What You Should Know
- Pune-based healthcare AI company Froncort.AI received $2.3M in grant funding, $200k in Google Cloud credits, and pro bono technical engineering expertise from Google.org.
- Selected as one of 15 global recipients in Google.org’s Impact Challenge: AI for Government Innovation, building on an initial validation study funded in 2025.
- Addresses a major global health challenge: India accounts for roughly 25% of the world’s ~11 million annual tuberculosis (TB) cases, where routine door-to-door community campaigns screened nearly 20,000 individuals to confirm just one positive TB case.
The Algorithmic Engine: Merging Earth Observation with Epidemiological Baselines
Founded in 2024 by physician-scientist and engineer Dr. Ujjwal Rao, Froncort.AI was designed to bridge clinical diagnostic logic with geospatial predictive modeling. Rather than deploying AI as an ambient charting tool or passive data dashboard, the core platform, AI-NETRA, generates actionable, micro-targeted field itineraries:
- Multimodal Geospatial Ingestion: The model synthesizes local public health surveillance metrics with Google Earth AI satellite models, layering environmental and topographical risk variables such as settlement density, structural informal housing, land use patterns, regional poverty indexes, migrant labor movement, rainfall levels, and local flood topography.
- Micro-Targeted 300-Meter Hotspots: The algorithm segments geographic sectors into high-risk 300-meter micro-zones, analyzing the intersection of environmental and demographic variables to predict clusters of undiagnosed disease.
- Turn-by-Turn Dynamic Screening Paths: The platform converts predicted hotspots into ranked, household-by-household work-paths optimized for physical field navigation. Instead of walking broad perimeter lanes door to door, health workers follow prioritized routes directly to household clusters demonstrating the highest clinical vulnerability.
- Sovereign In-Country Data Architecture: While global satellite and environmental representations are derived from Google’s Earth AI models, all statutory government health registry metrics and patient encounter records remain stored and processed strictly within domestic infrastructure in India.
In a formal validation study conducted under ethics approval with the State TB Office of Maharashtra, AI-guided screening teams matched the pulmonary TB case detection numbers of conventional campaigns while screening two-thirds fewer people. The targeted routing lowered the public health detection cost from approximately $29,000 to $10,000 per confirmed case, with 92% of participating frontline workers reporting a substantial drop in physical workload.
Dr. Ujjwal Rao, Founder and CEO of Froncort.AI, framed the operational imperative: “The world does not lack experts. It lacks their reach. AI-NETRA puts that expertise in the health worker’s hands at the door: where to go, whom to screen, and when. Google.org brings more than funding: Google’s Earth AI models and engineers we wouldn’t have access to. Add our clinicians’ judgment and a startup’s speed, and the world’s 2030 goal of ending these epidemics comes back within reach.”
