Yash Kanoria
Professor
Columbia Business School
Yash Kanoria is a professor in the Decision, Risk and Operations division at Columbia Business School, where he studies how marketplaces and online platforms are designed. With Amine Allouah, Omar Besbes, Josué Figueroa, and Akshit Kumar, he built ACES, a framework for auditing how AI shopping agents choose products, and found that those choices can reshuffle with each model update.
Testing AI Shoppers at Columbia
"The first GPT model we evaluated was three times more likely to pick the product in the top left than random chance would predict," Kanoria told The Innovator in 2026, describing a product grid in which listing position was the only difference between items. Gemini and Claude models from the same generation showed strong position biases of their own, one favoring the top right and another the top center.
The paper behind those findings, What Is Your AI Agent Buying?, audits agents with ACES across providers and model versions. Agents tended to concentrate demand on a few products, penalized sponsored tags, rewarded platform endorsements, and weighed price, ratings, and reviews very differently from one model to the next. "Companies selling online should be prepared for this kind of unpredictable volatility," he said.
Training at IIT Bombay and Stanford
Kanoria earned a BTech at IIT Bombay in Mumbai, India, and a PhD in electrical engineering at Stanford University in 2012, advised by Andrea Montanari. His honors include the MSOM Young Scholar Prize, the Lanchester Prize, the SIGecom Test of Time Award, an NSF CAREER Award, and a Best OR Paper Award he won as an Amazon Scholar for scalable supply chain algorithms.
At Columbia he teaches Business Analytics in the MBA core and an elective on Digital Marketplaces. His personal site collects his papers on matching markets and online platforms.
Advice for Sellers Facing AI Agents
In the same Innovator interview, he called the seller's answer Generative Engine Optimization and said "anybody who cares about their search rankings now also needs to care about how chatbots and AI tools perceive them." "In the short run, the main lever companies have is the product listing," he added, and he expects separate listings written for people and for specific AI tools.
He also warned that "some second-tier brands may be in for a struggle, because AI tools tend to concentrate on a few brands or products." A student group in his Digital Marketplaces elective found that AI tools favored certain heritage luxury brands while some second-tier luxury labels barely registered.