How these 5 organisations are using AI for social good

We are witnessing a new era, powered by Artificial Intelligence (AI), that is transforming lives across every domain.

From revolutionising education, making healthcare accessible, transforming data-led governance, to making our cities smarter, AI is shaping India’s progress on all fronts.

According to government data from last year, over Rs 10,300 crore has been allocated over five years for IndiaAI Mission with 38,000 GPUs deployed. Six million people are employed in the tech and AI ecosystem, and AI could add $1.7 trillion to the country’s economy by 2035.

At the heart of this revolution are several startups and organisations that are using AI for social good. They are focused on innovation that would uplift communities, bridge systemic gaps, and make technology accessible to those who have historically been excluded from its benefits.

Here are some startups/innovations using AI for social good.

Shishu Maapan - Wadhwani AI

Focused on AI for social impact, Wadhwani AI sees AI as a tool to drive large-scale, equitable, and human-centred impact. One of its innovations is the Shishu Maapan tool.

Under the country’s Home-Based Newborn Care (HBNC) programme, ASHAs visit newborns on the seventh, 21 and 42 days after birth to monitor their growth. The traditional scales used for this purpose proved challenging as they were impractical to carry around.

The Shishu Maapan AI tool can be integrated into a state’s existing ASHA application or accessed through the standalone Shishu Maapan app, which hosts the AI model.

Using a smartphone, a frontline worker records a short video of a newborn placed unclothed on a flat surface beside a wooden ruler. The AI processes the video in real time to extract key anthropometric measurements such as weight and height, which appear instantly on the app.

If internet connectivity is unavailable, the data is stored offline and automatically synced once the connection is restored. To ensure privacy, the video is deleted immediately after the measurements are extracted, and only the data is transmitted to public health administrators for real-time monitoring. Currently, 450 ASHAs have been trained in Dadra and Nagar Haveli, Daman and Diu.

Karya

In 2016, Vivek Seshadri and Manu Chopra started Karya as a project at Microsoft Research.

The founders saw that many rural workers depended on seasonal physical labour, despite rising smartphone access and cheaper data. They began exploring how to bring well-paying digital work—especially language-based tasks—to rural India.

In 2019, their first research paper showed that even first-time smartphone users in rural areas could accurately perform text transcription tasks. After four years of research, they spun out Karya from Microsoft Research as a non-profit in 2021.

Early pilots in Aamle (Maharashtra) and Soda (Rajasthan) saw strong uptake. The work was easy to learn, paid significantly more than manual labour, and offered dignity in low-wage settings.

Today, Karya covers 28 states, has impacted over 35,000 people, has 50% of its workforce being women, and has distributed over $800,000 in direct wages to its workers. Its workers have completed more than 35 million digital tasks on their smartphones.

Apurva.AI

Apurva.ai is a not-for-profit AI-driven platform founded by Anand Rajan that aims to transform how the development sector tackles complex problems. Drawing on Rajan’s experiences—from corporate work to grassroots farming—the organisation was born from the belief that traditional top-down solutions often fail because they overlook the insights of those closest to the challenges. 

With mentorship from leaders like Nandan and Rohini Nilekani, Apurva.ai focuses on bridging the collective wisdom gap by capturing and amplifying voices from communities and ecosystems, rather than relying solely on expert assumptions.

At its core, Apurva.ai uses a simple yet powerful framework—Listen, Learn, Act—to integrate authentic community experiences with broader institutional knowledge. This approach starts with gathering lived experiences, deepens understanding by connecting those insights with ecosystem data, and then informs actions that are contextually relevant and grounded in real needs.

The platform’s suite of tools, like Apurva Lens, Apurva Compass, and Apurva Thread, helps organisations uncover patterns, synthesise knowledge, and turn insights into solutions that reflect local realities.

ARMANN

Mumbai-based nonprofit ARMMAN works at the intersection of technology and public health to reduce maternal and infant mortality. Founded by paediatrician Dr Aparna Hegde, the organisation focuses on reaching underserved women—especially those in low-income and rural communities—with timely, stage-based health information during pregnancy and early motherhood.

Through large-scale mobile programmes like mMitra and the Government of India’s Kilkari service, ARMMAN delivers free, voice-based health messages directly to women’s phones in regional languages. These messages cover nutrition, antenatal care, immunisation, breastfeeding, and newborn care—ensuring critical information reaches women who may not have consistent access to frontline health workers.

In collaboration with Google DeepMind, ARMMAN has integrated predictive AI into its outreach systems to identify women likely to disengage from services. By targeting and optimising call delivery, the nonprofit has improved engagement and retention rates—demonstrating how artificial intelligence can strengthen last-mile healthcare delivery and potentially save lives at scale.

Adalat AI

Adalat AI is a justice-tech non-profit organisation founded in 2023 by Utkarsh Saxena and Arghya Bhattacharya to address India’s mounting case backlog and administrative strain in courts. Adalat AI works within the justice delivery system itself, building AI tools that reduce clerical burden and improve courtroom efficiency.

Its core offering is real-time speech-to-text transcription designed specifically for Indian courtrooms, trained to understand legal terminology, regional languages, and diverse accents. By converting spoken proceedings into structured, searchable text, it significantly reduces the time spent on manual documentation.

Beyond transcription, the platform integrates into court workflows to streamline case records, drafting, scheduling, and document management. The broader goal is to strengthen the justice system’s capacity, making proceedings faster, more transparent, and more accessible across states.