Inspiration :- Rural Indian farmers face numerous challenges: fragmented workflows, language barriers, delayed payments, and heavy paperwork—all of which hurt their incomes and food security. We were inspired by their resilience and wanted to harness agentic AI to empower them with automation they can trust.
What it does:- KrishiSahayak is an end-to-end AI agent system that automates the entire farm-to-market journey. It delivers multilingual crop planning, responsive agronomy advice (including vision-based disease detection), market discovery, digital contracts, credit/insurance pre-qualification, logistics coordination, and rapid payment settlement—customized for local needs.
How we built it :- We leveraged IBM’s open-source Granite models for language, vision, and code, orchestrated by the Agent Development Kit (ADK). Data pipelines were created to unify weather, soil, satellite, mandi prices, and transport info. The system is accessible by WhatsApp/IVR and a lightweight mobile app, built to work offline.
Challenges we ran into :-
- Integrating diverse, real-time data sources and vendor APIs 2.Optimizing AI for rural low-connectivity environments 3.Multilingual user experience design and local trust-building 4.Ensuring responsible AI with human-in-the-loop for sensitive workflows.
Accomplishments that we're proud of:- 1.Built a fully agentic, end-to-end automation workflow for farming operations 2.Achieved multilingual support and vision-based crop health diagnosis 3.Enabled rapid onboarding of FPOs with robust safety and explainability features 4.Demonstrated measurable gains in price realization and efficiency in pilot testing.
What we learned :-We learned how crucial local context and user feedback are for rural deployment, the power of open-source AI and agentic frameworks for scalable automation, and the importance of robust safety and auditability in agricultural fintech.
What's next for KrishiSahayak:- We plan to expand across states, support more crops and languages, integrate with new scheme APIs, and grow into a national platform for smart, resilient, and inclusive agriculture. Further plans include carbon tracking, financial inclusion, and custom marketplace integrations
Built With
- amazon-web-services
- and-vision-models)-cloud-&-platforms:-ibm-cloud-(watsonx
- cloud-functions)
- code
- deployment
- docker
- edge
- fastapi
- for
- github-actions-(ci/cd)
- ibm-watsonx.ai-databases:-mongodb
- ivr/sms-gateway-(twilio)
- javascript
- kubernetes
- languages:-python
- logistics-apis
- mandi/fpo-market-data-apis-other-technologies:-whatsapp-business-api
- ndvi/satellite-data-feeds
- offline
- postgresql-apis-&-integrations:-openweathermap-api
- react-native-ai-models:-ibm-granite-(multilingual-llm
- typescript-frameworks:-ibm-agent-development-kit-(adk)
- upi-payment-api
Log in or sign up for Devpost to join the conversation.