with AI startups.

FAQ

What tasks can Deutsche Telekom’s AI agents handle?

The AI agents manage routine customer service tasks, including checking or rebooting routers, explaining billing inquiries, and answering general queries.

How many languages do the AI agents support?

The system supports 40 to 60 languages, depending on the use case.

What are the challenges in scaling AI adoption for customer service?

While AI agents quickly reach 20-40% adoption for new use cases, scaling beyond that—particularly from 40% to 60%—becomes significantly more difficult.

When will the real-time translation feature launch?

Deutsche Telekom plans to launch its real-time call translation assistant later this calendar year.

How does Deutsche Telekom avoid vendor lock-in?

The company has built over 2,000 apps on its Lovable platform and develops its own evaluation tools to maintain control over its AI infrastructure.

Related on LazyFounders

• Lovable
• More ai news
• Booking.com regrets building AI tools in-house as frontier models catch up