Rising at 3 a.m. to trek 30-40 km to the local wholesale market or mandi, then haggling with 12-13 traders just for one fruit type, followed by grading, loading, and hauling it back to open shop by 8 a.m., that's the relentless daily grind for most of Delhi's neighborhood fruit retailers, locally known as "thelawalas" who end up spending up to INR 500-600 in pre-work costs before they earn a single rupee in India's capital city, says Rohit Nagdewani of supply-chain company Fresh from Farm.
After spending years as a controlled-atmosphere farmer immersed in mandis through 2021, he observed that this operational stack drained small, unorganized fruit and vegetable pushcart hawkers. "99% of fruit sales still happen offline at these local spots, so eliminating all the pre-dawn steps involved in fruit procurement was crucial," Rohit reflects on the idea behind his supply-chain service.
© Fresh From Farm
"Eliminate it, procure straight from farmers, sort and clean at a 20,000 sq ft tech-enabled facility in Chhatarpur with ripening chambers for bananas and mangoes, cold storage, and machine-packed produce, then deliver to over 600 retailers via a 40-45 vehicle electric fleet by 7:30 a.m," Rohit shares a slice of Fresh From Farm's behind-the-scenes operations. "Every pincode is within 90 minutes, enabling these unorganized retailers to skip the pre-dawn chaos entirely."
His team is now experimenting with AI through PhalNetra.ai, their in-house predictive procurement intelligence layer built on four years of transaction data from these same retailers. "It predicts what volume and quality will sell next week, down to benchmark prices, sourcing spots, and where to offload excess. This led to Fresh From Farm cutting its wastage from 11% to under 2% since we procure only what moves." Their retailer network also gets suggestions for procurement quantities to dial back on their own waste moving, moving from random gut-feel sourcing to data-backed orders, he explains.
© Fresh From Farm
Rohit reveals that the full pilot launch at 99% confidence level is about 6-8 months away. "The AI system currently pulls near-real-time data from app orders placed 6-11 p.m. nightly, learning continuously and factoring in hyperlocal disruptions like truck breakdowns, elections, or weather shifts that are reflected in market prices. Our data patterns also reveal insights like watermelons dominating Noida, avocados faring better in Gurgaon than South Delhi, and blueberries selling more in New Gurgaon."
Looking ahead, Rohit plans to layer in one-click city expansion tools and real-time procurement alerts, for instance, understanding how landslides in Himachal Pradesh impact apple costs. "No similar artificial intelligence or machine-learning demand-planning tool exists for unorganized fruit trade in India or abroad," he mentions, adding that they will be expanding into a new city within three months and targeting seven more across India in 18 months."
For more information:
© Fresh From Farm
Rohit Nagdewani
Fresh From Farm
Email: [email protected]
www.freshfromfarm.in
