- Manual quality checks continue to dominate Indian packhouses
- AI-powered grading could reduce quality disputes
- Agrograde raising Pre-Series A to scale autonomous packhouses
Despite rapid automation across sectors such as e-commerce and grain supply chains, fresh produce remains heavily reliant on manual processes.
Kshitij Thakur, founder and CEO of Indian startup Agrograde, estimated that more than 99% of packhouses in India still conduct activities such as quality checks, grading, sorting, weighing and packing manually.
“The post-harvest supply chain in India is going through an inevitable crisis,” Thakur told AgNavigator.
He highlighted that the existing post-harvest supply chain was developed for a different era and falls short of today’s requirements for scale, transparency, traceability, efficiency and consistent quality.
The sector is also grappling with an ageing workforce, as the average age of employees across packhouses has surpassed 45.
Thakur said this dependence on subjective human assessment has created a trust deficit across the fresh produce trade, particularly in crops such as onion and potato, where quality directly influences pricing.
“India is one of the leading producers of onion and potato, and yet the majority of packhouses still run on manual, subjective operations. There is a huge trust deficit in the trade. Quality is still determined manually, which introduces subjectivity and directly affects the price a producer gets,” Thakur said.
This uncertainty often results in price discounts, rejected consignments and lower returns for producers.
“In the absence of objective quality, trust in the trade becomes person-driven rather than quality-driven, and at every node the risk from uncertainty about quality compounds. That risk premium gets adjusted against the purchase price, which is why producers never get a fair price for their produce,” he said.
The challenges to digitising
According to Thakur, many companies have tried to implement post-harvest technologies but failed to adapt to India’s unique post-harvest challenges.
He noted that in India, produce from multiple crop varieties often arrives at collection centres and wholesale markets in mixed lots containing dust and other foreign material.
These handling and aggregation points frequently operate in hot, dusty conditions, adding further complexity to quality assessment and sorting processes.
Thakur added that the physical characteristics of each crop play a critical role when designing any technology that will handle thousands of tonnes of produce every season.
He cited onions as an example, noting that imported grading systems often damaged the crop’s delicate outer skin.
“Technologies developed by global leaders in this space have failed miserably in Indian packhouses because they were not designed to handle this kind of variation and were not designed for Indian crop varieties in the first place,” Thakur said.
Wider adoption of post-harvest automation would require greater grading capacity near farm gates, smarter storage intake systems and improved financing for FPOs and mid-sized packhouses, he added.
Local for local
Agrograde is positioning its technology as a locally designed alternative to imported systems that have struggled in Indian operating conditions.
To address these challenges, Agrograde spent four years and six product iterations developing its AI-powered optical grading technology.
Unlike conventional colour-sorting systems, which can struggle with unwashed and mixed-variety produce, its models analyse texture, boundaries and context to distinguish between defects and natural variations.
According to Thakur, Agrograde’s Vector series can sort unwashed, field-fresh onion and potato lots with up to 96% defect detection accuracy and switch between potato varieties without recalibration.
The company now has 130 machines deployed across 14 states and four crops.
Agrograde is currently raising a Pre-Series A round to expand manufacturing and accelerate development of autonomous packhouse operations, beginning with onion and potato supply chains where its technology is already deployed.
