🧠 AI Ed Wiki

📄 Research Article

Prompt Coach: An Empirical Evaluation of an Agentic Tutor for Learning Prompt Engineering in Software Development

● high · 2026-07-08agentic-ai, cs-education, feedback-loop, formative-assessment, intelligent-tutoring, llm, professional-training, scaffolding, ai-literacy

Prompt engineering is a critical yet undertaught skill for software developers, poorly served by traditional instruction because of its evolving, interactive, context-dependent nature. The authors introduce Prompt Coach (PC), an agentic tutor embedded in-flow within a developer's IDE that teaches prompt crafting through Socratic guidance. PC scores prompt quality across multiple dimensions and surfaces targeted questions that guide self-correction, grounded in the developer's codebase and the target LLM's behavior. An early empirical study with 15 professional developers showed statistically significant improvements after a single 60-minute session, with the largest gains on dimensions developers commonly overlook, plus strong trust and adoption readiness.

  • In-flow Socratic tutoring: PC delivers Scaffolding and Feedback Loop support directly inside the IDE, contextualized to the developer's code and the target model.
  • Empirical gains: Statistically significant prompt-quality improvement after one 60-minute session, aligning with findings in Prompt Based Programming Lesson about prompt fluency as a learnable skill.
  • Self-correction emphasis: Targeted Socratic questions steer developers to repair their own prompts rather than receive answers, consistent with Formative Assessment practice.
  • Professional context: Extends CS Education and Professional Training beyond novices to working developers.
  • Trust and adoption: Participants reported high trust and unanimous belief PC improved their skills, contributing to AI Literacy for practitioners.

    • Connected Concepts

  • Scaffolding
  • Feedback Loop
  • Formative Assessment
  • CS Education
  • Professional Training
  • AI Literacy

    • Connected Articles

  • Prompt Based Programming Lesson

    • Citation

    Mehra, R., Singi, K., Kaulgud, V., Sharma, V. S., & Choudhury, S. G. (2026). Prompt Coach: An Empirical Evaluation of an Agentic Tutor for Learning Prompt Engineering in Software Development. arXiv:2607.06074.