Chris Churchman, head of Goldman Sachs' digital platform Marquee and co-chair of the Global Banking & Markets AI Working Group, recently stated in a podcast that once enterprise AI products are launched, the worst outputs will be scrutinized by the market and clients. He emphasized that "roughly correct" AI is unacceptable in high-risk scenarios such as finance. He highlighted that AI productization requires addressing issues such as factual accuracy, computational verifiability, access control, and timely error cessation. He also proposed four productization boundaries: not substituting demos for acceptance testing, not betting on temporary model flaws, redoing workflows, and not outsourcing human judgment.