The AI framework for facilities and real estate from OpenAI + OfficeSpace

Bots that take action on your behalf. Hours of research condensed into minutes. Labor-intensive tasks on autopilot.

AI in the workplace has moved from concept to reality, and teams are beginning to act. No longer confined to research labs and IT departments, it’s now in the hands of workplace teams thanks to tools like ChatGPT.

Although these tools are evolving fast, not every industry has embraced the shift. 

Why the AI adoption curve is steep for facilities and real estate

In real estate, structural barriers make AI adoption harder. Long building lifecycles, budget constraints, and complex stakeholder relationships create friction. 

“What we’re seeing across the board is that AI adoption in workplace management is still early, but picking up momentum. Most teams are in the ‘crawl’ or ‘walk’ phase: running small pilots, testing tools, and figuring out how to make it all work with the data they have.” — Andres Avalos

Additionally, organizations across industries feel they lack sufficient data to make AI tools effective. Only 14% of Commercial Real Estate (CRE) leaders say their data is ready to generate results, according to Deloitte. “A lot of teams have data from booking systems, badges, sensors, but it’s messy, siloed, or just not usable yet,” shares Andres. “In some orgs, people don’t even have permission to use AI tools yet, let alone budget or support to roll them out,” he adds.

And one of the largest barriers to leveraging AI? Simply knowing where to start.

In our webinar with OpenAI, in partnership with IFMA, OfficeSpace Software’s Chief Product Officer Andres Avalos spoke with OpenAI’s Head of Strategic Occupancy Planning Zig Wu and Workplace Data Engineer Jeremy Schreiner to explore tangible ways facilities and real estate teams can start their journey with AI.

Read highlights from the conversation, including a primer on the AI agents you’ve been hearing so much about, how teams can use AI to drive efficiency and support strategic decision-making, and key steps to make your AI rollout achievable and successful. 

Generative AI vs. agentic AI: A quick primer

Most teams are familiar with generative AI: tools like ChatGPT that create original content based on natural language prompts. AI agents, on the other hand, are systems that can act autonomously to perform tasks on your behalf. 

“The best analogy is: Generative AI is the brain and agents are the arms and legs that go out and interact with the world.” — Jeremy Schreiner 

Not your grandma’s chatbots

Though “chatbots” have existed since the 1960s, today’s chatbots go far beyond simulating conversation. They’re what Schreiner calls the ‘mouthpiece’ of the agent, interfacing with users while automating behind the scenes. Employees can delegate work to AI agents just by providing instructions in plain language; they’re capable of tasks like entering JIRA tickets and sending emails without your oversight. 

By combining both generative and agentic AI, workplace teams can take on more strategic work and reduce time spent on manual, repetitive tasks.

AI use cases for facilities and real estate teams

Challenge: Space supply vs. demand

“There’s often a lag from the point of time we decide to acquire real estate and fit out space, while headcount changes happen in much shorter timeframes,” shares Zig Wu. Predicting space demand and aligning spaces with employee behavior is a challenge for workplaces across industries, especially as in-office policies continually change, making attendance unpredictable. 

When space is insufficient or doesn’t match how people work, collaboration often declines.

Challenge: Office moves

Without clear communication and coordination, workplace experience suffers during relocations. Employees feel disoriented, costs run high, and facilities teams are left managing complex logistics.

These processes are ideal use cases for AI. The OpenAI team uses AI to analyze headcount growth, forecast space needs, and dynamically adjust layouts based on collaboration patterns.

They’ve also built custom GPTs to simplify move communications. These “virtual assistants” require no coding, can be trained on your internal documentation, remember context, and be saved or shared with teams for ongoing use. For Zig’s example, he created a Comms GPT that automatically writes and sends employees messages in the communications director’s tone of voice, saving 90% of the time previously spent on this type of communication.

At OfficeSpace, custom GPTs support internal knowledge-sharing and product education. For example, our product teams can upload research, meeting notes, and feature benefits to reduce repetitive questions and ensure consistent messaging across teams.

The crawl-walk-run framework for successful AI adoption

To help teams adopt AI in a manageable way, OpenAI and OfficeSpace developed a simple framework.

Crawl: Start small, get comfortable.
Zig recommends experimenting with tools like custom GPTs or ChatGPT’s Deep Research feature to get familiar with AI. This will help you understand the tools’ capabilities and better identify a few time-consuming tasks that feel unnecessarily manual in your workflow. 

Next, audit your data sources and develop relationships with data owners to build a foundation for experimentation.

“Bring us in early,” says Jeremy Schreiner on working with data engineers in your organization. “Introduce us to the problem you’re trying to solve, rather than the solution you’re looking for.”

While many teams are becoming more data-driven, critical workplace data is still scattered across systems, from desk booking to CAFM to HRIS. Making sure data is accurate, centralized, and up-to-date can make AI pilots more successful. 

Walk: Test and demonstrate quick wins.
Use pilots to build momentum with stakeholders. What worked for McKesson’s Director of Real Estate Operations Technology was demonstrating to executives the real, practical applications of AI in real estate. For their team, it was an AI agent that could forecast headcounts and decide which projects to prioritize. 

Other examples include a Slack bot to automatically create tickets when an employee asks for help in Slack. Within ChatGPT, you can even explore specific GPT use cases for your industry and role.

Regardless of the systems you use to develop your AI projects, you want to make sure you’re using tools from trusted vendors who are mindful of data governance and security. 

Run: Make AI part of how you work.
Once you’ve centralized your workplace data from facilities, IT, HR, and finance, you can make real-time, AI-powered decisions, broaden use cases, and continue to iterate and evolve.

For example, Zig and Jeremy are working on developing a bot that combines utilization and seating capacity insights from OfficeSpace with headcount forecasts to proactively alert the team before they run out of space.

“In this stage, AI shifts from being an experiment to a strategic partner in running your workplace,” — Zig Wu

6 steps to start your AI journey:

• Identify your top constraints
• Map out your data and tech partners
• Understand your organization’s AI guardrails
• Consolidate and clean your available data
• Build your first GPT or agent to solve a real task
• Expand to cross-functional use cases tied to business goals

Learn how to leverage custom GPTs and Deep Research, discover AI experts, and expand your AI knowledge with our AI Field Guide for FMs.  

What’s next: The future of AI in the workplace

In the not-so-distant future, and within the OfficeSpace platform, AI agents will be able to:

• Suggest floor plans based on usage trends
• Manage energy consumption based on in-office attendance
• Coordinate office days based on collaboration patterns
  

These capabilities will deliver measurable cost savings, enable smarter planning, and create more personalized workplace experiences—all of which support talent strategy and long-term space optimization.

The takeaway: With the right mindset, the right strategy, and the right tools, AI adoption can be fast, practical, and transformational.

Want to hear more from OpenAI and OfficeSpace?
Join us live at IFMA World Workplace in Minneapolis, MN, on September 19 where Zig and Andres will dive deeper into how high-quality data powers successful AI strategies. Learn more.