Inspiration:

The central idea of the Agri-Tech Agent was to be a small, but positive difference to a community that are often neglected by well-known tech solutions. Our background was what inspired us, coming from a background in agriculture and understanding the issues facing farmers at a subsistence level where agriculture is the principal economic activity of the community. We recognized that these famers often do not have even timely information on crop disease, price volatility in markets, and uncertainty surrounding climatic events. While this lack of knowledge puts the communities under a direct threat financially, it also adds stress that contributes to their financial burden. Our goal was to design a tool that could provide solid information to counter these gaps, without being invasive, to those that need the information most.

What it does:

The Agri-Tech Agent is more than just another app; it is a complete, AI-powered ecosystem which seeks to provide small-scale farmers with the levels of information and service previously only provided to commercial farms. The Agri-Tech Agent is a personal digital assistant that communicates with the user via an easy-to-use interface like a mobile app or messaging chatbot.

1. Proactive Crop Health Management:

Rather than waiting until there is a problem, the agent takes a proactive approach to crop health. The agent enables the farmer to detect pests and diseases in real-time:

  • Pest and Disease Detection: A farmer simply takes a picture of the diseased leaf or pest and sends it digitally to the agent. With some computer vision models, the agent analyzes the picture in real-time and determines the pest or disease in seconds...the farmer receives an immediate diagnosis.
  • Local Recommendations: Not only does the agent name the issue, it provides action, local recommendations. The agent recommends specific pesticides or organic treatments that are available in the local market, providing the correct dosage, application method and precautions for safety.
  • Predictive Alerts: The agent uses real-time weather, in conjunction with historical pest patterns in the farmers' region, to produce predictive alerts for impending pest outbreaks or conditions for a particular crop disease. The agent sends alerts advising the farmer preventive measures to protect crops prior to being affected.

2. Intelligent Weather and Climate Advisories:

Unpredictable weather is a major source of risk for farmers. The agent helps farmers mitigate this risk with hyper-local, timely information.

  • Real-time Weather Forecasts:
    First-hand Weather Forecasts: The application provides a detailed weather forecasting function, relating specifically to the individual farmer's location, from as short as 5-7 days. With real-time updates, this application helps farmers determine what they want, and the most effective weather forecasts are the most critical part of this, whether that is likely weather, sunlight or rain, or extreme temperatures.
  • Extreme weather alerts: The agent provides notifications and alerts regarding extreme weather events, such as rainfall, hail occurrences or unseasonably high-temperature heatwaves, alongside possible generative directions as to ensuring protection of produce, with suggestions including draining excess water from fields, or instructions to cover fruit and vegetables known or thought to be possible vulnerable.

3. Market Price Optimization:

Farmers often struggle to get fair prices for their produce due to a lack of market information. The Agri-Tech Agent addresses this directly.

  • Real time Mandi Pricing:
    The agent pulls real time prices from neighbouring mandis (agriculture markets) for crops available for sale in the current day. Upon communication of a readiness to sell the crop, the farmer simply states what is the best price for my crop and the agent collects and presents the data to the farmer for their decision making.
  • Route and Logistics Advice: The agent not only presents the farmer with the best price but also provides the farmer with optimal mandi to sell based on distance and demand, enabling the farmer to maximize revenue.

4. End to End Automated Workflows

This is where the agent is more than a simple query-response agent thanks to the Agent Development Kit (ADK) as it can manage and automate the whole process.

  • Resource and Inputs:
    The agent could promote nearby sellers of seeds, fertilizers and other farm inputs, provide sister prices, and even ensure they purchased good quality inputs.
  • Data-Driven Actions: The agent logs every interaction and stores and utilizes responses for all data points over time. The agent is creating a history of a farmers profile, which analyzes farmer data trends over time. It could tell a farmer (for instance) that they produced a yield increase of \(10\%\) the previous season using a certain mix of fertilizer, and hence that is likely what they were doing right and therefore do it again.

Essentially, the Agri-Tech Agent will turn a farmer's smartphone into a useful always-on tool, delivering not only information, but personalized, proactive and actionable intelligence that enables farmers to farm more efficiently, mitigate risk, and ultimately drive their income.

What's in front of us:

We have set the stage for this hackathon through a robust feasibility study and exploration process. we've validated the viability of our concept and have a clear trajectory. Our next phase is clearly moving from an articulated vision to a physical, functional product.

The first task is a Minimum Viable Product (MVP) build. The first features of the MVP will be image-based disease detection and real-time mandi price notifications, both powered from our tuned AI model.

This will allow us to gather the critical input from farmers and test on location. Our focus is to build a scalable, significant solution that will genuinely improve the lives of those in rural India, then add more excellence, both in terms of features (e:g: predictive weather notifications) and in automating workflows.

Built With

Share this project:

Updates

Submission history