The Agentic Frontier: Artificial Intelligence in Real Estate

The real estate sector's complex, multimodal, and spatially dependent data creates challenges that artificial intelligence is unusually well positioned to address. This review distinguishes real estate from general finance across four pillars: spatial dependence, asset heterogeneity, the primacy of unstructured data, and distinct institutional and theoretical foundations. It traces the field's evolution from automated valuation and return-prediction models toward advanced natural language processing, computer vision, and, increasingly, explainable, causal, and agentic systems capable of automating complex workflows. The paper closes with a forward-looking agenda centered on causal inference, multimodal data synthesis, and the ethical governance needed for fair and transparent property markets.