Relaunch of FinanzOnline and integration of chatbot “Fred”



Policy areas

Organisation name FinanzOnline, Bundesministerium Finanzen

Share this article via:

Contact person: Dr. Angelika Schätz

Good Practice Certificate:

FinanzOnline was the first transactional portal of the Austrian public administration, launched in 2003. It is still the flagship of e-government applications in Austria. Through the course of time, a large number of services have been integrated into FinanzOnline, where the overview for the users is often lost.

As part of the ‘Customer Service 4.0’ strategy project, Danube University, Krems identified potential for improvement in the entire customer service of the tax administration, thus also for FinanzOnline. It quickly became clear that a significant revision was needed, especially for those users who rarely use the portal but make up a large portion of all users. These are the citizens who want to process their employee tax assessment or applications for family allowances via FinanzOnline.

For citizens to enjoy using the new portal, early adoption into the development process is essential. Internal analysis has been held on the BMF co-creation platform (, where assistance for the tax return was one of the highest-ranked ideas. The other findings were for the implementation process and several design-thinking workshops were carried out with the involvement of the target audience. The goals and the scope of the necessary changes were further elaborated upon and modified where needed.

The technical implementation of the changes was carried out in 2019 in the form of an agile development method (Scrum). Since the culture of the public administration requires a certain amount of formalisms – even with an agile development method – a steering group of technical and IT members was set up in addition to a steering committee. The results from the ongoing rapid prototyping were accepted and commissioned for implementation.

All goals set for the restructuring of FinanzOnline were set to be implemented in 2019. The Customer Service 4.0 project dealt extensively with the customer service issues and with the issue of telephone information. The number of calls has doubled to around five million per year over the past 15 years. As the number of staff decreases, it is more and more difficult to deal with in a satisfactory time. Accordingly, another winning idea from the co-creation platform e3lab was taken up, with a chatbot answering all questions about employee assessment. It became clear that a chatbot would take the strain off the telephone lines in the medium term and that customers could also be offered a new 24/7 service. As part of the design-thinking workshops, the basics as well as the name and avatar of the chatbot were developed. At the same time, an editorial team was set up and trained to load Fred with the necessary knowledge. Citizens are used to online services which they can use to deal with their concerns quickly and easily; administration services not often state-of-the-art. With the relaunch of FinanzOnline, a very user-friendly and transparent interface was implemented with the involvement of the target audience. Even better, an interactive assistant was also integrated for the most important service – the employee tax assessment – which guides the user through the tax return. This should be as easy as ordering a product in an online shop.

Fred was one of the first chatbots deployed by the public administration in Austria. The way information is provided to the citizens was new to the public administration and required a change of culture. Traditionally, the user had to search static websites for tax-related information. With Fred, the citizens enter a dialogue with a software agent. The citizen describes their request and the algorithm of the bot delivers rather small, easy-to-understand items of information from which the citizen can explore further by asking more questions. This contrasts with many public administration websites that often comprise exhaustive texts to cover every legal detail, which is not even relevant for the majority of cases. Because of the size restrictions and the dialogue-oriented character of the communication information provided by a chatbot, it has to be structured and formulated in a very different way.

Also interesting