Your Technology Should Get Better the Longer You Use It

One of my biggest frustrations with real estate software is how much information we put into it and how much work we still have to do ourselves. We enter contacts, record conversations, track transactions, and spend years using a system that still depends on us to catch what went wrong and figure out how to fix it. Having that information stored somewhere is useful, but I want our technology to help us do something with it. If we keep running into the same problem, we should be able to use what happened to improve how we handle it next time. That’s a big part of what we’re building at Incyte and why learning and adapting matter to me.

When I talk about a system learning, I’m talking about reviewing what happened, identifying where something fell short, and using that information to make an improvement. Agents do this all the time. You leave a listing appointment knowing there was a question you could have answered better or a concern you didn’t fully address. You think through it, talk it over with someone, and prepare differently for the next appointment. Experience becomes valuable when you actually use it to change something. I want the technology supporting our agents to benefit from that same kind of review instead of repeating a mistake until someone has time to investigate it.

That’s where I want us to go beyond the automations agents are used to. An automated follow-up can do exactly what you programmed it to do and still deliver a poor experience if the timing, message, or next step is wrong. Unless someone reviews the results, it can keep doing that without anyone realizing there’s a problem. We’re designing our systems to help identify those problems and bring them forward for review. That gives us a better opportunity to understand what needs to change, evaluate an adjustment, and see whether it actually helps. The goal is to make improvement part of how the system operates.

Morgan is a good example because conversations give us a lot to learn from. The different versions of Morgan have specific responsibilities throughout the company, including the version that handles inbound calls and texts. We need to know whether someone got their question answered, reached the right agent, or left the conversation frustrated. A completed interaction doesn’t automatically mean we did a good job. We’re building Morgan to help surface those issues so we can spend more time improving the experience and less time digging through transcripts trying to find out what happened.

For example, if callers keep asking a question that Morgan isn’t handling well, I want us to recognize that pattern before it becomes a recurring complaint. The system should help identify the issue, give us enough context to understand it, and support a change we can review and test. As we connect the different versions of Morgan, the goal is for a useful improvement in one area to inform other areas where it applies. That doesn’t mean every version should respond the same way or have the same responsibilities. It means we should be able to learn from what’s happening across the company. That’s the direction we’re building toward, and I want to be clear that the complete process is still in development.

The same thinking applies to our internal data tools and CMA. When an agent reviews an analysis and makes a correction, there’s a reason behind that change that may help us improve the next report. Maybe a comparable property wasn’t appropriate, an important detail was missed, or an adjustment needed more explanation. I want us to capture that feedback and understand it. A correction still needs to be evaluated before it becomes something the system applies elsewhere, because what makes sense for one property may not make sense for another. The value comes from understanding why the agent made the change and whether it reveals something we can improve.

I’m also very particular about the oversight involved. I don’t let AI delete an email from my own inbox without my approval, and I’m certainly not interested in letting it change how we communicate with clients without someone reviewing what changed. Trust but verify and inspect what you expect are principles I apply to the brokerage, and they apply to our technology too. Our approach is to keep meaningful changes subject to human review. Morgan’s inbound role stays within that responsibility, and our agents still review their market analyses and own the recommendations they give clients. Better technology should help an agent make informed decisions while keeping professional judgment and accountability with the agent.

The quality of the information matters just as much. If the notes going into Console are incomplete or we never hear about an interaction that went poorly, we have less to work with when we try to improve it. That’s why our agents’ involvement is so important to what we’re building. They’re using these tools in real conversations and real transactions, and they see things we need to know about. Their feedback helps us understand where the technology is useful and where it still creates more work than it should. I want them to know that when they bring something to us, it can influence what we build and how it works.

For our agents, the reason to care about any of this is whether it makes their day better and helps them run a stronger business. A follow-up that needs less manual attention, an inquiry handled more effectively, or an analysis that requires less time to prepare can give an agent more room to take care of clients and pursue new business. We have pieces of this working today, and we still have work ahead before everything operates the way I want it to. I’m willing to do that work because I believe the technology we provide should become more useful as we gain experience with it. If you want to see where we are, come spend some time with us, and I’ll walk you through what’s running, what we’re developing, and how our agents are helping us make it better.