Built for the classroom most software ignores.
Thirty-eight children, one teacher, a shared phone at home and a connection that drops. That is the room this was designed in.
Why this exists
A teacher in an Indian school spends roughly four hours a week on work that is not teaching: making a worksheet, setting a test, marking it, and writing something honest for a parent. None of that is the job they trained for, and all of it is the first thing to get rushed when there are thirty-eight children in the room.
Every product aimed at this problem is either an operations system — fees, attendance, timetables — or a content library. A school already has both. What neither of them does is the four hours.
Sahayak is built around that gap, and around a second finding that shaped it more than anything else: in the largest study of an AI tutor in schools, the biggest problem was not that the tutor gave bad answers. It was that students did not open it. A chat box is a thing you have to remember to go to. So the help here appears where a student already is — the chapter they just opened, the question they got wrong, the homework due tomorrow — and not in a box they have to think of first.
What we will not build
- • A public profile for a child. There is no social graph here and there will not be one. A twelve-year-old with a bio and a visibility setting is a safeguarding incident waiting for a date.
- • An AI that writes into a student’s record. It drafts; a teacher commits. The one deliberate exception is machine marking, and we explain that one rather than hide it.
- • Engagement mechanics.No streaks, no leaderboards between children, no notification designed to pull a child back in. Attendance is the school’s job, not a growth metric.
- • Selling anybody’s data.Student data is the school’s. It trains no model and it reaches no advertiser.
Talk to us
We are a small team in Pune, Maharashtra. A demo is a conversation with someone who built the thing, not a scripted call.