How schools across Bharat can adopt AI without buying hardware

For a decade, "technology in schools" in Bharat has meant a purchase order. Smart boards. Interactive flat-panel displays at a lakh or more each. Tablet trolleys. A server room. The pitch was always the same: buy the hardware, and modern learning follows.

The results have been uneven. Panels gather dust when teachers are not trained. Tablets get locked in cupboards. And the capital sits on the balance sheet whether or not it moved a single exam score. If you are a principal or trustee who has lived through one of these cycles, you are right to be skeptical the next time someone says "AI" and reaches for a quotation.

The good news: modern AI does not require any of it. Here is how a school can adopt AI as software, not capex - and start seeing value in weeks, not budget cycles.

Why the hardware model is the wrong model for AI

Hardware-led ed-tech tied your outcomes to the weakest link in the chain: the device in the room. If the panel was in Section A, Section B did not benefit. If the projector bulb died, the lesson died with it.

AI tutoring inverts this. The intelligence lives in the cloud and reaches the student through devices they - and your teachers - already own:

• Any web browser on a school computer, a teacher's laptop, or a shared machine in the staff room.
• Android phones, which are already in the vast majority of households across Bharat, for students studying at home at 10 pm.
• No installation, no lab, no server. Nothing to procure, rack, cool or insure.

The right question stops being "what do we need to buy?" and becomes "who do we want to give access to, and when?"

Works on the networks you actually have

A fair objection: "Our internet is not great." AI built for schools across Bharat has to assume that from the start, and good products do.

Pravachak streams the tutor's response token by token over ordinary HTTP, so even on a slow 4G connection a student sees the tutor "typing" in real time instead of staring at a blank screen. If a mobile connection drops mid-explanation, the session resumes from where it left off rather than starting over. And when the AI gateway is briefly unreachable, the student is handed a cached practice set from the same lesson instead of an error message.

We do not promise offline AI inference - that genuinely is a hardware problem. What we promise is that a normal school network and a normal phone are enough.

A capex-free adoption path, step by step

Here is the sequence we recommend, and the one our partner schools follow.

1. Start with one class, free

Do not roll out school-wide on day one. Pick a single class and a willing teacher, and run a free four-week pilot. Pravachak goes live for those students on real lessons - no hardware, no long contract, no risk. The goal of the pilot is one irreversible moment: a teacher who saves three hours on lesson planning in week one, or a measurable score lift in the pilot class by week four.

2. Let teachers feel the win before students do

The fastest path to adoption runs through your teachers, not around them. Before students ever open the tutor, a teacher can generate a board-aligned 5E lesson plan in about thirty seconds, produce a ten-question quiz from that plan, and assign it to the class. When a teacher gets an evening back in week one, you have a champion, not a mandate.

3. Onboard the way your school is structured

Adoption fails when a tool does not match how a school is organised. Pravachak onboards top-down: the school is set up first, then teacher rosters are bulk-imported from a CSV, then each teacher imports their own class. Students get an invite and a two-minute setup. A pilot milestone checklist tracks the first teacher login, the first lesson plan, the first quiz and the first tutor session so leadership can see momentum building in real time.

4. Roll out grade by grade, on outcome data

When the pilot produces its outcome anchor, expand - one grade, then the next - carrying the dashboard with you. Because seats are counted per student, you grow the licence as you grow adoption, not the other way round. There is no stranded hardware if a section does not take it up.

What it costs, and why it is not what you think

Priced per student per year and passed through a school's existing fee structure, AI tutoring is a rounding error against what families already spend on private coaching - and a fraction of a single smart-panel purchase. The cost lands on the "digital tools" line, not the capital budget, which is precisely why it does not require a board meeting to approve a pilot.

If you want to see the tiers for a school your size, our pricing page lays out four plans with per-student seats and a free pilot on every one.

The bottom line

You do not need to buy anything to start. AI adoption in 2026 is a software decision, a training decision and a change-management decision - not a procurement project. Start with one class, let a teacher feel the win, and let the outcome data make the case for going wider.

Ready to run a free pilot in one of your classes? Talk to our team and we will have you live in weeks - on the devices and network you already have.