Earlier this year, the corridors of Bharat Mandapam buzzed with the phrase “populationscale impact” as the India AI Impact Summit brought together civil society, government, and experts to shape AI’s trajectory for humanity. However, this is not a mere agenda item limited to panel discussions. It is showing up in real-world products and services, being developed by a rapidly evolving ecosystem.

India is charting a new path, radically different from the prevailing global trends. In a nation home to a sixth of the world’s population, and across much of the Global South, aspiration sits uncomfortably alongside uneven access, low digital literacy, and gaps in last-mile connectivity. Big Tech services have been mostly designed for scale and not necessarily to serve the diversity of needs and constraints that are intrinsic to the Global South. Deploying AI for scale needs a very different playbook, one informed by lessons from India’s own experience. A recent analysis of nearly 110 home-grown startups and non-profits building AI for impact at scale reveals a decentralised landscape. It is shaped by solutions built around communityspecific problems, rather than chasing scale for its own sake.

In contrast to the ‘bigger is better’ ethos of Silicon Valley, a new generation of founders is building full-stack AI start-ups tailored to domestic realities. A distinct playbook for populationscale impact in India is emerging, defined by three conditions: localisation, offline-compatible design, and voicefirst AI. The first condition is localisation. In India, impact often means reaching rural populations, where barriers are not just technical, but economic, linguistic, and infrastructural. Solutions must be designed for low-connectivity environments, intuitive interactions, and affordability. Moreover, the hinterland is not a monolith and........

© The Statesman