AI, Intent & the Future of CRE Search: An Interview With Realmo’s Head of Analytics Ian Arguno

Commercial real estate (CRE) search and analysis is challenging due to the scattered nature of information. Properties, listings, broker descriptions, public records, and other relevant data often reside in different formats, such as databases, PDFs, and conversations with brokers. Inconsistent terminology and the varying relevance of the same building to different users further complicate the process.

Realmo, a commercial real estate intelligence platform, aims to address these challenges by developing an AI-powered assistant called Rey.

Rey is designed to be an AI teammate for CRE searchers. Instead of requiring users to navigate complex databases and understand property categories, Rey takes the user's goal and helps them find, compare, and understand the most suitable properties. By understanding the user's intent, Rey can analyze the available information and present relevant opportunities.

Unlike general-purpose AI models like ChatGPT, Rey is specifically tailored for CRE. It can understand why a user is searching for a particular type of property and translate their requirements into structured search criteria. Rey accesses a vast database of over nine million US properties and more than one million active listings, providing up-to-date information on property characteristics, market signals, and analytical products.

When a user submits a complex request to Rey, the assistant separates mandatory requirements from preferences and normalizes them. It then searches Realmo's inventory and property data, evaluates both structured and descriptive information, ranks candidates, and explains the evidence behind each result. Rey recognizes when available evidence is insufficient for a definitive conclusion and presents the relevant evidence rather than presenting an assumption as fact.

Teaching Rey concepts such as "good location" or "undervalued" involves understanding the user's specific use case, priorities, and constraints. For example, a good location for a restaurant may differ significantly from that of a warehouse. Rey analyzes factors such as visibility, traffic, demographics, parking, and local demand to determine a good location for the user's particular business.

Similarly, understanding terms like "undervalued" requires comparing price, size, income potential, market assumptions, comparable properties, alternative uses, and hidden risks.

Rey can discover properties that conventional filters might miss by searching across various signals, including property descriptions, current use, layout, zoning context, and surrounding property data. Rey labels these discoveries as exact matches and broader alternatives, providing users with a comprehensive view of potential opportunities.