Bharat in Betastartup field guide
← Back to the atlas
No. 271Problem statementPublic services

Develop an AI/ML-driven system for topic clustering/modelling to enable auto-categorisation of received grievance reports for sharing of the same with last-mile officers concerned.

Source

Source summary

Use AI/ML topic clustering to categorise citizen grievances and route them to the relevant last-mile officers. The source also allows mechanisms to share reports with registered officials and monitor their progress, not just a classifier detached from the redressal workflow.

No. 170Problem statementPublic services

GIS Solution - GIS solutions across sectors - agriculture, water, forestry, urban affairs, road transport etc.

Use GIS to analyse locations, relationships and patterns for decisions and resource management across agriculture, water, forestry, urban affairs and road transport. The source invites cross-sector applications rather than one water-treatment or infrastructure product.

No. 174Problem statementPublic services

Preventing Human Intervention - Innovations to mechanize cleaning of septic tanks, drainages and manholes

Mechanise cleaning of septic tanks, drains and manholes to remove the need for people to enter or directly clean hazardous sewage environments. The source frames the ask around worker dignity, deaths and exposure to disease, rather than general municipal efficiency alone.

No. 250Problem statementPublic services

Government of India offices receive multiple communications on paper in regional languages. These documents (both printed and handwritten) to be digitized using OCR and then translate then worked upon and then needs to be translated back and responded in original regional language.

Help government offices digitise printed and handwritten communications in Indian languages, translate them for processing, and respond in the sender's original language. The requested open-source, cloud-based OCR service includes batch processing, image preparation, editable text, metadata, templates, forms and table extraction, handwriting recognition and multiple output formats. Targets are word-level accuracy above 95% and latency below one second per page; the source presents these as requirements, not demonstrated performance.