Weekly test reports piled up unread — an AI read them for him

🎓 Rakesh Yadav, 19, Khagaria district, Bihar. ₹2.4 lakh borrowed for JEE coaching in Patna (one and a half times his father's annual salary as a primary school teacher). Mock percentile: 84. Needed 96 for JEE Advanced. Batch C-7 had 184 students. Each got a 30-page test report every week. His reports sat unread in a corner. An AI changed that.

Weekly test reports piled up unread — an AI read them for him

🚨 The problem

Coaching factories print data, not clarity. Thirty pages of chapter-wise breakdowns, percentile trends, comparison charts — data that tells the story of exactly which topics are bottlenecks. But in a batch of 184, nobody reads a 17-year-old's report. The first page (rank, percentile) matters. Pages 8–14 (where the real weaknesses are) go unseen. Rakesh had all the data he needed to fix his ceiling. He just didn't know how to read it.

🚀 How GabFORGE helped

A cousin (software engineer, Bengaluru) and an AI system did three things:

  • 🔍 Read the reports systematically. Instead of flipping to percentile, the cousin read pages 8–14 and found it: rotational mechanics 35%, electromagnetic induction 32%. Everything else was 75–85%. Two chapters were the wall.
  • 💬 Installed an AI on his phone. Taught Rakesh three questions to ask every Thursday: "What's the worst chapter this week? Give me 5 key formulas for it. How much improvement this month?" AI read PDF, answered in Hindi.
  • 📞 Kept his mother informed. Every Saturday, AI sent Rakesh's mother a voice message in Hindi: "Rotational mechanics 35% last week, 42% this week, target 50–60% by week 3." Truth-telling, backed by data. No more "it's going well, mama."

January to May: rotational mechanics 35% → 71%, EM induction 32% → 68%, percentile 84 → 96 ✅. JEE Advanced May 2026: AIR 4,127, NIT Patna Mechanical Engineering. The factory produced a report; a cousin read it; an AI kept his mother in the loop; four months of focus on two chapters moved the ceiling.

🇮🇳 Why this matters

Coaching factories operate at scale. They generate data but not insight. A student drowning in information (30-page reports weekly, 184 classmates, no individual attention) needs someone to read the data and translate it: "You're strong at X, weak at Y, here's what to focus on." An AI reading PDFs in Hindi did that. A mother receiving weekly progress reports stopped worrying about money spent and started asking about physics gaps. A family that had borrowed ₹2.4 lakh had someone actually looking at whether it was money spent well. NIT Patna is not an IIT, but for a boy from Khagaria who couldn't afford to make mistakes, it was the outcome the investment was designed to buy.

Read the full story →

The long version has the unread reports in a corner, the cousin's technical eye reading pages 8–14, the three Hindi questions, his mother's Sunday calls, the truth-telling moment, and the AIR 4,127 result.