AI agent for social research

What's behind it

To make this work, several agents work together in the background:

• Interview agent: conducts semi-structured interviews independently and responds specifically to answers.
• Moderator-Agent: ensures that the conversation does not stray from the topic.
• Memory Agent: memorizes the most important findings in order to build on them later.

The result: less manual work for the researchers and at the same time more consistent, higher-quality data.

Why this is important

In social research, qualitative interviews take a lot of time - from conducting them to evaluating them. Our AI-supported approach helps the PH Lucerne to make this process more efficient and scalable without compromising on quality.

More interviews. More time for the essentials. More value from the data.

Background

Cooperation between teachers is considered an important success factor for everyday working life and for students' learning processes. It is now a professional standard and part of teacher training. Nevertheless, many prospective teachers feel inadequately prepared for such collaborative activities.

How and with whom do students learn collaborative practices? This is precisely what the Lucerne University of Applied Sciences and Arts is investigating in the current research project - from a socio-constructivist and social-network-oriented perspective.

How the data collection setup works

Round 200 students from the primary and secondary level 1 study programs at the Lucerne University of Teacher Education are taking part. They visualize their personal learning networks in a Web app - so-called egocentric network maps.

Afterwards, the students discuss their map in an interview. And this is where our AI agent comes into play:

• One group discusses the map with fellow students.
• The other group conducts the interview with our Chatbotwho acts as a digital interviewer.

The chatbot is built in such a way that it takes on the role of a researcher and stimulating follow-up questions based on the contents of the network map and the answers given.