Rethinking qualitative analysis with AI

When the ground of qualitative research is shifting

Large Language Models are not just another tool. They signal a shift in how knowledge work, interpretation, and sense-making will be organised. Ignoring this does not protect qualitative rigor — it risks making it irrelevant.

Going back to the roots

In their grounded theory approach, Strauss and Corbin emphasized the importance of immersing oneself in the data and developing concepts through direct interaction and reflection. Although they referred to this process as coding, it was always much more than simply jotting down words in the margins of a transcript.

Looking at the examples used by Strauss (1987), his notes also included analytical thoughts, sometimes quotes from the original document, references, instructions for theoretical sampling, and ideas on the elements of the coding paradigm.

The original idea was that researchers develop concepts and ultimately theories through constant comparison and reflective thinking. Direct engagement with the data was the focus, supported by memo writing to gain deeper insights.

By using AI, we can now revisit this original methodology. AI enables us to interact directly with the data and supports reflective thinking without the need for the technical step of extensive coding. In this way, the research remains closer to the original idea of how to conduct an analysis that is grounded in the data.

Strauss, Anselm L. 1987. Qualitative Analysis for Social Scientists. Cambridge: Cambridge University Press.

Questions, context and interpretation

Researchers develop concepts through constant comparison and reflective thinking. Dialogue with AI can support this direct engagement with the material, with the researcher bringing contextual knowledge, methodological judgement and responsibility for interpretation.