AgriChain AI

InspirationInspiration

India’s agricultural sector feeds millions but suffers from inefficiencies in the farm-to-market supply chain. Smallholder farmers often lack access to real-time market data, logistics options, and fair pricing, which results in both financial losses and food waste. Witnessing these challenges inspired us to create AgriChain AI — a solution that empowers farmers with actionable insights and reduces post-harvest losses.

What it does

AgriChain AI uses Generative AI agents to optimize the farm-to-market journey. It collects real-time data from weather APIs, mandi (market) price feeds, government agriculture databases, and logistics providers. The system then:

Predicts demand and price fluctuations.

Recommends the most profitable selling locations.

Suggests cost-efficient logistics routes to minimize spoilage.

Notifies farmers in local languages via SMS, WhatsApp, or mobile apps.

This creates a transparent, farmer-friendly ecosystem that also benefits consumers through stabilized prices and better availability.

How we built it

We leveraged IBM’s open-source Granite series of models, data preprocessing toolkits, performance benchmarks, and the Agent Development Kit (ADK). Our pipeline integrates real-time data feeds, processes them with predictive models, and deploys AI agents that can automate notifications, bulk aggregation, and logistics coordination. The system was designed to be lightweight and mobile-first, ensuring accessibility even in low-connectivity regions.

Challenges we ran into

Integrating multiple heterogeneous data sources (weather, prices, logistics).

Designing AI models that can handle regional crop diversity and demand variability.

Ensuring the platform works for farmers with limited digital literacy and connectivity.

Optimizing for both farmer profit and consumer affordability without bias.

Accomplishments that we’re proud of

Built a functional prototype that integrates prediction, recommendation, and automation in one seamless workflow.

Enabled multilingual, farmer-friendly interfaces.

Demonstrated real-world scalability by aligning production, pricing, and logistics.

What we learned

How to leverage Generative AI agents beyond text, for real-world decision-making.

The importance of designing AI systems that are inclusive and accessible in rural contexts.

That combining AI-driven insights with human decision-making creates the most sustainable outcomes.

What’s next

Expanding partnerships with government agriculture boards and cooperatives.

Incorporating blockchain for transparent farm-to-market transactions.

Extending support to more crops and regional languages.

Piloting AgriChain AI in rural districts to measure direct farmer income impact.

What it does

How we built it

Challenges we ran into

Accomplishments that we're proud of

What we learned

What's next for AgriChain AI

Built With

• docker
• flask
• github-actions-(ci/cd)
• hugging-face-transformers
• ibm-agent-development-kit-(adk)
• ibm-cloud
• ibm-granite-models
• javascript
• javascript-frameworks:-flask-(backend)
• kubernetes-databases:-postgresql
• logistics-apis
• logistics-apis-cloud/platforms:-ibm-cloud
• mandi-price-feeds
• mongodb
• mongodb-other-tools:-ibm-agent-development-kit-(adk)
• postgresql
• python
• react
• react-(frontend)-ai/ml:-ibm-granite-models
• scikit-learn-apis:-government-agriculture-databases
• twilio/whatsapp
• weather-apis