Tobacco Board Partners With Constems AI to Bring AI-Powered Grading to Andhra Pradesh Auctions

The Tobacco Board of India, Guntur, has partnered with deep-tech Vision AI company Constems AI Pvt. Ltd. to introduce AI-powered grading of Flue-Cured Virginia (FCV) tobacco bales across select auction platforms in Andhra Pradesh.

The initiative will use computer vision and deep learning to make tobacco grading faster, more consistent and transparent. FCV tobacco is currently graded manually by qualified classifiers, a process that relies heavily on human expertise and judgement.

Since the grade assigned to a bale directly affects its market value, the partners believe an AI-assisted system could reduce subjectivity and improve consistency.

Constems AI will deploy its proprietary CAInatics Vision AI engine, alongside 3N, its Large Vision Modal Model (LVMM), to analyse images of tobacco bales from different sources and generate grading results.

The output will be delivered to graders through a mobile application, allowing assessments to be made in real-time.

The system will evaluate visual characteristics including colour, texture or grain, and visible leaf condition, generating individual parameter scores as well as an overall grade.

“This is India-built deep technology working inside a real marketplace, helping create greater transparency, consistency and confidence for farmers, buyers and the wider tobacco ecosystem. For us, this is also part of a larger vision: to build AI that does not remain confined to controlled environments, but becomes a cognitive vision layer for the real world, helping industries, institutions and communities see better, decide faster and act with greater confidence,” said Amit Srivastava, Constems AI Co-Founder and CTO.

The Secretary of the Tobacco Board of India said, “ AI-based grading of tobacco would reduce manual intervention and automate the only existing manual operation in the e-auction system. Successful implementation of AI-assisted grading is expected to enhance transparency, consistency, and efficiency in tobacco classification at auction platforms and enhance the confidence of both growers and buyers.”