Proxet was proud to sponsor and participate in a recent webinar hosted by Realcomm, Before the Bots: Structuring Your CRE Data for an Agentic Future. The event brought together a great lineup of industry experts to discuss how commercial real estate firms can build the essential data foundation required for the new age of AI-driven agents. The consensus was clear: the quality of the results you get from AI is decided entirely by the quality of the data you feed it.
The entire session was full of useful insights, offering practical advice on everything from data governance to organizational strategy to data architecture. One of the most vital points — and a clear plan for the future — was made by Gabriel Safar, Proxet’s Head of Real Estate Solutions. He pulled the discussion into a definitive framework, stressing that a thoughtful approach to your data is not merely through technology improvement, but a business requirement closely tied to company goals.
Data maturity is at the core of AI
Gabriel's presentation focused on the concept of data maturity — and specifically a level that organizations must reach to successfully use AI agents. This is one of the most important topics for CRE: agentic AI doesn't work well on messy data; it needs a validated foundation of information coordinated across all systems.
He further emphasized that reaching the "Advanced Stage" of data maturity is vital to move past simple reporting and gain predictive and prescriptive abilities. This maturity is not just about tools; it's a measure covering four areas:
- People: Having the right roles and abilities to manage and interpret data
- Processes: Setting up workflows to move, validate, and standardize information
- Standards: Defining clear terms, business logic, and definitions, e.g., "What is a property?"
- Technology: Using the right applications and platforms for collecting, storing, and processing
Simply put, if your firm still finds itself asking, "What happened?" you are in the "Reactive" or "Foundational" stages. The goal is to move to "Advanced," where you can ask, "What will happen?" and have an AI agent take action on the prediction.
To understand the full progression of these data maturity stages in detail, check out our real estate data best practices whitepaper: Modern data practices for real estate.
Data architecture: the conceptual data conveyor belt
To reach the advanced stage of data maturity discussed above, Proxet outlined the structure of a strong data foundation — a key process that changes raw data into useful knowledge and usable applications:
- Collect: Getting data from all necessary systems (e.g., Yardi, Salesforce, building IoT systems, files, external data, etc.)
- Correct: Validating and cleaning the information
- Normalize: Adding the semantic layer, terms, and business logic to standardize and match the data, creating a single "Golden copy" that truly represents your business
- Capitalize: Running functions and deploying AI agents on top of the trusted, normalized data to deliver value (e.g., predictive maintenance, risk scoring, automated reporting)
This setup ensures that the output from your AI — your data products — is believable and reliable.
The path to ROI: connecting AI to business goals
The biggest mistake organizations make is putting in place technology without aligning it to a clear business objective. Gabriel stressed that the way to get funding and resources is to tie AI projects directly to the executive company goals.
He provided a way to decide where to start:
- Define corporate goals: These must be related to business value, not technology (e.g., "Be the highest-rated multi-family owner for customer satisfaction," or "Double the next fund size through high-quality investor reporting").
- Inventory use cases: List specific AI actions that will directly affect those goals (e.g., automating deal memo review, predicting tenant default risk).
- Plot impact vs. feasibility: Map the use cases on a chart. Here's where you can identify projects to prioritize.
Crucially, you only need to build out the data processes (and infrastructure) to the extent needed to execute your high-priority use case(s). This avoids huge, multi-year projects and brings results faster, building support across the organization through clear success. We dive deeper into how this use case-driven approach avoids huge, multi-year projects in our recent blog post.
A special thank you to our co-speakers
We sincerely thank all the panelists for their excellent knowledge and deep commitment to improving the commercial real estate industry: Betsy Reed (CBRE), Kevin Stofman (Cherre), Justin Segal (Boxer Property), Andrew Weakland (W.P. Carey), and Stuart Appley (Bay Tech Advisors).
Where to go from here
The main advice from all speakers was shared: you must begin now. Avoid the trap of perfect data compliance; instead, engage your organization by providing meaningful, data-driven value that matches what they care about.