Chris Horn (Deriv) draws a key distinction: proof of concept asks “Can we build this?” while a prototype asks “Should we build this?” He emphasizes discovery, clear KPIs, and the willingness to stop if value isn’t proven. Deriv used a medallion architecture to get AI-ready, prioritized internal AI before customer-facing use, and fostered an AI-native culture through curiosity, training, and incentives.