India’s Cropin Technology has partnered with Google's Gemini generative AI chatbot to launch Sage, a real-time intelligence platform for the agricultural sector.
Thanks to Google's Gemini AI capabilities, the Bangalore-based start-up has developed an advanced agricultural intelligence platform that aims to predict future yields and help stakeholders make data-driven decisions to improve food production globally.
Founder & CEO Krishna Kumar tells us more.
ATN: Cropin at present leverages AI, machine learning, and data analytics to provide comprehensive agricultural intelligence solutions, aiming to improve productivity, sustainability, and profitability across the agricultural value chain… what distinguishes Sage?
KK: Firstly, it is the world’s first real-time Gen AI-powered agri-intelligence platform. This on-demand Global Agri-food Map divides the world’s agricultural landscape into grid-based sections, with options of 3x3 meters, 10x10 meters, or 5x5 kilometers, delivering data and intelligence with unprecedented accuracy and speed. Each grid holds valuable information and insights to guide food production planning in the short, medium, and long term. Cropin Sage allows agri-food businesses to use natural language to ask questions and gain valuable insights, exploring the past, present, and future potential of crops. It provides insights on critical topics such as crop suitability, long-term supply chain sustainability, production planning, and sourcing in a climate-impacted world.
Sage helps CPG players and food processors future-proof their supply chains. In a climate-impacted world, agri-food decision-makers struggle with the unpredictability of agricultural production. Sage removes this guesswork, offering past and current yields from global cultivation operations and predicting expected yields for the current and next year. It pinpoints ideal locations for cultivating specific crops anywhere across the globe in easy-to-comprehend grids. This granular intelligence empowers food producers to optimize their supply chains, eliminate uncertainties in procurement planning, and build resilience for the future.
Sage enables governments to secure regional food security. It provides policymakers with historical, current, and future insights on crop practices, irrigation needs, climate patterns, and soil. With this comprehensive intelligence, governments can make informed decisions to future-proof production and guarantee food security at the country level.
Sage is truly an AI-in-the-loop farming companion for agronomy decision-makers and the ultimate digital twin for stakeholders in the global food system. Its diverse use cases make it an invaluable tool in addressing the pressing challenges of the agri-food sectors.
ATN: From where are you getting the technologies that power the platform such, specifically the global climate data, a global crop knowledge graph, earth observation data and advanced crop models?
KK: Firstly, we are an AI-first agri and food system player with 14 years of experience in the field. Over these years, we have generated trillions of data points and built the world's most powerful crop knowledge graph. This graph spans 350 crops and 10,000 varieties across 103 countries, powering our flagship product, Cropin Cloud. We have digitized 30 million acres of farmland, positively impacting over 7 million farmers worldwide.
We collaborate with global geospatial and satellite companies, both government-owned and privately held, to leverage satellite data for farm monitoring at a global scale. Our advanced crop models are proprietary, developed from analyzing how different crops grow under varying climatic conditions, their reactions to weather changes, diverse management practices, and crop performance in varied soil types, among other factors. This extensive dataset has been built over a considerable period. Additionally, we utilize publicly available climate data and engage with weather companies based on the specific requirements.
Google is an investor in Cropin, making this partnership both strategic and long-term. We use Google Cloud to host our services, and Google has been a trusted technology partner for various agri-food transformation initiatives we are driving today.
At the heart of Cropin Sage is the Gemini model (Gemini Flash 1.5), which plays a pivotal role in transforming user interactions. The Gemini model generates corresponding SQL queries when an enterprise user inputs a text prompt or query. By leveraging Google Gemini, Cropin Sage offers a seamless and intuitive user experience for our enterprise customers, making complex data accessible and actionable for agri-food stakeholders. This on-demand, Gen AI-powered agri-intelligence platform provides unprecedented scale and insight, available to all enterprise customers who come on board to leverage this transformative technology.
ATN: How many customers are currently using it? How much does it cost?
KK: We have at least half a dozen customers either in the pilot stage or with signed contracts to use Sage. Due to non-disclosure agreements with our customers, we cannot provide more details. The cost of deploying Sage is highly variable, depending on factors such as the number of crops for which intelligence is needed and the scale of deployment (whether it's for 100 hectares or an entire country), among other considerations.
ATN: How actionable is the data they receive?
KK: They don't need additional resources or an army of analysts to make use of the data. This is a major advantage that Cropin Sage offers our customers. You can give any prompt that comes to mind and receive ready-to-use insights within 2-3 seconds.
Cropin Sage provides access to global cultivable lands through small, easy-to-analyze grids, enabling agribusinesses and governments to make informed decisions based on historical, present, and future data on cultivation practices, crops, irrigation, climate, and soil. Sage goes beyond just collecting historical data on agriculture, crops, and weather patterns. It leverages powerful AI tools to predict future yields, offering comprehensive regional intelligence patterns for specific crop value chains in a climate-impacted world.
This platform is the perfect Gen AI companion for agri-food businesses, providing them with the ability to make informed decisions based on data, trends, and ground realities. Sage seamlessly navigates through a vast array of proprietary and public datasets, linking grids and crops from around the world. It integrates and analyzes these datasets to deliver comprehensive insights. By condensing months of data processing and analysis into seconds, Sage transforms complex information into actionable insights with an intuitive user interface featuring charts, data, graphs, and infographics using GenAI capabilities.
ATN: A few bottlenecks for the wider adoption of AI technology in ag have been identified: data silos; fragmented data sources and legacy systems; hallucinations; and potential biases. How have you deal with these kinds of challenges?
KK: We have been dedicated to the digitalization and building predictive intelligence of the agri-food businesses for over 14 years. Our agricultural data models have been meticulously built over these years, grounded in local agronomy, management practices, soil conditions, and other critical factors. Our models are built generically but are fine-tuned for specific crop varieties and agro-climatic regions. We have also validated our data with enterprise customers across various regions over the years.
We've leveraged our partnership with Google, specifically Google Cloud & Gemini, to build innovative and responsible AI solutions that ensure fairness, transparency, and safety. The architecture of Cropin Sage is specifically designed and trained to handle various crop varieties. Cropin Sage avoids hallucinations by relying on our proprietary data and not public data alone, with built-in guardrails to maintain accuracy and reliability.
Cropin Cloud, the industry cloud for agriculture and food systems solves these problems for the industry and our customers. Cropin Data Hub helps our customers break data silos, connect legacy systems, and organize all data in a structured and harmonized manner in one place, making it insight-ready. It allows for the ingestion of diverse data types from both public and private datasets.
