Alex Liss is vice president of data science and AI at Huge, where he develops ways to make digital products more useful, trustworthy, and measurable. His work spans simulated user research, AI-driven search, and the business economics of generative models.
Liss combines experience in analytics, customer experience, marketing strategy, and machine learning with graduate study in computer science at Georgia Tech, specializing in artificial intelligence. His technical interests include algorithmic personalization, reinforcement learning, and conversational systems.
- Invisible interfaces: Liss argues that AI should reduce friction instead of adding unreliable chatbots and unnecessary complexity. His approach to AI-assisted experience design treats trust as a consequence of clear interfaces that help people accomplish actual tasks.
- Intelligent Twins: Using Huge’s Live audience-data platform, Liss develops simulated personas grounded in demographic, psychographic, and contextual information. Intent mapping translates audience characteristics into goals; computer-use and computer-vision models then test whether interfaces support those goals. In an experimental audit of global sports websites, simulated casual fans and enthusiasts exposed friction in navigation, content architecture, and engagement. He emphasizes reproducible instructions, controlled comparisons, and human oversight, positioning simulation as a complement to human research.
- Generative Engine Optimization: Liss argues that websites now serve human visitors and AI systems that retrieve, interpret, and recommend information. His writing on AI-ready websites calls for clearer content architecture, structured information, and measurement centered on completed tasks. His work on AI-mediated search explores how organizations can remain discoverable as language models reshape search.
- Value Per Token: Liss measures AI workflows against business outcomes instead of token consumption, output volume, or production speed alone. As automation makes routine production cheaper, he places greater importance on contextual judgment, precise problem definition, and whether a system delivers meaningful customer value.
