Using AI in Qualitative Analysis: Where It Helps, and Where the Researcher Still Has to Drive

The question of how, and how far, to use artificial intelligence in research is moving quickly from theory into everyday practice. Generative AI is already a capable assistant in qualitative analysis, and the academic conversation about its responsible use is maturing alongside it. A forthcoming special issue of The Service Industries Journal takes up exactly this question, and five open-access papers from that conversation are collected at the end of this piece.

The emerging consensus is not that AI should be avoided, nor that it can be handed the work wholesale. It is something more disciplined: AI belongs in the process as a tool that helps produce higher-quality results, not as a substitute for the researcher's own judgement.

Where AI is already useful

In qualitative work, the most immediate gains show up in coding. AI can assist with both open and axial coding of text, returning the number of items analysed, working definitions for each theme, and supporting excerpts drawn from the source material. It can hold to a pre-established coding frame or leave the coding open, and it can associate emerging themes with prior research rather than treating them in isolation.

In teaching, the same capability is proving valuable. Used in a research methods setting, AI helps students learn both quantitative and qualitative analysis by giving them a fast, interactive way to see how coding and interpretation actually work. In one such use, students coded online reviews, social media posts, and Reddit threads drawn from various sources, and surfaced genuinely insightful themes in the process. One of the more useful features is that AI handles images and video nearly as readily as text, which opens multimodal material to forms of analysis that were previously slow and labour-intensive.

The working rules

Using AI responsibly in qualitative work

The tool changes what is possible; it does not change who is responsible for the result.

1 Keep a human driving the process AI assists with coding and interpretation, but the analytical decisions, and accountability for them, stay with the researcher.
2 Proof-check every output As with any process, AI output has to be verified against the source material rather than accepted at face value.
3 Disclose how and how much The methods section should state precisely how AI was used, including for transcription, and the extent of that use.
4 Treat it as a tool, not a substitute AI should be employed to produce higher-quality results, not to replace the researcher's intellectual work.
5 Cite it like any other tool Where AI is used, it warrants acknowledgement in the same way any other analytical instrument would.

One genuine cost sits outside the methodology: the environmental footprint of AI energy use is a real consideration in weighing when to reach for it.

The disclosure question

The single most important habit is transparency. In the methods section, it is advisable to disclose exactly how AI was used, including for transcription, and the extent of its involvement. The interaction between researchers and AI is a new phenomenon, and the norms around it are still forming. In a few years there will likely be clearer, community-accepted guidelines for its appropriate use at different stages of a project. For now, the best available approach is to be completely transparent with readers, and to keep the work an original reflection of the researcher's own intellectual effort. Those qualities are the enduring ones, whatever the tooling around them turns out to be.

Five open-access papers to read

The following papers, drawn from the conversation feeding into a forthcoming special issue of The Service Industries Journal, are all open access.

Liu & Benckendorff (2026): "'Simulate' to 'stimulate': generative AI for causal and anticipatory service research." The Service Industries Journal. tandfonline.com →
Ali, Kasturiratne, Ameer & Bhaskar (2026): "Generative AI in digital engagement: a quasi-experimental study of tourist sentiment." The Service Industries Journal. tandfonline.com →
Ivanov (2025): "Responsible use of AI in social science research." The Service Industries Journal. tandfonline.com →
Do (2025): "Generative AI in service research: promise or peril?" The Service Industries Journal. tandfonline.com →
Imschloss, Sarstedt, Adler & Cheah (2025): "Using LLMs in sensory service research: initial insights and perspectives." The Service Industries Journal. tandfonline.com →