A feedback report a professor opens more than once.

Most course feedback arrives as a number out of five, months late, with no indication of what to do about it. It is read once, resented, and filed. This is built to be the other thing.

Request a briefing → Read the platform brief

Theirs, and nobody else’s.

A report, not a score

From their own students, on their own course. What worked, what did not, and what to change, with the evidence for each.

It is theirs

An administrator does not see it.

The same shape every time

Every report has the same sections in the same order, reporting the same six areas. Nothing about the shape of the document depends on the data or on a model. Two professors in different departments get reports that can be laid side by side and read the same way.

It describes the figures. It does not produce them.

Every score, benchmark, percentile, chart, count and priority in the report is computed by code and can be reproduced by hand.

The AI writes four kinds of sentence and nothing else: the summary paragraph, the sentence naming a strength or a concern, the one line description of a theme in the written comments, and the wording of a suggested action.

It is handed finished statistics and asked to describe them. It cannot produce a number. It cannot change a ranking. It cannot reorder a priority.

And each section of the report states how it was written. If the model was unavailable when the report was built, the report says so, on the page, rather than presenting templated wording as analysis.

The limits are the point.

Not an appraisal input

The report is the professor’s. It is not sent to a head of department.

No colleague is named

A comparison against the department is shown as a number, never as a name, and is withheld entirely when the group is small enough that the top score would identify one colleague.

No demographic breakdowns

Not because they are switched off, but because that data is never collected, since collecting it would sit badly against UGC equity norms.

A minority view still reaches the teacher

A theme raised by very few students is still shown to the professor, and still suppressed everywhere else.

The same responses, compiled for each body that asks.

The same responses produce the compiled documents each body expects: AICTE 360 degree appraisal, NAAC key indicator 1.4.1, NAAC key indicator 2.7.1, and the NBA course exit survey.

Where a framework asks for something a student survey cannot answer, the document says so in place rather than leaving a blank. The NAAC student satisfaction survey, for instance, is administered by NAAC and not by us, and our output marks those items as not collected rather than pretending to them.

Results below a minimum number of responses are suppressed, and every document carries the name of whoever generated it.

One module of a platform the whole campus signs in to once.

What the institute gets

A record that was true when it was made, because it came out of people going about their work rather than out of a form somebody was asked to fill in.

Where the data sits

On our infrastructure in India, or on hardware inside your own campus. The AI that answers questions is a model we run ourselves, so nothing is sent to an external AI provider.

How to look further

The platform brief is the whole argument in one document, architecture and procurement included, with the limitations stated alongside the capabilities.