BuyerIQ
www.buyeriq.ioGary Tan posted on Twitter on a Friday about how LLMs can simulate purchase intent. That night I built BuyerIQ. By Sunday I had seven paying customers.
You paste your website link or describe your product. BuyerIQ runs synthetic buyer simulations across 100+ demographic segments, predicting who will buy and why. The whole process takes under 30 seconds.
The methodology draws on LLMs Reproduce Human Purchase Intent via Semantic Similarity Elicitation of Likert Ratings. The researchers measured how closely simulated survey responses reproduced human purchase-intent ratings. Their benchmark describes the research method; it is not a measured accuracy rate for BuyerIQ. Research details.
BuyerIQ has generated several thousand dollars in revenue and helped several hundred founders with early-stage market validation.
Notes & sources
What the research measured
Maier and colleagues report 90% of human test–retest reliability for semantic similarity rating, using 57 personal-care product surveys with 9,300 human responses. These are the paper’s results, not a measured accuracy rate for BuyerIQ or a test of 9,300 completed purchases. Read the research paper (version 3) →
Seven paying customers in the first weekend
This launch result comes from my first-person account of building BuyerIQ. The retrospective also describes its later revenue and customer growth. My BuyerIQ launch retrospective →
