How to run a small teacher trial that tells you something
For coordinators and principals: a five-step trial plan built around your teachers' topics, your phones and replies that teachers check, with a clear decision at the end.
6 October 2026 · 7 min read · Product information as of 6 October 2026
The situation
A principal hears about a teacher assistant app from a coordinator. The questions arrive in the right order: does it cover our topics, does it work in Marathi, will it run on the phones our teachers actually own, and can we trust what it says?
The wrong way to answer them is to roll it out to every teacher and see what happens. The right way is a small trial with a decision at the end.
Before you start: what the current version is
VIDYA Copilot is an early Android app, not yet on the Play Store. It answers one school-topic question at a time, by typing or voice, in Hindi, Marathi or English. Replies are written on the phone from what the AI has learned, carry a warning that they came from memory, and need checking. The marking tool, handwritten-paper reading, textbook grounding and class study reports are not available in the regular app.
If your trial needs any of those four, wait. If it needs a reliable first response to teacher questions, read on.
The five steps
The company’s operations view, which COO Vaibhav Jain is responsible for, is that a trial should be small enough to finish and specific enough to decide from. The suggested steps below follow the supplied product documentation.
1. Choose a small group of teachers and a few topics. Three to five teachers, the subjects they actually teach, and chapters they know well. Decide in one sentence what you want to learn: for example, “Are replies in Marathi clear enough for Class 7 science preparation?”
2. Check how to get the app and which phones to use. Access is arranged directly. Confirm the app version (the regular build or the new design being tested), the setup steps and the phones available. Use the phones teachers already own, not a demonstration device. Results from one test phone may be different from another.
3. Review actual replies. Ask teachers to compare each reply with their books and notes. Three questions per reply: Are the facts right? Is the explanation clear? Is the language suitable? Keep a simple record, one row per question.
4. Check everyday controls. Try voice with the internet off. Try quiet mode in the staffroom. Stop a spoken reply halfway. Ask two unrelated questions in a row to confirm that each is treated as new. These are the controls teachers will use every day, so test them in the places they will use them.
5. Decide from the findings. Continue, narrow the scope to fewer topics or teachers, or wait for improvements. Any of the three is a good outcome from a trial. The bad outcome is a trial with no decision.
What the record should look like
| Teacher | Topic | Language | Phone | Facts right? | Clear? | Language OK? |
|---|---|---|---|---|---|---|
| Teacher 1 | Fractions, Class 6 | Marathi | model __ | |||
| Teacher 2 | Photosynthesis, Class 7 | Hindi | model __ |
Fill it in by hand. Twenty rows will tell you more than a presentation.
Privacy during the trial
The app writes replies on the phone and has no internet permission. The phone’s own listening and speaking tools may use the internet and have their own settings. The teacher app does not connect to a school information page or to the student app’s school system. Do not send student answer papers; testing of marking tools needs a separate agreement and checks on how the papers will be used.
What a trial cannot tell you
A trial of this size shows whether the app is useful for your teachers on your phones. It does not show time savings, learning gains or exam results, and the supplied documentation claims none of those. Decide on what you saw.
Takeaway
Small group, known topics, real phones, teachers checking replies, one decision. The for schools page has the same steps with the questions to put to us. When you are ready, talk about a school trial.
Facts in this article come from the supplied product documentation dated 6 October 2026. No customer quotes, adoption figures or results beyond the reported tests are claimed. Leadership views are presented as company perspective, not as direct quotations.
