AI that drafts. A person who decides.
Every AI feature in EduCore does one of two things: draft something a person reviews, or answer a question from data your school already has. Nothing on this page writes to a student record, sends a message, or reaches a parent without a human approving it first.
Grounded in your data. Reviewed by a person. Never guessing.
01
AI never touches your database directly. It works from real records already pulled through your school's normal permissions — it can't see, or answer with, anything a given user couldn't already access.
02
Where AI drafts something — a comment, a newsletter, a message — it produces a first version, not a final one. Nothing reaches a parent or a student until a teacher or staff member reviews and sends it.
03
Not everything labeled "automation" is AI. A scheduling algorithm and a hand-weighted risk score are automation, not a language model — and this page says which is which.
Where EduCore actually uses a language model.
Four features, all built on the same rule: the model drafts, a person decides.
AI-assisted report card comments
For every learner, a short narrative comment drafted from that student's actual marks this term — teachers edit and approve it, it's never sent to a parent unreviewed.
Newsletter tone polish
Rewrites the tone of a term newsletter template to sound warmer, while keeping every merge placeholder and real figure exactly as written — staff still approve and send.
Fee-alert message polish
The same tone-polish step, applied to fee-reminder message templates before they go out to parents — a draft to review, not an auto-send.
Ask Educore AI
Understands a plain-language question well enough to match it to one of a fixed set of school questions, then answers from your school's real, permission-scoped data — never invented. Available to School Owner and Principal accounts.
Ask Educore AI
Answers a fixed set of school questions, from your school's real data
Rules-based work, done automatically.
These save real time too — they're just not a language model. Worth being precise about, since the difference actually matters for what you can trust each one to do.
Timetable auto-generation
Fills a school's weekly timetable automatically from each subject's periods-per-week requirement, respecting stream and teacher double-booking rules — a scheduling algorithm, not a language model.
WhatsApp parent bot
Answers simple, common questions like a child's fee balance or attendance straight over WhatsApp, matched by keyword rather than free-form understanding — anything it doesn't recognize is hands straight to a staff member.
At-risk student ranking
Ranks students by a hand-weighted score across attendance trend, exam trend, payment lateness, and discipline cases — a transparent formula school staff can reason about, not a trained model.
WhatsApp — parent bot
Keyword-matched, not free-form — hands off to staff when unsure
Everything on this page is live in EduCore right now.
Nothing above is a roadmap item or a future promise — it's what the product actually does today, for schools already using it.
See how it handles a real school day.
A short walkthrough, using questions like the ones your staff would actually ask.