Govt needs to expedite creation of DPI in agriculture sector: Cropin CEO Krishna Kumar

Cropin co-founder and CEO Krishna Kumar at Davos 2025

The government needs to fast-track the process of creating digital public infrastructure (DPI) for the agriculture sector, said Krishna Kumar, co-founder and CEO of agriculture technology (agritech) startup Cropin.

“The government is doing their bit and they are on the right track while talking about agritech DPI. One of our requests is to expedite this. While progress is faster than before, it still quite slow from our lens,” Kumar said, speaking to Moneycontrol at the World Economic Forum (WEF) in Davos on January 20.

According to Kumar, the creation of DPI for the agritech space will accelerate growth in the sector.

In September 2024, the government had approved the Digital Agriculture Mission, with a budgetary outlay of Rs 2,817 crore to support digital agriculture initiatives. Moreover, in the Union Budget 2024, the Centre proposed Digital Public Infrastructure (DPI) for the agriculture sector, aiming to bring more than six crore farmers into the formal land registry system.

One of the core components to building this DPI is the Agristack – a single data pool or a repository of all farmer information, including their identity, land records, coverage, income, insurance, loans, crop details and revenue history.

According to Kumar, the development of this Agristack will open up further avenues for the creation of products and solutions for farmers.

“Agriculture starts from land ownership. Once you digitize the farmer, and land ownership, you can build a lot of businesses on top of it,” Kumar said.

“For instance, if a grower wants a loan, a bank needs to underwrite. (For this) the bank needs more data and risk analysis needs to be done. Same for insurance products, or a marketplace. A lot will be built on it (agri stack). Once you have an agri stack, you will see a lot of solutions and products getting built on top of it,” he explained.

Improving crop output using LLMs

Cropin leverages artificial intelligence (AI) to predict crop yields and assess the impact of climate risks on supply chain. The Google-backed startup, which is present in 103 countries, counts Pepsico, World Bank, Walmart, and McCain, among others, as its clients.

To offer more nuanced and detailed information to farmers, the company has been experimenting with large language models (LLMs) to offer a suite of specialized offerings.

Last year, Cropin partnered with Google’s Gemini AI chatbot to launch the agri-intelligence tool Sage to predict crop output and weather patterns based on historical data. This year, it also launched Aksara, an open source micro language model for farmers, which offers farmers inputs and advisory functions.

“(With Aksara) we wanted to establish that you don’t have to burn the whole world’s resources to localise it…So I think this will be the way to go. And with the verticalization of these LLMs, you’ll see a lot of products getting launched in this direction,” Kumar said.

Unlike horizontally trained LLMs, which offer general knowledge across a broad range of topics, vertically trained LLMs are tailored to excel in a specific domain, enabling them to provide more accurate and insightful information.

“For instance, one of the things large language models are solving at scale is communication and dialect. This was one of the biggest challenges. Earlier, we used translators to do that which was very tough. But now, it’s on the fly,” Kumar said.

However, to ensure that this cutting-edge technology is accessible to farmers with limited resources, Kumar said that Cropin is developing its offerings specifically for the last mile.

“We have to make it (technology) very user friendly and actionable for them. Farmers don’t have to pay for it, the enterprises do. And while enterprises are benefiting from this, the growers are then end consumers of the technology,” he said.