Beyond Automation: Prompt Design and Trustworthiness in AI-Assisted Inductive Coding

Abstract

This study introduces a structured prompt framework for AI-assisted inductive qualitative data analysis, demonstrating how carefully designed prompts can guide each stage of the coding process while preserving methodological rigor. A distinctive contribution of the study is that it provides a set of empirically tested, ready-to-use prompts that researchers can directly adapt to their own qualitative studies, offering a practical and replicable workflow rather than merely describing general principles of AI-assisted analysis. The prompts systematically support inductive code generation, refinement of code definitions, deductive coding, data summarization, visualization, and reliability assessment. Rather than treating artificial intelligence as an autonomous analyst, the proposed framework conceptualizes qualitative analysis as an iterative and reflexive collaboration between the researcher and AI, in which interpretive authority remains with the researcher. Grounded in the principles of credibility, dependability, and confirmability, the framework emphasizes the continuous verification of AI-generated outputs at every stage of the analysis to ensure fidelity to the raw data and prevent the uncritical acceptance of automated interpretations. The study argues that the trustworthiness of AI-assisted qualitative analysis depends not on automation itself, but on the quality of prompt design and the researcher’s systematic oversight throughout the analytical process. By providing validated prompt templates, the study offers a transparent, practical, and reproducible methodological framework for rigorous AI-assisted qualitative research across diverse contexts.