Learnr

Making exam prep a habit, not a decision

Pay first, learn later

A large fee was charged before aspirants could feel whether the product was worth it, the first friction the whole product had to solve.

Long lessons, small concepts

A single idea sat buried inside a long lesson, costing hours to learn something that needed minutes.

The quiz came too late

Practice arrived after a stretch of videos, so concepts had already faded by the time aspirants could test themselves.

Nowhere to ask, in the moment

A doubt mid-lesson had no path to resolution inside the app, so aspirants left to search, and once gone, many didn't return.

A homepage with stories, trending and subject-wise entry

Three discovery patterns, all required on one home surface, the design problem was hierarchy: which one greets a first-time aspirant vs. a returning one, without the homepage becoming a bazaar.

Short videos, organised into modules

Bite-sized content was mandated, but so was structure, every video had to belong to a module, and the experience had to stay engaging, not become a folder of clips.

A quiz on every video

Practice couldn't be a separate mode or an end-of-chapter test. Each video needed its own quiz, the open question was timing, weight, and how to keep it from feeling like homework.

An AI chat assistant

Aspirants had to be able to chat with an AI bot. Whether it lived in its own tab or inside the learning flow was ours to figure out, and became one of the highest-leverage decisions in the project.

Hinglish as the platform language

The product had to be built vernacular-core for Tier 3–4 aspirants from day one, which shaped content, UI copy, and Riya's responses before a single screen existed.

A vertical scroll, YouTube Shorts–style

Video, description, play/pause, like, and share sit exactly where a Shorts or Reels viewer already expects them. Nothing about consuming a Learnr video needed to be learned, the only new thing was the content underneath it.

Stories, borrowed from Instagram

Stories are how aspirants already expect to hear “what's new” in short form: a quick, skippable preview rather than a full session. Borrowing that pattern let new modules get noticed without competing for feed real estate.

Home divided by category

Home is split into rows by category the way a streaming home screen is “trending, subjects, history” because aspirants already know how to scan a shelf of options and pick one, rather than parse a single continuous feed.

Riya, placed inside the video and the quiz

A doubt happens the moment a concept doesn't land, mid-video or mid-question, not later, and not in a separate tab. Riya sits exactly where the doubt appears.

Structured modules

Each module groups its chapter's videos and quiz questions with the progress and sequence. A visible content and attempted count tracks exactly how far into the chapter an aspirant has gotten, giving bite-sized content a sense of “next” without forcing anyone through a rigid, locked path.

V1 – Course-first, manual, generic

We bet that structuring everything around a chosen exam, course pitch, batch selection, then content, was the fastest way to make Learnr feel credible. Personalisation and AI support didn't exist yet.

V2 – Personalised, AI enters, quiz moves into the feed

We bet that removing manual batch selection and introducing an AI doubt-solver would deepen engagement, and that practice belonged inside the scroll, not off in a separate tab.