Commercial real estate firms are pouring millions into artificial intelligence tools despite minimal returns on those investments. The industry sits in an awkward phase where experimentation dominates strategy, yet actual operational change remains elusive.
CRE companies justify these expenditures on AI's theoretical potential. Vendors pitch machine learning for property valuation, lease analysis, tenant screening, and market forecasting. Large brokerage firms, asset managers, and institutional investors have all launched pilot programs. Some have hired dedicated AI teams. Yet most honest assessments reveal that existing implementations have not fundamentally altered how deals get done, how portfolios get managed, or how properties get valued.
The disconnect stems from reality meeting hype. AI works best with clean, standardized data. Commercial real estate runs on fragmented systems, legacy contracts, and highly localized markets. A building in Manhattan operates under different rules than one in Des Moines. Commercial leases vary dramatically in structure and terms. Property databases lack consistency. These real-world complications slow AI adoption far beyond what cheerleaders anticipated.
Fear of missing out drives spending nonetheless. If competitors invest in AI, firms worry about falling behind. Marketing departments leverage AI announcements to signal innovation to clients and investors. Technology budgets expand because executives expect breakthroughs, even when evidence of transformation remains thin.
For landlords, tenants, and investors, this spending spree creates false expectations. Brokers market "AI-powered" market analysis and valuations that rely on the same underlying human judgment as before. Property managers promise AI-driven efficiency gains that materialize slowly, if at all. Lenders and equity investors hear AI pitches from borrowers seeking to appear cutting-edge.
The honest outcome: CRE firms will continue experimenting. Some use cases will eventually prove valuable. Portfolio analytics, predictive maintenance, and anomaly detection show promise. But the transformation narrative oversells current capabilities