The context​

Germany’s data ecosystem is shaped by its federal structure, with responsibilities distributed between the national government (Bund), the 16 federal states (Länder), and local municipalities. This decentralized system has led to a fragmented approach to data management, where each ministry and authority operates largely independently, maintaining its own datasets with limited cross-departmental exchange. A cultural transition has only just begun within public administration regarding the strategic use of data, with many authorities still treating data as a passive resource rather than an active tool for policymaking and governance. In most ministries, non-personal data is only collected and processed within individual departments and is often locked away in departmental silos without a central knowledge management system to provide an overarching view of available data assets.  

Prior to 2021, federal ministries faced a significant lack of dedicated data teams. Only a few ministries had  units responsible for data science, data governance or analytics services to support decision-making. Institutional structures to enable data-driven government action, such as data labs, were largely absent in downstream ministries and federal authorities, restricting opportunities for experimental and innovative use of data. Compounding this issue was a shortage of skilled personnel with expertise in data analysis, visualisation and interpretation (although public administration institutions in general have traditionally lacked dedicated positions for this line of work).  

Recognizing these challenges, the German Data Strategy 2021 sought to modernize public administration by establishing data laboratories within ministries to foster cross-departmental collaboration, develop data literacy among staff and encourage the use of artificial intelligence (AI), predictive analytics and data-driven policymaking. The initial funding for the data labs originates from the German Recovery and Resilience Plan (DARP). This plan is financed by the European Union through the NextGenerationEU recovery instrument, which was established in response to the COVID-19 pandemic.  

‍

Data labs: Theory of Change

The primary goals of data labs are to strengthen evidence-informed policymaking, enhance data literacy within federal ministries and improve ministerial efficiency through data-driven applications. Additionally, they foster cross-departmental and inter-ministerial collaborations that support these objectives.

One of the main challenges of the data labs is that they still need to demonstrate their value which means they must balance short-term, high-visibility projects with long-term, impactful initiatives. Prioritizing quick successes can risk diverting attention from broader, more strategic goals that require sustained effort—such as the development of a robust data governance framework.

Data labs typically engage in four key activities:

1. Development of data products

Common examples include the use of large language models (LLMs) or the aggregation of data into dashboards. For instance, one of the data labs is developing an AI assistant for government officials, in the form of a cloud-based, agentic AI system. This means it can act autonomously to complete tasks and make decisions. The AI assistant is designed to seamlessly integrate with existing sources of knowledge and data. The assistant will have various functionalities, such as a general chatbot, an internet search function and access to internal documents and databases, with the aim of making it easier for employees to access the right information at the right time. The AI system is designed to be user-friendly and secure, ensuring that it meets the specific needs of its users in the context of government. The development of the AI system is iterative and will incorporate real user needs to continuously enhance its functionality.  

The vision for the AI system is to expand its capabilities and integrations with other agentic tools, allowing it to leverage new models as they become available. As the AI system evolves, it will perform more complex tasks, which will boost efficiency and allow for more resources to be channeled towards achieving the ministries’ goals. It is important to note, however, that the agent will not make independent decisions; the final say will always rest with employees (“human-in-the-loop”).

2. Data literacy training

The training conducted through the data labs will equip government staff with the essential skills to navigate the complexities of data and AI and better understand how they can effectively use data and AI in their daily work. The training modules encompass a broad range of topics, including foundational knowledge of data and AI, data visualization, ethical and legal considerations, as well as practical skills in prompting large language models. The Data-to-Policy Navigator—a powerful online tool developed by Germany’s Federal Ministry for Economic Cooperation and Development (BMZ) and the United Nations Development Programme (UNDP)—serves as a valuable resource for these types of training programmes as it offers a curated collection of use cases that can be used as practical examples. By focusing on both theoretical understanding and practical application, these trainings foster a culture of data-informed decision-making within Government.

3. Building data infrastructure

In collaboration with its implementing agencies, one of the data labs is in the process of developing a shared data space—a secure, structured environment where multiple stakeholders can access, exchange, and use data according to shared standards and rules. The data space initiative seeks to establish a robust technical infrastructure to support data exchange, alongside the implementation of essential data governance frameworks and user access protocols. The overarching aim is to minimise information loss and enhance the impact of developed applications. To this end, the initiative seeks to leverage shared data for the collaborative development of applications across organisations, including the sharing of existing data and AI products. At present, the data space remains in a developmental and testing phase.

4. Establishing new partnerships and promoting interministerial cooperation

Interministerial collaboration is a primary focus of the data labs. They act as catalysts and intermediaries between data analysis, policy processes, and technical implementation. Additionally, team members in the data labs regularly get together in a peer-to-peer network that facilitates knowledge exchange on projects and products across ministries. Beyond government, the networks also engage with civil society, academia, and the broader public to strengthen cooperation and innovation in data-driven policymaking.

How can better data contribute to better policy?

The data labs have advanced a culture for data-informed policymaking by making the use of data a collective responsibility across all departments. The decentralized nature of the data labs addresses the unique characteristics of each ministry—such as work cultures, data affinity, and infrastructure—by translating these differences into tailored applications and training formats. Innovation is fostered through collaboration and mutual learning among a diverse range of actors. By working with agile principles (inspect, adapt, improve fast), the data labs shorten development cycles, enhance adaptability to a rapidly changing world, accelerate the creation of practical solutions, and better prepare ministries for successful implementation. In doing so, the data labs equip ministries with the tools, skills and flexible processes needed to harness data and AI effectively, thereby contributing to enhanced data- and evidence-informed policymaking and enabling more targeted, effective and responsive public policies.

Where do we go from here?

In 2026, the current funding of the data labs expires. The political leadership of each ministry must decide whether they continue to finance their data lab through the ministry’s annual budget. Until then, the data labs will continue to develop data products, AI products and trainings, while initiating long-term transformational change towards a more innovative, effective and data-informed government.