India
Education
Implementing Organisation
I-Stem (Inclusive Science, Technology, Engineering and Mathematics Foundation)
India, Delhi, New Delhi
Implementing Point of Contact
Kartik Sawhney
Founder and Director
Contributor of the Impact Story
Carnegie India
Year of implementation
2025
Problem statement
Persons with disabilities face compounding barriers in accessing education, employment, and public services due to inaccessible digital content, lack of relevant career guidance, and exclusionary online systems. Over 96 percent of digital content is inaccessible, and most career guidance platforms do not account for disability-specific pathways, accommodations, or lived experiences. As a result, many individuals are unable to discover viable careers, complete learning independently, or perform essential tasks such as applying for jobs or government schemes. Traditional solutions rely heavily on human intermediaries, are not scalable, and often reinforce dependency rather than autonomy. Additionally, many users face low digital literacy, language barriers, or lack consistent internet access, making mainstream AI tools unusable. There is a need for a scalable, multilingual, multimodal AI system that is accessible via voice and messaging, grounded in disability-specific data, and capable of supporting learning, decision-making, and task completion independently. This use case addresses these challenges by leveraging open-source AI models to deliver inclusive career discovery, adaptive tutoring, and task automation through accessible channels, enabling users with disabilities to access opportunities with dignity and independence.
Impact story details
I-Stem is a nonprofit organization founded to address systemic barriers faced by persons with disabilities in accessing education, employment, and digital public infrastructure. India has over 26 million people with disabilities, with unemployment rates exceeding 70 percent, largely due to inaccessible digital content, lack of career information, and exclusionary systems. I-Stem was founded by persons with lived experience of disability to ensure that solutions are grounded in real user needs and dignity. I-Stem builds and deploys AI-powered accessibility infrastructure through its NClude platform, which operates across web, WhatsApp, and IVR channels to reach users regardless of digital literacy or device access. The platform focuses on three core areas: 1. career discovery through lived-experience-based guidance, inclusive learning through adaptive tutoring, and task automation to enable independent navigation of inaccessible digital systems such as job portals, government schemes, and application forms. 2. Since inception, I-Stem has partnered with government ministries, state governments, UN agencies, and over 70 nonprofits. It has converted more than 2.3 million pages of inaccessible content into accessible formats and currently serves over 50,000 users with disabilities, including rural users through IVR-based access. The organization combines open-source AI models, including Meta’s Llama, with human-centered design and responsible AI principles to ensure equity, privacy, and transparency. 3. I-Stem’s work aims not only to provide access, but to create pathways to economic participation, independence, and representation by enabling people with disabilities to learn, apply, and succeed on equal footing.
AI Technology Used
Key Outcomes
Economic Value Creation, Access
Reach, Inclusion
Equity, Knowledge
Skills Impact, User Experience
Satisfaction
Narrative Outcome
Persons with disabilities face compounding barriers in education, employment, and public services due to inaccessible digital content and exclusionary systems. I-Stem provides AI-powered career discovery, accessible learning, and task automation. Over 64,000 users with disabilities have accessed the platform. Task completion accuracy reaches 92 percent. User confidence in career selection has increased significantly. For individuals navigating systems not designed with accessibility in mind, the platform provides pathways that would otherwise be blocked.
Impact Metrics
Total Number of Users with disabilities accessing the AI-powered platform
Baseline Value
Not Applicable Users
Post-Implementation
More than 64,000 users Users
Rate of accuracy of task completion on the platform
Baseline Value
Not applicable Percentage
Post-Implementation
92 % accuracy of task completion on the platform
Confidence in career selection done by the platform
Baseline Value
3.3 user rating
Post-Implementation
4.65 user rating
Implementation Context
India
Persons with disabilities including blind and low vision users, mobility-impaired users, women with disabilities, rural and underserved populations, and youth and job-seekers
Key Partnerships
Union Ministry of Social Justice and Empowerment, Government of Uttarakhand, UNICEF, and over 70 disability-focused NGOs
Replicability & Adaptation
1. The solution can be adapted to other geographies by localizing language models, integrating country-specific datasets (jobs, schemes), and aligning with local disability policies. 2. The use of open-source models like Llama enables low-cost replication while maintaining data sovereignty.
Supporting Materials
* The data presented is self-reported by the respective organisations. Readers should consult the original sources for further details.