Walk into any commercial real estate (CRE) conference today and one topic comes up in almost every conversation: AI.
Ask most brokers what AI actually means for their business, and the answer gets vague fast. Maybe they tried ChatGPT to summarize a lease or clean up a draft memo. Maybe they tested a tool that looked good in a demo but never made it into a real workflow.
That pattern isn’t unique to a single firm. According to JLL, 88-92% of investors, owners, and occupiers are currently piloting AI. Yet only about 5% of CRE firms report achieving all of their AI program goals, per AI Daily. Most of the industry still sits in the gap between experimenting with AI and building workflows that create real value.
The real question is how CRE professionals can close the gap between AI experimentation and AI adoption that improves how they work.
The Adoption Curve: Where CRE Actually Stands
The 2026 ZIPDO Education Report found that 72% of commercial real estate companies have adopted AI technologies, with growing integration into core workflows like lease abstraction and analytics. The global AI market in CRE is projected to reach approximately $4.5 billion by 2027, growing at a 25.4% compound annual rate from 2022. In the U.S. specifically, the market was valued at roughly $1.1 billion in 2023 and is forecast to reach $3.8 billion by 2030.
Those are big numbers, but day-to-day AI adoption in CRE is far less uniform. AI adoption in CRE is happening at very different speeds depending on asset class, firm size, and willingness to change workflows.. Some investors are pasting property data into general-purpose tools to see what surfaces. Others are deploying purpose-built platforms for lease abstraction, market analytics, or deal pipeline management. A smaller group is building custom agentic workflows on top of models like Claude or Gemini.
The common thread is that firms extracting real value from AI are treating it as infrastructure, not as a standalone feature. The ones stuck in pilot mode are still waiting for the tools to do the work on their behalf.
What AI Actually Does Well in CRE (And What It Doesn’t)
The most common misconception in the industry is that AI understands deal context out of the gate. It doesn’t. A better way to think about AI is as a new analyst on day one.. iIt is capable, fast, and able to process far more data than any human, but it needs training, context, and review before you can trust its outputs on anything consequential.
Where AI is already delivering measurable results in several commercial real estate workflows:
- Property valuation speed: AI tools have been shown to reduce property valuation time by up to 70 percent through automation and pattern recognition across comparable data sets. (ZIPDO 2026)
- Lease abstraction: 58 percent of property managers used AI for lease abstraction in 2023, making it one of the earliest and most widely adopted use cases in the industry. (ZIPDO 2026)
- Investment analytics: 41 percent of CRE investment firms had integrated AI analytics platforms by 2024. These tools help surface markets and deal opportunities that manual research would likely miss. (ZIPDO 2026)
- Process automation: Morgan Stanley analysis cited by Forbes suggests approximately 37 percent of CRE processes can be automated by AI, representing billions in potential efficiency gains across the sector.
AI still requires significant human oversight in complex lease negotiations, multi-tenant deal structures, and any analysis that depends on hyperlocal market knowledge that isn’t well represented in the model’s training data. The technology is improving quickly, but for decisions involving millions of dollars, human review remains non-negotiable.
The practical takeaway: deploy AI aggressively in high-volume, repeatable workflows and cautiously in high-stakes ones. Keep humans firmly in the loop to make the high-stakes calls.
The Mental Shift Most CRE Professionals Haven’t Made Yet
The most productive question to ask when approaching AI adoption isn’t “how do I use this?” It’s “what specific problem am I trying to solve, and what would it be worth to fix it?”
That reframing matters because the “how do I use AI” question leads to chasing tools instead of solving workflow problems. You end up with a half-dozen subscriptions, a few promising demos, and no real workflow change. The problem-first approach forces you to get specific: Which part of the deal pipeline is slow, repetitive, or error-prone? Where is manual data entry creating errors? What research are you not doing because it takes too long?
The second shift is about scope. A common failure in AI implementation is trying to automate an entire role instead of a specific task. The firms seeing returns are breaking workflows into smaller, discrete tasks and automating one step at a time. AI is excellent at unlocking data and making it actionable. Connecting those insights to downstream decisions is the next layer, and it’s still evolving.
Model improvements are released on a monthly cadence right now, not every few years. What you can’t automate cleanly today may become practical sooner than most firms expect. That doesn’t mean waiting. It means starting with something small and building from there.
Reframe: From Artificial to Abundant Intelligence
At Crexi, we have started using the phrase “Abundant Intelligence” to describe what AI actually makes possible. The old model of deep market insight required a full research team and weeks of lead time. That cost shaped which opportunities firms could realistically pursue. Abundant Intelligence means any broker or investor on a platform can access that same depth of analysis, in real time without waiting on a full research process.
Agentic AI: The Next Phase of CRE Dealmaking
Basic AI assistance, pulling a summary or running a comp, is already table stakes for firms paying attention. The more significant shift is agentic AI, which refers to systems that do not just surface information but act on it. In commercial real estate, that looks like tools that pull from offering memorandums, research reports, and archived deal communications to produce structured outputs with far less manual assembly.
For brokers, the practical impact spans the full deal lifecycle. Creating a broker opinion of value (BOV) that previously took two days can now be drafted in a few hours. Generating an OM, tracking letter of intent (LOI) terms, and flagging deal milestones are all areas where agentic tools are compressing timelines. In commercial real estate, delays cankill deals. Anything that moves a transaction forward faster has direct value.
General-purpose models like Claude and Gemini give practitioners a strong foundation to build on. But building a production-ready workflow on top of a general-purpose model typically takes weeks to months of integration work. Most CRE professionals don’t have that bandwidth, nor should they need it. The more practical path for most firms is using purpose-built CRE platforms that handle the integration layer and support CRE-specific workflows.
For proptech and software companies, agentic AI is forcing a critical question: where does your value actually live? Firms whose value is primarily in workflow design face real competitive pressure, because agentic tools can replicate a lot of that. Firms with proprietary, integrated data are in a different position. The companies that crack customer access, data integration, and reliable workflow execution together are positioned to become the foundational infrastructure of the next era of CRE technology.
What AI-Fluent Looks Like in Three to Five Years
The professionals who are building AI fluency now aren’t doing it as a side project. They’re treating it the same way the industry treated CRM adoption a decade ago: a foundational capability that eventually becomes expected of everyone. By 2028, the gap between AI-enabled workflows and traditional workflows will be visible in deal velocity, client retention, and earnings.
In practical terms, top-performing brokers and investors in three to five years will likely use AI to handle recurring work such as market screening, document review, comp analysis, and pipeline tracking, while they focus on the things that actually drive revenue: relationships, negotiation, and complex judgment calls. Transaction timelines will compress. Firms that can move faster will win more business.
The cost of sitting out isn’t abstract. It shows up in lost listings, slower closes, and clients who find brokers who can deliver more in less time. The window to get ahead of this is real, but it’s not unlimited. Early adopters are spending the time and money right now to figure out what works. Waiting much longer means stepping into a market that has already moved.
Four Ways to Start Using AI More Strategically in CRE
These aren’t theoretical recommendations. They reflect the approaches working for CRE firms that have moved past the pilot stage.
- Identify your highest-volume, lowest-judgment task and automate it first. Lease abstraction, comp research, and market screening are all strong starting points for AI adoption in commercial real estate. Pick the one that consumes the most time for the least strategic value and build from there.
- Set a review standard before you deploy. Decide in advance what “good enough to use” looks like for AI outputs in each workflow. This prevents both over-reliance on unverified outputs and the paralysis of checking everything twice.
- Train the tool on your market, standards, and workflow context. AI gives better results when it understands your market, your standards, and your deal criteria. Time spent upfront on context-setting compounds into faster, more accurate outputs over time. This is the investment most firms skip, and it’s why their results plateau.
- Choose purpose-built CRE platforms over general-purpose tools for production workflows. General-purpose models are excellent for exploration, drafting, and experimentation. For workflows that need to be reliable and repeatable, platforms built around CRE-specific data get consistent results faster and with less maintenance overhead.
The Bottom Line
The data is clear: AI in commercial real estate is no longer experimental. With 72 percent of CRE firms having adopted AI technologies, this looks less like a short-term trend and more like a long-term infrastructure shift. The question isn’t whether to engage with it. The question is whether you’re building the capability now or catching up later.
The tools are imperfect. They still require investment, context, and human oversight. But the same was true of email, CRM software, and online listings platforms, all of which eventually became baseline expectations. AI is on that same trajectory, and it’s moving faster than any of those prior shifts did.
Start with a specific problem. Build one repeatable workflow around it. Review the outputs. The window to get ahead of this is real, but it is narrower than it may seem.
