Whether for perfect “matchmaking” for funding calls or neutral digital assistants for experiments: the cloud platform Microsoft Azure supports researchers from a wide range of disciplines at the University of Vienna. For 3 years now, the ZID has been funding innovative research projects that use Azure and helping those involved to get started with the platform.
Last year’s funding period has expired – time for the next round, a look back, and ahead at old and new projects.
Review of 2025: Focus on text and dialogue
The 2025 funding was characterised by projects that developed and used text-based applications, such as these 2 examples:
SciTextMatch impressively demonstrated how AI can ease the burden of everyday research. The project team used the Microsoft Foundry AI platform via Azure to analyse the potential for collaboration between researchers, for example by comparing publications.
“Our goal was to automate the identification of promising collaboration partners and suitable funding programmes. In this specific case, the focus was on potential collaborations between the University of Vienna and the University of Toronto, which were also suggested for relevant calls for proposals under the EU’s Horizon Europe funding programme. In this way, we want to enable and promote collaboration exactly where it makes the most sense,” explains Lilian Nowak from the Research and Publication Services Department of the Vienna University Library.
At the same time, Can Çelebi from the Vienna Center for Experimental Economics used Microsoft Foundry for his research in experimental economics. In this context, a chatbot can be used to improve participants’ understanding of the experiment, thereby reducing noise caused by incomplete comprehension of the underlying task mechanics. In this role, it is essential that the chatbot supports understanding while avoiding any information that could bias participants’ beliefs, preferences, or decisions and thereby affect the experimental results.
As part of his project, Çelebi developed additional LLM-based agents that simulate participants with different characteristics and interact with the instruction bot in order to stress-test it. This makes it possible to examine whether the bot can be pushed into giving information that could influence participants rather than simply helping them understand the task. “Neutrality is very important in research. Thanks to the possibilities offered by Azure, we can now better determine whether an AI remains in the role of a neutral helper in the context of an experiment,” explains Çelebi.
Outlook for 2026: from IT security to vaccination advice
A look at the 8 new funded projects once again reveals a wide variety of application scenarios for Azure services in research:
- Thematic diversity: The spectrum ranges from physics, where complex simulations are to be carried out in the field of magnetism research, to law, where the thicket of Austrian constitutional law is being analysed using language models.
- Focus on security: 2 projects are testing the robustness of AI models against attacks or attempting to automatically detect phishing web pages through a combination of image and text analysis.
- Social impact: A project by the Faculty of Psychology is testing how AI chatbots can improve vaccination counselling by presenting medical information in an understandable way.
Azure as a multi-tool
The use of the cloud is becoming more diverse: “In the context of AI, the projects in the first funding rounds exclusively used various GPT models. Now, the entire range of Microsoft Foundry is increasingly being taken advantage of, including a variety of different language models such as Claude, Grok, Mistral and DeepSeek,” reports Fabian Jusufi from the Coordination Digital Transformation staff unit at the ZID.
He adds: “Some projects use the scalable computing resources of the Azure Cloud. For example, a project at the Department of Political Science connects these resources to the image analysis capabilities of Azure Vision models. This allows several thousand high-resolution photos of election posters to be evaluated.”
The projects in this year’s funding round are now starting their implementation in Azure. Fabian Jusufi and the entire Coordination Digital Transformation staff unit are looking forward to accompanying the researchers as they take their next steps into the cloud.