At Redmenta, we often get asked how our AI-based system actually works.
AI in education is everywhere, but not all AI is created equal. At Redmenta, we don’t just generate content; we engineer it for real learning.
When you open an AI Assistant on redmenta, you can specify: the curriculum, the subject, the grade level, and even upload your own learning materials. Based on these inputs, the AI Assistant generates content, usually starting with an outline.
But those “ideas” aren’t coming from a magical digital brain. What we call an “idea” is actually a carefully orchestrated output of prompt engineering, a series of detailed instructions that shape the final product.
Think of a prompt like a recipe. It tells the AI how to interpret the ingredients, how to prepare the final product, and what kind of structure to follow. For example, if a teacher requests a lesson on photosynthesis for 8th graders, the AI follows rules like:
When we build a new Assistant, the real work begins with writing instruction chains, and it’s not just a technical task, but also a pedagogical one. Here are a few core questions that shape every Assistant:
What is its pedagogical purpose?
Is it used for formative assessment, concept reinforcement, critical thinking, or inquiry-based learning?
Which instructional principles apply?
We align every assistant with Redmenta’s pedagogical pillars: Differentiation, Mastery-based progression, Cognitive scaffolding, Reflection and feedback, and Real-world relevance.
How should it interpret uploaded content?
Should it prioritize a teacher-uploaded document, or the national curriculum objective? What happens if they seem to differ?
How should learning tasks be structured?
Should tasks begin with retrieval, or build toward synthesis? Should there be creative, open-ended elements?
How much flexibility do teachers need?
Most assistants are designed to be adaptable. Teachers can keep, remove, or regenerate tasks, and steer the AI’s logic interactively.
AI cannot read your mind. It doesn’t “think” like a human. If you want high-quality, pedagogically sound results, the prompts have to be extremely precise. For instance, our Practice Task Assistant runs on over three full pages of prompt logic. It includes: task structure rules, alignment with the learning goals, language and tone models, options for interactivity, and response logic for errors and misunderstandings.
Every Assistant on Redmenta is guided by the following instructional design principles:
At Redmenta, we don’t believe in “black box AI” — systems that operate mysteriously, without transparency or teacher input. Instead, our AI Assistants are built on a foundation of transparent, teacher-controlled, and evidence-based design. Every assistant operates through carefully crafted prompts that follow clear pedagogical principles, allowing teachers to understand, customize, and guide the AI’s output.
