What Real Estate Teams Can Learn from How Tech Uses AI

Tech's AI Experiment: Surprising Results

Software developers have been using AI assistants for several years now. The outcomes aren't what you'd expect. Individual productivity goes up. Code quality improves. But there's a catch: delivery stability actually drops. Rollbacks and emergency fixes become more frequent. That sounds contradictory, doesn't it? How can quality rise while stability falls?

Nathen Harvey, who leads research at DORA, thinks the answer lies in team incentives and visibility. Developers might focus on shipping new features fast, without checking whether those features hold up in production. AI amplifies their speed, but it also amplifies their blind spots.

Why Real Estate Should Care

You might wonder what this has to do with real estate. More than you'd think. Whether you manage residential properties, commercial spaces, or a mixed-use portfolio, you deal with complex systems. Maintenance teams, leasing agents, finance departments, contractors. Legacy processes, like paper-based inspections or outdated property management software. And constant pressure to do more with less.

AI is creeping into real estate too. Automated valuation models, chatbots for tenant queries, predictive maintenance, smart building systems. But just like in software, AI won't fix a broken organization. It might even make things worse.

AI Isn't a Fixer, It's an Amplifier

Harvey calls AI an amplifier. If your systems are solid, AI amplifies the good. If they're messy, AI amplifies the chaos. You see a productivity bump in one corner, but gains get swallowed by bottlenecks downstream.

Say you adopt an AI-powered property inspection tool. It speeds up your inspectors, so they cover more units per day. But if your maintenance ticketing system is a mess, those inspection reports just pile up in a shared drive. No one sees them. Nothing gets fixed. Tenants stay unhappy. The tool didn't help because the system around it was weak.

Seven Team Profiles: A Cautionary Tale

DORA's research looked at hundreds of software teams and grouped them into seven profiles. One profile is called "Process-Constrained." These teams have high burnout, lots of collaboration friction, and low personal efficiency. They spend little time on high-value work. But here's the odd part: their delivery stability is actually pretty good—better than most teams. The catch is their throughput is low. They're stable but sluggish. They're not creating much value.

Does that sound like any real estate teams you know? Maybe a property management group stuck in weekly meetings and paperwork. They get the monthly reports out on time, but they never seem to improve occupancy or tenant satisfaction. They're busy, but not effective.

Seven Capabilities That Matter

DORA also identified seven capabilities that separate teams who get real value from AI from those who don't. These are worth borrowing for real estate:

• Clear and consistent communication about how AI will be used.
• A healthy data ecosystem—your numbers are clean, connected, and trusted.
• AI can actually access the internal data it needs.
• Good version control practices—you know who changed what and when.
• Small batches of work—not giant, risky changes.
• A user-centric mindset—you're building for real people, not just processes.
• A high-quality internal platform—a foundation that handles the complex stuff.

Small Batches Beat Big Bangs

One of the most counterintuitive findings is that AI models tend to generate large changes in one go. They'll produce a massive pull request that touches dozens of files. That's the opposite of what DORA has found works best: small, incremental changes that can be deployed independently.

In real estate, think of a renovation. Instead of closing down an entire wing for a month to do everything at once, tackle it floor by floor. Or unit by unit. You get faster feedback, less disruption to tenants, and you can catch problems early before they cascade.

When you use AI to help plan a renovation, you can ask it to break the work into smaller chunks. The same principle applies to policy changes, new software rollouts, or marketing campaigns. Small batches reduce risk and increase learning.

Platforms: The Hidden Hero

Another key insight is the role of a high-quality internal platform. In software, a platform gives teams access to tools, workflows, and models in a consistent way. But more importantly, it absorbs complexity. If your company has a security policy, the platform enforces it automatically on every new system. Developers don't have to think about it.

For real estate, think about your property management software. A good platform would automatically enforce lease terms, compliance rules, and safety standards. It would handle the boring but critical stuff so your team can focus on actual problems—like why a certain property's occupancy is dropping or how to improve tenant retention.

Without a solid platform, you end up with a patchwork of spreadsheets and emails. People make mistakes. Policies get ignored. And AI tools can't work their magic because they don't have a clean foundation.

Start With a Self-Assessment

DORA's research isn't a prescription. Harvey emphasizes that you can't just copy their recommendations. Every organization has its own context, its own customers, its own history. The findings should be treated as hypotheses to test in your own environment.

So, where do you start? First, take a hard look at your current state. How are your teams structured? What's the culture like? Are people burned out? Is collaboration a struggle? How good is your data? Do you have a clear AI policy, or is it every person for themselves?

Then, pick one or two areas to improve. Don't try to fix everything at once. Maybe you start with cleaning up your data so that eventually, an AI can actually use it. Or maybe you focus on breaking down big projects into smaller ones. The key is to build a culture of continuous improvement. Keep iterating, keep learning, and let the research guide you—but don't let it dictate.

The Human Element

Ultimately, any organization's goal is to serve people. In real estate, that means your tenants, buyers, sellers, investors. AI can help you prototype and test more ideas faster. Instead of just one plan for a new community space, you can develop five and let residents vote with their feet. You can simulate different pricing strategies and see which one actually drives occupancy.

But AI won't tell you what people really need. That requires empathy, conversation, and a willingness to be wrong. The best results come when you combine AI's speed with human judgment. Use AI to explore possibilities, but keep the human in the loop to evaluate, refine, and decide.

So, before you invest in the next shiny AI tool for your real estate business, take a step back. Look at your systems. Fix the bottlenecks. Clean up your data. Make sure your team isn't burning out. Then, and only then, let AI amplify what you've built. That's the real playbook.