Background
Anyone who wants to apply for public services often faces a real obstacle: complex forms, unclear responsibilities, and long waiting and processing times. For vulnerable and disadvantaged groups in particular, finding the right service can be difficult. At the same time, administrative staff are under heavy pressure every day due to high volumes of communication, incomplete application documents, and time-consuming processing procedures. Digital forms barely improve the situation when they simply transfer paper forms one-to-one onto a screen, because the process remains just as complicated.
In recent years, however, Artificial Intelligence has developed enormous potential to fundamentally transform exactly these kinds of processes. In particular, agentic AI makes it possible for essential parts of an application to be completed automatically. Intelligent systems can guide citizens through complex application processes or even fill out the applications themselves, ensuring they arrive at the administration more complete and better structured – to the benefit of everyone involved. Beyond Forms aims to explore what this could look like in practice.
Project goal
Beyond Forms is a two-year research project exploring how the application process between citizens and public administration can be improved for both sides with the help of Artificial Intelligence – and what it means to truly move beyond the classic digital form.
The vision behind the name: In the future, citizens could apply for public services with the help of an AI agent without having to fill out a form. Instead, an intelligent system would guide them through the process – clearly, proactively, and tailored to their individual life situation.
Concrete goals include:
- Easier access to public services, especially for disadvantaged and vulnerable groups
- Reduced workload for administrative staff and relief from high communication volumes
- Shorter processing times through more complete and better structured applications
- New access pathways beyond text-based forms
- Proactive and personalized recommendations for services that match citizens’ life situations
Approach
Beyond Forms follows a two-track approach. On the one hand, existing application processes are analyzed in depth together with the administration to understand where concrete obstacles arise for both citizens and staff, and where there is potential for improvement. On the other hand, the project explores technically how agentic AI systems could fundamentally redesign these processes – for example through chatbots, speech and image recognition, or automated document analysis.
Based on these insights, prototype components will first be developed and tested using real application processes, and then iteratively improved.
As an initial use case, the project focuses on access to public services for disadvantaged and vulnerable groups and explores how this access could be improved with the help of AI. In close collaboration with the Senatsverwaltung für Arbeit, Soziales, Gleichstellung, Integration, Vielfalt und Antidiskriminierung (SenASGIVA), existing barriers will be analyzed and new solution approaches developed.
Data protection and security have the highest priority throughout the project. An infrastructure will be established to ensure that sensitive data is protected. All results of the project will be published as open source.
Funding & Collaboration
Beyond Forms is implemented at CityLAB Berlin with funding of one million USD from Google.org, the philanthropic arm of Google. During the first phase of the project, the CityLAB team works for six months together with an interdisciplinary team of Google.org Fellows with expertise in technical development, UX, design, and project management. Afterward, CityLAB Berlin will continue the project independently until the end of 2027.
What’s next?
At the end of 2025, initial potential use cases were identified. Beyond Forms is currently in the phase of technical exploration and service design. In the coming months, prototype components will be developed and tested within real application processes. By mid-2026, a prototype is expected to demonstrate the feasibility of the approach and prove the added value of the solution. The next step will be to further develop it toward product maturity so that individual components can gradually evolve into practical solutions. By the end of 2027, the results will be fully documented and published as open source so that other cities and administrations can use them.
Team
CityLAB Berlin, Google.org Fellows, Senatsverwaltung für Arbeit, Soziales, Gleichstellung, Integration, Vielfalt und Antidiskriminierung (SenASGIVA)
Contact
Björn Mohr (Projektmanager Verwaltungsinnovation)
