Podcast: Using AI to Make Better CRE Acquisition & Underwriting Decisions

Steven Song is the founder and CEO of Diald, an AI-powered decision intelligence platform for commercial real estate. Before founding Diald, Steven worked on both the investing and development sides of real estate, and built his career across architecture and urban planning, training at Carnegie Mellon and the University of Pennsylvania. He was a founding principal at SCAAA, a global strategy, planning, and design firm, and is a partner at Axle Companies, a family office focused on real estate investment and social impact ventures. Steven is based in Los Angeles.

(02:26) Why CRE Decisions Are Still Judgment-Driven
(04:41) The Signal That Killed an Atlantic City Deal
(07:44) Contextual Drift: The Risk Nobody Models
(10:20) Diald’s approach
(11:59) AI Token Costs and Asking Better Questions
(14:27) Tools vs. Workflows
(15:52) Diald’s Underwriting
(17:58) Killing Bad Deals Earlier
(19:18) How AI Upgrades the Analyst Role
(20:44) Where General Purpose AI Fails at Underwriting
(24:23) Does AI Make CRE More Efficient or More Competitive
(26:02) What Underwriting Looks Like in 5 Years
(27:14) Where Human Judgment Still Matters
(29:06) The Local Signals Investors Miss
(31:15) Collaboration Superpower: Denise Scott Brown and Reyner Banham