For many real estate brokers, the dream of artificial intelligence (AI)-powered efficiency has crashed against a hard reality; agents are reluctant to use tools that force them to change habits.
Tyler Morton — broker-owner of REMAX Victory + Affiliates in Beavercreek, Ohio — learned this lesson the expensive way.
After his first AI platform Victoria failed because agents refused to download “just another app,” Morton rebuilt from scratch. The result was Victoria’s update, Tori 2.0 — a brokerage operating system that unifies disjointed systems from messaging to documents.
Morton sat down with HousingWire to share hard-won lessons on building an AI infrastructure that agents actually use.
Editor’s note: This interview has been edited for length and clarity.
Jonathan Delozier: You mentioned that the first version of your AI platform failed because agents had to change their behavior and log into another tool. What are the biggest lessons brokers should learn about AI adoption before they invest heavily?
Tyler Morton: We tried to make it as easy as possible. I mean, even with the fact that I initially built the app with the agents in mind, thinking it would make their lives easier, they didn’t want to download just another app. That’s not where their mind goes when they have a pressing question. They still want to call the broker and get a quick answer, or shoot a text, a Facebook message or even an email. That’s where the agent lives.
We tried to make it as easy as an agent texting a question to Victoria, and it would give an instant response. But they just weren’t programmed to do that, right? They’ve been doing business the same way for years.
Delozier: So how did you fix it with Tori 2.0?
Morton: We really tore it all down. I started looking at the platform and all the gaps or holes that it had for other tools that we needed to integrate and tried to think about where agents actually live. I said, ‘Let’s break this down and look at the entire stack that we’re using for all of this stuff. I can probably rebuild 90% of it relatively easily.’ And I say relatively easily with 2,000 hours of vibe coding experience and tens of thousands of dollars. We don’t call them mistakes, we call them tuition, because I certainly learned from it.
Delozier: What does that unified system do for agents day to day?
Morton: One of my most-used functions of Tori OS is an agent goes in and asks a question, and proactively, the system, based on our knowledge base, is answering those questions in stream. If an agent asks, “Hey, what plumber do you recommend in Westchester, Ohio?” it’s going to look at all of our previous conversation history and say that John recommended ABC Plumbing six months ago. It might say, “You may want to check to see how that job went,” or “Carol recommended 123 Plumbing.” It will give them that entire list, and it’s just learning and growing from that.
Agents know what they’re used to, and so that’s what we built. We built the voice-to-contract model with that in mind. What we sought to build was — using Dotloop’s API — a simple voice-to-contract where agents can be driving back from their showing and hit a button in the app and say, “Write me an offer for 123 Main Street for $500,000, two weeks for inspections and close it in 30 days.”
Delozier: You’re running a pilot with Amazon on something even more advanced. What does that look like?
Morton: We are running that through a company [Amazon] acquired called Bee, or bee.computer, using conversational intelligence. Instead of an agent still reacting to a conversation they had earlier with their clients, there’s a bracelet that can listen in real time. Through what we’ve built in our platform, it can understand the context of the conversation — possibly even who they’re with — and know that they’re sitting at the kitchen counter at 125 Main Street and the client wants to write an offer for $560,000. It will go ahead and proactively create that contract for them.
Delozier: What’s the separation right now between practical AI use cases that agents want versus flashy stuff that ends up being more marketing?
Morton: The flashy is the cool. I just did a webinar an hour ago for agents on building their first AI employee. But one of the biggest things that I’ve taught over the last year and a half, two years, is when you don’t know what to ask AI, ask AI what you should be asking. We were building [tech] all the wrong ways — thinking of how we wanted things to be built and not where the agents were.
Delozier: Where should brokers build their own technology versus relying on outside platforms?
Morton: For years we’ve been pigeonholed into buying technology, and now you could have a working concept of a simple tool within 24 hours if you really spent the time doing it. That’s not to say that everyone should, or could, because there is a learning curve. I remember banging my head against the wall because I spent six weeks developing the first version of Victoria — only to realize that I’ve got to pay a developer to get this up and running.
I think a lot of companies are going this way; don’t try to be everything to everyone. Be very narrow and great at what your tool does. I don’t want to reinvent IDX. I don’t want to reinvent e-signatures. I want to be the hub that ties all these different things together, so that users can pick whatever they’re used to using. As soon as an agent or a broker has to learn a new system, that’s where you create the friction that causes lower close rates.
Delozier: Looking at the next five years, what will be the biggest differentiator for brokerages with AI?
Morton: I don’t think it’s even five years. I certainly can’t predict out that far. I don’t even know what the next five months hold. You’ll see more consolidation. The ones that are not adopting AI or looking at the possibilities of agentic AI — they’re not able to streamline and they’re not able to scale. I’m not saying that AI is going to replace X, Y and Z. I just think that the ones that do adopt it will be light years ahead in terms of employees. You literally can scale a larger operation with very minimal employees with the technologies that we have available today.
