evidoria

← Back to browse

Good practice Imported

I-Stem — AI Document Accessibility Platform Reaching 31,000+ Blind and Low-Vision Users in India

India · New Delhi · See the India profile · See the New Delhi profile

Evidence: Observational / pre–post Top 31% 60/100 · Ask Evidence Copilot about this practice

I-Stem's AI platform converts STEM documents, exams and textbooks into screen-reader-accessible formats at 92% accuracy. Used by IIT Delhi, Ashoka University and Telangana's K-12 system, reaching 31,000+ users, ~9,000 of them women.

31,000 users
Total users reached (cumulative to date)
9,000 users
Female users reached (cumulative to date)
92 %
Document conversion accuracy (n/a)
I-Stem — AI Document Accessibility Platform Reaching 31,000+ Blind and Low-Vision Users in India

Details

Maturity
Scaling
Promoter
I-Stem (Inclusive STEM)
Period
2018–present (AI mobile platform since 2021)
Keywords
assistive technology, higher education, K-12 education, nonprofit

Context

I-Stem, founded in 2018 in New Delhi by blind engineers Kartik Sawhney and Shakul Sonker, built an AI and computer-vision platform that converts inaccessible STEM documents, exams, and textbooks — including tables, mathematical notation, and diagrams — into screen-reader-compatible formats for blind and low-vision users in India.

Objectives

The platform aims to close the accessibility gap in STEM education materials, extending reach beyond urban centres and supporting exams and textbooks in multiple languages.

Activities

A mobile app launched in 2021 extended reach into semi-urban and rural areas, and a 2024 grant added voice-interface access via WhatsApp and IVR; the platform has been deployed at IIT Delhi, Ashoka University, and across Telangana's state K-12 system for English and Telugu textbooks.

Results

I-Stem reports serving 31,000+ users, including roughly 9,000 women, with a 92% accuracy rate for document conversion; these are self-reported adoption and technical-accuracy figures, as no independent third-party study of learning outcomes has been published.

Implementation

Indicative cost
Medium (€50k–€500k) — Funded via GSMA grants (2021, 2024) and a 2025 Meta/Nudge Institute grant; no public unit-cost figures published.
Time to results
Long (> 3 years)
Staffing & skills
I-Stem (founders Kartik Sawhney, Shakul Sonker)

Conditions for success

  • Mobile app to reach semi-urban/rural users
  • Voice-interface (WhatsApp/IVR) access for low-literacy or low-connectivity users
  • Partnerships with universities and state education systems for institutional deployment

Where it fits

Governance type
NGO/social enterprise
Scale
national (India)
Income level
lower-middle-income

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful