Table 2. Prompting strategy for AI-assisted code application.

Steps	Action	Description
1	Codebook optimization	The revised codebook was uploaded to the AI model. The model was then asked to optimize it for AI use of qualitative coding.
2	Prompt optimization	The model was presented with the draft prompt and instructed to “Optimize this prompt for qualitative analysis by AI.”
3	Data ingestion	A new chat session was initiated where the 10% dataset was uploaded, as well as the AI-optimized codebooks.
4	Coding	The article’s text was given to the AI model with the optimized prompt to code the sentences. After the coding for that article finished, it was asked to refine the coding. This process was repeated for each article.
5	Extraction	The codes were transferred manually to ATLAS.ti (ATLAS.ti Scientific Software Development GmbH) to create a fully coded file with AI-assisted coding. To ensure data integrity during the manual transfer of AI-generated codes into ATLAS.ti, a systematic cross-verification process was performed against the raw AI outputs.