SimGym: A Framework for A/B Test Simulation in E-Commerce with Traffic-Grounded VLM Agents

Han Li, Vibhor Malik, Zahra Zanjani Foumani, Alberto Castelo, Shuang Xie, Ailin Fan · 2026

Read SimGym: A Framework for A/B Test Simulation in E-Commerce with Traffic-Grounded VLM Agents by Han Li, Vibhor Malik, Zahra Zanjani Foumani, Alberto Castelo, Shuang Xie, Ailin Fan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-19219.txt.

Description

Han Li and 19 colleagues present SimGym, a tool for A/B test simulation on e-commerce storefronts with vision-language model agents in a browser. It combines persona generation from clickstream data, a browser agent with perception, memory, and guardrails, and a protocol comparing simulated shifts with real buyer shifts. Validated on UI theme tests, SimGym agents reach 77 percent directional alignment with add-to-cart shifts in real traffic. It shrinks test cycles to under an hour, from weeks before, enabling rapid tests without exposing real buyers to candidate variants.

Catalog record

CREATORHan Li, Vibhor Malik, Zahra Zanjani Foumani, Alberto Castelo, Shuang Xie, Ailin Fan
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46607
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.19219

Browse the full library · About the Tower of Babel