JMIR Publications

ABSTRACT

Background:

Clinician educators may wish to develop small tools for educational record management but find it difficult to begin without programming experience. Generative artificial intelligence (GenAI) can support development, but educators must verify that proposed configurations preserve educational meaning.

Objective:

To present a practical workflow for clinician educators collaborating with GenAI to develop educational record management tools and critically review AI proposals.

Methods:

The tutorial draws on the development and redesign of a surgical residency training management system. A worked example using synthetic departmental education records illustrates cumulative and placement-specific review. An implementation appendix and an English-language Notion demonstrator support the exercise. Actual development principles are distinguished from specifications reconstructed for teaching.

Results:

The workflow comprises six steps: defining the educational question and scope, reviewing record structure, building a minimum implementation, testing exceptions, defining review actions and responsibilities, and extending the tool while preserving existing meanings. Each step connects clinician judgments, AI tasks, acceptance criteria, and a tangible output. The example addresses missing links, uncertain placement attribution, incorrect relationships, and activity-date criteria while preserving historical completion records. The demonstrator includes six practice pages and independently preserved intermediate states. Saved-view results and selected exception manipulations have been checked; complete visual inspection and independent reproduction by users without development experience remain unverified.

Conclusions:

The workflow connects educational requirements with implementation and review of AI-assisted tools. Its central principle is to preserve completion evidence while distinguishing placement attribution and retrieval criteria. Independent reproducibility and educational effectiveness require further evaluation.