This Startup Built A Treasure Trove Of Crop Data By Putting A.I. In The Hands Of Indian Farmers Posted on : Oct 15 - 2018

Tushar Kamble knew there was something wrong with his chilli plants when he noticed they were smaller than average, and the leaves were starting to curl.

Last year the former academic, who’d taken a bold leap into farming in his 30s, asked his neighboring farmers for advice one morning at the border between his seven acres of land and theirs.

Kamble, 38, got lots of different opinions about aphids and disease. Then he tried an app called Plantix, and used it to take a photo of his chilli plant. The app cross-referenced it against a database 50 different species using image recognition, a type of machine-learning, and within two minutes he had a different answer: his chillis weren’t getting enough water, and they’d benefit from a micronutrient spray to.

Within a few weeks, Kamble’s chillies had grown to a decent size. Today he is using the app to help him change his chemical usage so he can go organic, and export his crops to Europe. “I don’t depend on other people to help me,” he says with a smile. 

Boosting the yields of one farmer's crops may seem trivial, but smallholding farmers like him make up a very long tail of global food production.

Some 70% of the world’s food comes from smallholders like Kamble and there are around 500 million of them globally, says Simone Strey, co-founder and CEO of Peat, the Berlin startup that makes Plantix. Her company is using images from Kamble and others to build a bigger picture about the health of crops across India and beyond.

Peat, which Strey started with her husband and machine-learning specialist Robert Strey in 2015, now has 620,000 monthly active users, and 80% of them are in India.

Sandip Shinde is another farmer who uses Plantix on his 25 acres of land where he cultivates tomato, cauliflower and sugarcane, and despite his 13 years of experience, he believes the apps’ automated suggestion system and chat-network with other farmers has helped boost his yields too.

But what’s good for smallholding farmers may also be useful for the world’s biggest agricultural and chemical companies. Peat has found a nexus between the two worlds for an initial business model, white-labelling its plant-recognition software for chemical company BASF, which has integrated it into its own crop management software called Maglis. 

Professional farmers who use Maglis tend to have much bigger fields than the smallholders in India, and they also access Peat’s software less often - up to around 12,000 times a month. The startup charges a fee for each “call” to its API, or every time a professional farmer uses Peat’s feature on the Maglis app. View More