How Deutsche Telekom Built a Governance-First AI HR Stack — 2.5M Chatbot Conversations, Skills Intelligence, and EU AI Act Readiness
When Deutsche Telekom's internal AI chatbot crossed 2.5 million employee conversations, it marked more than a usage milestone. It signaled that one of Europe's largest employers — with roughly 80,000 employees in Germany alone — had moved AI out of pilot programs and into the operational fabric of its HR function. Paired with a multi-year talent intelligence partnership with Eightfold AI, the telco offers a rare look at what a governance-first, enterprise-scale AI HR stack actually looks like in practice.
For HR leaders at large European organizations watching the EU AI Act's August 2026 high-risk compliance deadline approach, Deutsche Telekom's phased playbook is worth studying closely.
askT: An AI Assistant Embedded Across 30+ Departments
At the center of Deutsche Telekom's AI-HR architecture sits askT, an internal chatbot built on ChatGPT technology and approved for use with data classified as "internal." Originally designed to field HR service queries — benefits, sabbaticals, employee share programs — askT has expanded to cover specialist knowledge from more than 30 Telekom departments, assisting with report creation, workshop development, and general task support.
The adoption numbers are notable for a workforce of this size. Approximately 10,000 of Deutsche Telekom's 80,000 German employees — roughly 12.5% — use askT on a weekly basis as of late 2024. The chatbot has logged more than 2.5 million conversations since launch.
But Deutsche Telekom has not relied on the tool alone to drive AI fluency. In 2024, 30,000 employees improved their prompting skills through structured training programs. As of September 2024, 90,000 employees across the company have acquired AI-relevant skills — a figure that reflects the organization's broader commitment to building AI literacy at scale.
Eightfold: From Recruiting to Resource Management
Deutsche Telekom's partnership with Eightfold AI started with recruiting and internal mobility — using the Talent Intelligence Platform to analyze applicant data, accelerate hiring, and surface internal career pathways.
In March 2025, Eightfold published a case study detailing the partnership's latest expansion: Resource Management, an AI-powered solution for project staffing. Deutsche Telekom partnered with Eightfold to help shape this capability, which uses deep-learning AI to match employees with projects across borders and business units.
Ilja Bitterling, VP of Skills Intelligence and Performance Management at Deutsche Telekom, framed the expansion as a strategic necessity: "We knew that in order to create this end-to-end value, we needed to bring in the resource management use case."
The move reflects a broader industry pattern: organizations that start with AI-assisted recruiting increasingly find that the same skills data can power internal mobility and workforce planning. For Deutsche Telekom, the Eightfold platform now spans the full talent lifecycle — from candidate screening to project-level deployment.
Eightfold AI itself achieved ISO/IEC 42001 certification, providing an additional layer of vendor-side AI governance assurance.
EU AI Act Readiness: Audit Complete, No Prohibited Systems Found
Where many enterprises are still mapping their AI inventory, Deutsche Telekom has already completed an EU-wide audit of all AI applications. The result: no prohibited AI systems were found in use or in the company's products.
This is not incidental. Deutsche Telekom has maintained digital ethics assessments for AI technologies since 2020, embedded within a broader Privacy and Security Assessment process. The company also deploys ICARE Check, a tool designed to help employees handle artificial intelligence responsibly.
On the regulatory engagement front, Deutsche Telekom is a member of the EU AI Pact, voluntarily committing to three core actions: developing an AI governance strategy, mapping high-risk AI systems, and promoting AI literacy among employees. With the EU AI Act's high-risk compliance deadline set for August 2026 — with an August 2027 extended deadline applying to high-risk systems already subject to third-party conformity assessment — Deutsche Telekom's early moves position it well ahead of many peers.
Zsuzsanna Friedl, Chief HR Officer of T-Systems, underscored the cultural dimension of this readiness: "The key is to introduce all employees in all areas and functions to working with AI and to continuously promote their willingness to learn."
The Blueprint for Enterprise HR Leaders
Deutsche Telekom's approach rests on three pillars from its broader AI strategy — Taking (purchasing market-ready solutions like askT), Shaping (adapting tools for specific needs), and Making (building proprietary solutions). The company has deployed more than 200 AI solutions for business customers and employs approximately 1,500 AI specialists across generative AI, quantum AI, predictive analytics, and large language models.
For CHROs and HR directors at large European enterprises, the Deutsche Telekom case illustrates several replicable principles. First, starting with a broadly accessible tool — askT — created organizational familiarity with AI before introducing more complex applications like skills-based project staffing. Second, investing in upskilling alongside deployment (30,000 employees trained in prompting in a single year) ensured the workforce could actually use what was being built. Third, embedding governance from the start — through ethics assessments, compliance audits, and voluntary pact membership — turned regulatory readiness into a competitive advantage rather than a last-minute scramble.
With 76% of HR leaders believing that companies not adopting AI in HR within one to two years will fall behind, the window for building this kind of infrastructure is narrowing. Deutsche Telekom's phased, governance-first approach offers a concrete model for getting it done.
Sources
- T-Systems Management Blog — "How AI is Revolutionizing the HR Working World": https://www.t-systems.com/de/en/insights/newsroom/management-unplugged/how-ai-is-revolutionizing-the-hr-working-world-1077428
- Eightfold Customer Story — "Deutsche Telekom Seizes Opportunity to Apply AI to Project Staffing" (published March 4, 2025): https://eightfold.ai/customers/customer-stories/deutsche-telekom-seizes-opportunity-ai-project-staffing-eightfold-resource-management/
- Deutsche Telekom EU AI Act Governance Page: https://www.telekom.com/en/company/digital-responsibility/details/the-eu-ai-act-at-deutsche-telekom--1090710
- Deutsche Telekom AI Strategy — Shape, Take, Make: https://www.telekom.com/en/company/details/shape-take-make-ai-at-deutsche-telekom-1078506
What is askT at Deutsche Telekom?
askT is Deutsche Telekom's internal AI chatbot built on ChatGPT, approved for data classified as 'internal.' It has logged over 2.5 million employee conversations across more than 30 departments, covering HR queries, vacation requests, sabbatical applications, and IT help.
How is Deutsche Telekom using Eightfold AI?
Deutsche Telekom uses Eightfold's Talent Intelligence Platform for recruiting, internal mobility, and project staffing. The company helped shape Eightfold's Resource Management capability, which uses deep-learning AI to match employees to open projects across borders and business units.
Is Deutsche Telekom compliant with the EU AI Act?
Deutsche Telekom has completed an EU-wide audit of all AI applications and found no prohibited systems. It is a member of the EU AI Pact, has deployed digital ethics assessments since 2020, and uses the ICARE Check tool for responsible AI deployment.
What is the EU AI Act's high-risk compliance deadline?
The EU AI Act's high-risk AI compliance deadline is August 2026, with an extended August 2027 deadline for high-risk systems already subject to third-party conformity assessment.
What can enterprise HR leaders learn from Deutsche Telekom's AI HR strategy?
Deutsche Telekom's approach — deploying an accessible chatbot first (askT), expanding to skills-based talent intelligence (Eightfold), upskilling 30,000 employees in AI prompting in one year, and embedding governance from the start — offers a replicable blueprint for large enterprises navigating AI HR adoption.