Learning Differently / Dora
Education guidance, not diagnosis. An education technology platform with Dora, an AI-assisted guide that helps parents, teachers, and learners understand learning, behaviour, and attention patterns, then decide what to try next.
Overview
Dora sits inside an education platform designed for different learning journeys. The product’s job is to make observations usable: what has been noticed, what it might mean in educational terms, and what a parent or teacher can do this week. Clinical diagnosis stays outside the product boundary.
The Problem
Parents and teachers often notice that a child is struggling and then face a wall of generic advice, late assessments, or medical language they did not ask for. They need a structured way to observe patterns and choose a next step, without a product pretending to diagnose.
The Approach
Treat the conversation as a structured intake, not an open-ended chatbot. Collect context, reflect patterns, and recommend educational next steps with clear limits. Language is careful: screening and guidance, never a label presented as fact.
System / Product Design
The system separates observation, interpretation, and action. Observation is human. Interpretation is assisted and reversible. Action is a short list of educational steps, resources, or conversations, with an explicit note that medical or clinical assessment is a different path when it is warranted.
Technology
Next.js · TypeScript · AI · Django
Challenges
Trust and harm
A confident wrong answer about a child is worse than silence. The product has to show uncertainty and refuse diagnosis.
Different audiences
Parents, teachers, and learners do not need the same words or the same next step.
Key Decisions
Guidance over diagnosis
The product boundary is educational. That constraint is a feature: it keeps the AI useful and honest.
Workflow, not a chat toy
A repeatable screening and guidance path is more valuable than an unbounded conversation.
Outcome
The platform is in development. The current work is the guidance workflow, language boundaries, and the first supported learning journeys.
What I Learned
- In education and wellbeing, what the product refuses to claim is as important as what it offers.
- AI is useful here as a structured mirror, not as an authority.