The AI-Native Agent: Douglas Elliman Builds an AI Data Company, and Why 'Using AI' Isn't Making Agents Faster (July 2026)

Shorter list this week, but the three stories all point at the same thing. Who owns the data, who actually gets faster from AI, and which parts of the job a human still has to do.

Every week I round up the AI-in-real-estate stories that actually matter to a working agent, with my honest take on each. Here’s what caught my eye.

1. A luxury brokerage just built an AI company to own its data

Douglas Elliman launched Elius on July 8, a separate AI intelligence company built with Google Cloud. The plan is to take decades of its proprietary luxury transaction data and turn it into pricing intelligence, buyer targeting, and new products it will eventually sell to customers outside the brokerage. The data stays Elliman-owned.

My take: Seeing a luxury brokerage at this level come out and say their data is one of their most important assets going forward tells you where this is headed. AI needs good data to work well, and Elliman is sitting on decades of it.

This is great. The companies that understand their data is the real asset, and move to own it, are the ones AI will actually be able to run on.

2. Everyone adopted AI, almost nobody got more productive

Recent industry data lays the gap out in plain numbers. 82 percent of agents now write listings with AI, up from 58 percent in 2024, but only about 28 percent say AI has made them significantly more productive. Adoption is nearly universal. The real gains still concentrate in a small group of power users, and most agents are producing about what they did two years ago.

My take: This is important to understand. Feeling productive from AI is not the same as being more productive because of AI.

In my own business, the ROI shows up when I point AI at the work that actually gets me clients. Making marketing assets, making better content, making more of it. That is where the biggest return is right now. The agents pulling ahead aren’t the ones using AI the most, they’re the ones aiming it at what makes money.

3. What AI actually does to the loan officer’s job

A wave of AI-native mortgage shops kept rolling out automation this month. reAlpha Mortgage enhanced its platform on July 21 to centralize borrower inquiries and automate routine follow-up before handing off to a licensed loan officer, and newer players like Ralo are claiming AI loan officers that can close in around 15 days. The pitch is that most of the loan process no longer needs a person.

My take: I hear the same argument from both sides. The AI-native mortgage shops say AI can do everything a loan officer does and replace them outright. The experienced loan officers point to the edge cases and the complex deals and say AI will never replace them.

My take is it lands in the middle. Most of the repeatable work gets automated and handled by AI, and the complex cases still need a human. That same model moves straight over to the buy side and sell side of real estate. And it comes back to the first story: the data you own is what makes any of it work.

---

The throughline

Put the three together and the picture is simple. Your data is becoming the most valuable thing you own, the return comes from pointing AI at the work that actually makes money, and the durable job is the complex, human part of the deal while the repeatable work gets automated.

That is what being AI-native actually means. Own your data, aim your AI at clients, and let it run the rest.

That’s the whole reason I do this every week.