When we sat down to design the Performance page that parents see for their child, the easy version of this feature already existed in a dozen other products: give parents everything. Every quiz attempt, every wrong answer, every question a student typed into an AI tutor at eleven at night. It reads well in a sales conversation. It does not read well from where a fourteen-year-old sits, trying to work out whether it is safe to admit they are lost.
Two views of the same child
A teacher and a parent are both trying to help the same student, but they are doing different jobs, and they need different information to do them. A teacher needs to know, in detail, where a student is getting stuck: which step in a quadratic equation trips them up, which questions they asked the AI tutor before giving up, whether a struggle looks like a bad week or a pattern worth flagging. That is fine-grained, professional work, and it requires fine-grained data. A parent's job is different. A parent needs to know whether their child is engaged, whether they are keeping up, whether something has changed recently, and whether a teacher has something to say about it. That does not require a transcript. It requires a clear picture.
So we built two views on purpose, not by accident of what was easiest to ship. A teacher's dashboard carries the detail: per-attempt telemetry from the AI tutor, the specific concepts a student struggled with, escalation flags when something looks like it needs attention, and the actual questions asked during a tutoring session. A parent's Performance page carries the outcome: subjects explored, number of sessions, an engagement level, an overall trend, quiz and puzzle results, and the remarks a teacher has chosen to share. Same child, same underlying activity, two different levels of resolution, and that difference is the whole point.
What the Performance page actually shows
In practice, a parent opens the app and sees each of their children listed, and for each one a snapshot: how they are doing across subjects, how consistently they are attending class, how recently they have engaged with schoolwork on the platform, and any remarks a teacher has left. From that page, a parent can generate an actual report card, printable, formatted the way a report card should be, pulling together quiz and puzzle performance, attendance, and teacher feedback into one document. It is meant to answer the question a parent actually has at the end of a term: how is my child doing, and is there anything I should know.
It is not meant to answer a different question, which is: what exactly did my child say when they were confused. That question, we decided, belongs somewhere else.
The line we drew on purpose
This is the part of the design that took the most conversation, because it is genuinely a trade-off and not a simple engineering decision. Full transparency sounds like the safer default. If a parent can see everything, nothing can hide, and in theory a struggling student gets noticed sooner. But that logic assumes a student's behaviour does not change once they know they are being watched, and for a teenager working through a confusing maths problem or an essay they are not confident about, that assumption tends to be wrong. If a student believes a parent will read the exact question they typed into a tutor at a moment of confusion, some of them stop typing the honest question. They ask something safer, or they stop asking at all, and the tool that was supposed to help them learn quietly stops working.
A parent doesn't get a transcript of everything their teenager asked an AI tutor in a moment of confusion. That level of detail stays with the teacher, who has professional responsibility for how to use it.
— PDP Shikshya, on the design of the Performance page
So the specific struggles, the specific questions, the moment-to-moment record of a student figuring something out, stays with the teacher. Not because parents cannot be trusted with it, but because the value of that information depends on someone with professional training and ongoing relationship with the student interpreting it, deciding what matters, and deciding what to do about it. A parent gets the outcome. A teacher gets the working.
A weekly digest, without the name attached
Alongside the Performance page, a parent can request a short weekly digest, an AI-generated summary of what their child has been doing that week: subjects covered, general engagement, anything notable at a high level. We built this the same way we built the rest of the feature, with the boundary in mind from the start rather than added afterward. When the digest is generated, no identifying information about the child is sent to the AI model doing the summarising. The summary is built from the underlying activity data and assembled without the model ever seeing a name, a school, or anything that would tie the pattern back to a specific student. The parent reading it knows exactly whose week it describes, because it is their child's data; the system generating the language does not need to know, and so it doesn't.
This sounds like a small technical detail, and mechanically it is, but it reflects a broader habit we try to keep: minimise what any AI system sees to exactly what it needs to do its job, and no further, even when the convenient path would send more.
A conversation happening well beyond one app
This is not a decision we arrived at in isolation. Right now, in school systems and education ministries in a number of countries, there is an active and genuinely unresolved conversation about how much of a minor's digital learning activity should be visible to a parent, versus how much needs to stay a private space where a student can be honestly confused without an audience. Some of that conversation happens under existing student-privacy regulation, in various forms across different countries, and some of it is happening informally, inside schools and among parents and teachers, as AI tutoring tools become common enough that the question can no longer be postponed. There is no settled global answer, and we do not think there should be one imposed from outside a school community. But the tension itself, transparency for parents against psychological safety for the student, is real, well recognised, and worth naming plainly rather than papering over with a feature list.
We landed on a specific answer for a specific reason: a teenager learns better when they believe asking a wrong question, or a stupid one, or the same one twice, does not become a topic at the dinner table. We would rather protect that belief and give parents a strong, honest, useful picture of how their child is actually doing, than give parents a raw feed and quietly lose the thing that made the tutor worth having in the first place.
What this means at home
In practical terms, this means a parent using PDP Shikshya can log in at any point in the term, see how each of their children is doing across subjects, check attendance, read what a teacher has written, and print a proper report card when it is time. It means the weekly digest can become a five-minute habit rather than a chore, because it is short and it is written for a parent, not for an audit. It also means that when a parent has a genuine concern, the conversation that follows is the right one: with the teacher, who has the full picture and the professional standing to act on it, rather than with the child, over a transcript that was never meant to be read by anyone but the two of them and, briefly, an AI trying to help.
None of this is presented as a finished, permanent answer. As families, teachers, and students use the Performance page and the digest in practice, we expect to keep adjusting where the detail sits and how much of it surfaces where. What will not change is the underlying principle: the level of visibility a parent gets should be chosen deliberately, for the child's benefit, not simply set to maximum because maximum was available.