General practitioners (GPs) in primary care—particularly in low- and middle-income countries (LMICs)—frequently encounter meaningful clinical problems but lack the methodological training to formalize them into research questions. This paper reports a five-stage, AI-assisted approach that embeds established frameworks—including the JBI Population–Concept–Context framework, scoping review methodology, and evidence-based questioning paradigms—into nine standardized AI prompts, guiding GPs through: practice observation and value assessment; information extraction and evidence-based transformation; literature search and knowledge summarization; research question prototype construction; and methodology selection and feasibility assessment. Built on human-AI collaboration with human primacy, the approach requires no prior methodological training. It was piloted through the Shanghai General Practice Research Network (SGPRN) and is most applicable in low evidence-density primary care settings.