AI in public administration for internal knowledge management: A case study by ONTEC AI and the Austrian Federal Computing Center

Public service faces numerous challenges. One of the biggest is the loss of knowledge due to retirements and the difficulty of recruiting new personnel. This is where the solution presented in this case study comes in: the “Digital Knowledge Butler,” an advanced innovation project by the Austrian Federal Computing Center, is a forward-looking solution to preserve the knowledge of departing experts and make it accessible to future generations of employees.

a robot dressed as a butler, serving books

The solution presented in this case study — the Digital Knowledge Butler — was developed based on AI-powered audio documentation.

The Austrian Federal Computing Center (BRZ)

The Austrian Federal Computing Center (German: Bundesrechenzentrum, hereinafter referred to as BRZ) is Austria’s leading provider of IT services, specializing in delivering innovative, secure IT solutions for the federal administration.

As a central pillar of the country’s digital transformation, the BRZ actively promotes the implementation of digital initiatives and significantly contributes to the modernization of public services.

Despite growing from 1,300 to 1,700 employees since 2020, the BRZ, like other public institutions, faces the challenge of a wave of retirements and the associated need to secure knowledge and recruit new talent.

Given the upcoming wave of retirements in public administration, it is essential to find efficient solutions to secure knowledge despite personnel changes within organizations

Matthias Lichtenthaler, Head of Digital Government & Innovation at BRZ

The challenge: preserving expert knowledge

With an impending wave of retirements in public administration, the urgency to secure the expertise of long-standing experts is growing.

Traditional methods of knowledge transfer are insufficient to effectively pass on extensive and specialized knowledge to new employees. The core problem is that often not everything is fully documented, and departing employees may leave the organization before their successors are onboarded. Thus, efficient knowledge transfer is often not feasible.

The BRZ faced the question: How can the wealth of employee knowledge be collected, processed, and made available to others in a simple, unbureaucratic way?

The solution: the Digital Knowledge Butler

With the support of ONTEC AI in AI consulting and development, the idea of the Digital Knowledge Butler was brought to life.

The Digital Knowledge Butler aims to digitize expert knowledge through 100-second audio briefings on specific topics. Employees can quickly record their knowledge or answers to specific questions on their phones.

These knowledge modules are managed by artificial intelligence, which not only answers questions but also identifies knowledge gaps. Experts can then be consulted to fill these gaps.

By the way: In collaboration with ONTEC AI, the BRZ has also developed another knowledge management project aimed at more efficient citizen communication.

ONTEC AI enables employees to document their valuable expertise via voice message using AI-powered audio documentation and make it available to the entire organization.

Technological implementation of the project

The solution is based on a well-thought-out data flow, where audio briefings are converted into text using speech-to-text technology. Originally, a knowledge graph was extracted from this, allowing the AI to answer questions or forward questions it couldn’t answer to human experts.

The underlying AI framework is the existing BRZ-AI Hub. It is designed to learn to recognize areas where knowledge is lacking or insufficiently documented.

Matthias Lichtenthaler, Head of Digital Government & Innovation at BRZ

Over the course of the project, the technological foundation of the Digital Knowledge Butler has significantly evolved. Initially relying on intent hierarchies and ontologies, the latest iteration uses LLM and semantic search methods to enable more precise and flexible handling of knowledge.

ONTEC AI’s hallucination suppression methods play a crucial role in identifying unknown knowledge in response to questions, allowing experts to address these gaps before they leave.

Despite changing project managers and evolving requirements, the project was characterized by high professionalism in approach and implementation.

Summary and key takeaways

The Digital Knowledge Butler exemplifies the successful combination of technological innovation and effective project management to proactively address challenges in public administration and secure the knowledge and experience of departing generations for the future.