Inside India’s plans to bring its agri economy into a tech stack

This is the reality for millions of growers in India, who are engaged in an occupation fraught with risks. When the rains are bountiful, holding the promise of a bumper harvest, the wholesale prices could crash. With no market intelligence to rely on, farmers often use the previous season’s price as the yardstick to choose what to plant. The upshot: periodic gluts and shortages.

Can digital smart farming services provide a solution? The agriculture ministry thinks so. Over the past few months, the ministry has set in motion an ambitious project to reset how farmers manage their businesses and the tools that they use to make decisions.

Named AgriStack, the project will collect granular data to provide growers with a range of customized services—on what to plant, where to sell, market information on price movements, and linkages to formal credit arrangements.

Each farmer will be provided a unique farmer’s ID, which will be linked to her Aadhaar number. It will contain details related to land ownership, the crops she grows, soil health and the benefits available under government schemes such as direct cash transfers, crop insurance and subsidized credit.

“Data is the new oil. When managed properly, it can fuel innovation and support several value-added services," said a consultation paper, titled India Digital Ecosystem of Agriculture-IDEA, released earlier this month.

Between April and June 2021, the agriculture ministry signed agreements with top tech developers and agribusinesses, including the India unit of Microsoft Corp. and Amazon Web Services, Patanjali Organic Research Institute, Star Agribazaar Technology Ltd (a digital marketplace for farm produce), and ESRI India Technologies Pvt. Ltd (a geospatial mapping platform). The ball has been set rolling on a pilot which will touch more than 800 villages in seven states.

The idea is to treat individual farms as a unit, connect the data points on AgriStack and then offer growers both public and private sector services on the platform, said Vivek Aggarwal, additional secretary in-charge of digital agriculture at the ministry of agriculture and farmers welfare.

Red flags

India’s rural economy is intrinsically linked to land, whose ownership is often deeply contested. Naturally, farm activists and digital rights groups have raised a slew of red flags about the impending big data project.

They have pointed out that linking land ownership data with Aadhaar may hit a roadblock because the process of digitizing land records is still a work in progress in a country where land disputes account for more than 60% of all civil litigation.

The critics argue that a blind reliance on technology may lead to exclusion errors in welfare programmes (similar to the experience in the food subsidy scheme). Most critically, not consulting either farmer groups or state governments runs the risk of a meltdown later— like it did after a set of farm reform laws were passed last year, leading to one of the longest running protests in post-independent India.

To be sure, there are positive spin-offs as well. Over the past one year, when the farm ministry compared enrolment in the PM Kisan direct cash assistance programme with that of the Kisan Credit Card (KCC), it found that more than 25 million PM Kisan beneficiaries did not have KCCs.

“It helped us connect more than 20 million farmers to concessional credit. That’s the power of data sharing and management," Aggarwal said. “With layers of data, we can provide plot-specific crop advisories, estimate yields and link produce with markets. The pilot project will provide the proof-of-concept… after solutions are identified and established, we can attempt a national level scale up."

The startup ecosystem is already upbeat about AgriStack. “It will be a huge win if digitized land records and cultivation details are added to the existing JAM and IndiaStack (Jan Dhan bank accounts, Aadhaar and mobile numbers)," said Mark Kahn, co-founder of Omnivore, a venture capital firm that has funded more than 20 startups since 2011.

According to Kahn, agri-tech startups received over $1.5 billion of investment in the past 5-7 years and AgriStack can open the taps for more inflows. “…startups had to invest in creating the data architecture. With an AgriStack in place, they can concentrate on commercial activities."

Artificial intelligence (AI) and machine learning (ML) from large datasets will not only support farm management, but can also help develop and design new crop protection products, said K. C. Ravi, chief sustainability officer at Syngenta India, a leading pesticides and seed manufacturer, which has over 250,000 Indian farmers registered on its Anantham app. Globally, 125 million acres of farmland is actively managed by growers who use the Syngenta group’s digital services.

Not everyone in the broader tech ecosystem has a rosy outlook on the impending shift though. “A stack is like a layer of bricks. You can make anything out of it. The question is what are the use cases? Who will the use cases benefit? Whoever builds these use cases will build it to their own advantage," cautioned Prashant Mehra, co-founder of Platform Commons Foundation, a non-profit that builds digital platforms as a public asset.

For instance, real-time data on tomato planting can be used to warn farmers of an impending glut or shortage, with an advisory to slash or raise area under the crop. But if the same data ends up solely in the hands of a set of large traders, it could help them manipulate markets. Farmers’ groups should proactively propose to the government use-case scenarios that will benefit them, Mehra said.

He also warned that anonymizing data on the AgriStack platform isn’t enough. The value of aggregate data is many times more than the individual data of a user—which is why cab aggregators hold enormous power over both drivers and riders on their platforms.

The AgriStack project may end up benefiting large corporations by providing them access to the data of millions of growers. Such vested interests can then sell farmers everything from chemicals to credit and pare their own current marketing and human resource costs, said Kavitha Kuruganti, convener of ASHA, a farm policy advocacy group.

More than advice and information, farmers need assured returns, said Ramesh Panghal, a vegetable grower from Haryana who recently incurred heavy losses following a sudden glut in supply.

“Also, what good advice will come from soil health data that is inaccurate?" asks Panghal, flagging what could emerge as a formidable task: capturing high quality data on crops and soil conditions from India’s 150 million land parcels.

Slew of MoUs

On 13 April, the farm ministry signed a memorandum of understanding (MoU) with Microsoft India to build a unified farmer service interface using Azure-based data and analytics services.

To provide a proof of concept for the model it develops, Microsoft will carry out ground surveys in 100 villages in six states—Uttar Pradesh, Madhya Pradesh, Haryana, Gujarat, Rajasthan, and Andhra Pradesh—to collect data.

Interestingly, according to a standard operating procedure document obtained under a Right to Information query filed by the Internet Freedom Foundation, it is Microsoft’s responsibility to ensure that the captured data is reconciled 100% with physical records. It remains unclear how Microsoft will reconcile discrepancies in land ownership data—an exclusive domain of the states.

The MoU states that Microsoft will bear its own expenses for the survey and the cost of building the technology platform. But what is Microsoft’s revenue model? Why is the company investing money in AgriStack? Microsoft India did not reply to queries from Mint.

Star Agribazaar signed a similar agreement with the ministry on 1 June, but with a larger set of responsibilities: not just digitally profile agricultural land and develop an advisory platform, but also enable market connect and financial services for farmers. That implies the data may have to be shared with service providers. In the pilot phase, Agribazaar will work in 724 villages in three states—Uttar Pradesh, Madhya Pradesh, and Rajasthan.

“The major challenge (for now) is the revenue model… we don’t know when revenues will flow and from where," said Amit Mundawala, co-founder and managing director of Agribazaar. “But we believe it will give very good fruits in the future… may be in two, three or five years. Meanwhile, we will get to connect with the farmer; know them better; get the complete data."

Mundawala added that the farmers’ data can only be shared after consent is taken for a specific purpose. “The data is owned by the government, and we have no rights over it. But farmers can sell their crops on our platform. We can help insurance firms with crop loss assessment (for a fee) and provide advisory services (which may include recommendations on which seeds, fertilizers and pesticides to buy)."

Responding to the query about why MoU partners were investing their own funds in the pilot phase of the project, Aggarwal, the additional secretary at the ministry, said: “The technology players have to provide a proof of concept (PoC) and show their model is working… so, when the national scale up happens, they have a good product to launch."

“These are companies who want to participate in the digital agriculture ecosystem… they could benefit in the long run, say, if you are a cloud service provider (like Microsoft and Amazon) or a trading platform like Agribazaar."

But how will companies reconcile land ownership data without the cooperation of state governments? “We are already working with the state governments under the National e-Governance Programme on Agriculture. (On AgriStack), we have released a consultation paper and are inviting suggestions from states, farmers and all stakeholders… If we had only done consultations without PoCs, it would have been a paper exercise."

Interestingly, the MoU with Patanjali Organic Research Institute notes that it has a “focus on future-proof technologies and innovation" and will provide the tools to build a platform with AI and ML technologies to provide farm management and farmer services.

Power and heartburn

The digital ecosystem (IDEA) consultation paper released early in June by the agriculture ministry recommends implementing the AgriStack project on a mission mode by establishing a National Mission for Digital Agriculture and setting up an Agri Data Exchange for “fast and purpose driven exchange of data".

But what is certain to cause heartburn among states is a suggestion to link a share of central assistance in agriculture to progress achieved in implementing IDEA.

Another potential conflict point could be opposition from farm groups on linking Aadhaar with land ownership data. If the farm laws passed by Parliament (in September 2020) could be misconstrued by farmers as a bid to take away their land, sharing land ownership details with private entities could be a bigger flashpoint.

The scale of the AgriStack that India is attempting—to bring close to 150 million land holdings on a single platform—has no parallels globally. So far, it has been confined to agriculture and technology giants developing their own data platforms to service farmers.

But the global experience is telling. A 2017 survey of farmers by the Australian government observed that growers displayed a general willingness to share data, but they also expressed “concerns over aggregated data in relation to privacy, financial advantage taken by other businesses, and the potential for it to be used to influence the markets, such as produce prices and land value".

Yes, digital technologies can be made to work for farmers, consumers, farm workers and the environment, said a January 2021 research paper by the international non-profit GRAIN titled Digital Control. “But technology does not develop in a bubble; it is shaped by money and power—both of which are extremely concentrated in the tech sector," the paper said.