We began the project by creating mockups of the FLASHCARDS-FR and françAIs web/app interfaces, accompanied by an initial concept document that helped define the project's scope, objectives, and overall structure.
Using these materials as a foundation, we developed a structured series of prompts for ChatGPT and Claude AI. These prompts were used sequentially to generate the Product Outline, Developer Specifications, LLM Coding Prompts, and AI API Prompts required to support the development process.
As expected, the journey was highly iterative. Multiple rounds of refinement were necessary to clarify requirements, expand feature definitions, correct omissions, and address inconsistencies within the original concept. Each iteration helped improve the quality, completeness, and precision of the project documentation.
Throughout this learning and development process, we produced six versions of FLASHCARDS-FR, four françAIs prototypes containing three lessons each, and a more advanced version of françAIs featuring thirteen lessons.
In our next post, we will explore in greater detail, illustrating the process, challenges, lessons learned, and key outcomes that emerged as we moved from planning and specification toward implementation.
