By Adam Baugh
AI is having a moment in commercial real estate—and like most “next big things,” it’s being oversold in some rooms and underutilized in others. For those of us working in land, zoning, and development across Arizona, the reality sits somewhere in the middle: AI is a powerful tool, but it’s not a substitute for judgment, experience, or local knowledge.
Let’s start with what it does well.
AI can meaningfully compress timelines in the due diligence phase. Tasks that used to take days—reviewing title reports, summarizing leases, pulling comps, or scanning municipal code—can now be done in a fraction of the time. In fast-moving submarkets like Phoenix, that matters. Speed can be the difference between getting a deal under contract or explaining to your client why someone else did.
It also improves the front end of decision-making. AI can ingest large datasets—demographics, absorption trends, infrastructure plans—and surface patterns that might otherwise go unnoticed. That’s useful in site selection and early underwriting, where you’re trying to answer a basic question: “Is this worth chasing?” The ability to run multiple scenarios quickly—adjusting rents, costs, or exit assumptions—adds another layer of discipline to the process.
And there’s a cost component. By automating repetitive work, AI can reduce reliance on outside consultants for initial screening and analysis. That doesn’t eliminate the need for lawyers, engineers, or planners—but it does allow teams to focus those resources where they add the most value. For smaller developers working in places like Mesa or Glendale, that can help level the playing field.
But let’s be clear about the limitations.
AI is only as good as the data behind it. If the inputs are incomplete, outdated, or just wrong, the output will be too—only faster and with more confidence. That’s a real risk in Arizona, where entitlement processes, zoning interpretations, and even political dynamics can vary significantly from one jurisdiction to the next.
More importantly, AI struggles with nuance. It doesn’t sit in pre-application meetings. It doesn’t read the room at a neighborhood meeting. It doesn’t know when a project that works on paper is going to run into resistance because of traffic concerns, school capacity, or a council member’s priorities. Those are the realities that shape outcomes here, and they don’t show up cleanly in a dataset.
There are also legal guardrails. AI can flag issues in contracts or title documents, but it doesn’t replace legal analysis. Real estate deals often hinge on small details and context—things that require interpretation, not just pattern recognition.
The bigger concern, frankly, is over-reliance. There’s a tendency to treat AI outputs as answers instead of inputs. That’s a mistake. In land and development, a bad assumption early in the process doesn’t just stay on a spreadsheet—it gets capitalized into the deal.
The takeaway is straightforward. AI is a force multiplier. It can make good teams faster and more informed. It can help you screen more opportunities and spend time where it counts. But it won’t entitle your project, it won’t build community support, and it won’t make the hard calls for you.
In Arizona, where local context drives so much of the outcome, the winners won’t be the ones who rely on AI the most—they’ll be the ones who use it well.