The Agency’s Zane Burnett: AI success starts with clean data, not hype

Editor’s note: This interview has been edited for length and clarity.

Jonathan Delozier: Where is AI making the biggest difference for The Agency right now — lead generation, marketing, transaction management, etc. — and what measurable results are you seeing?

Zane Burnett: That’s a good question. Right now, I would say that where AI is making the biggest impact is on operational efficiency, which is oftentimes not necessarily the first place people are looking. They want to see the immediate dollars and cents ROI. We’ve taken a pretty deliberate approach to how we implement AI in the sense that from the very start we didn’t look at it as something that we just needed to deploy. It is something that required an operational redesign of how we were going to implement it.

That started out with what most people think of as boring and monotonous work, which is laying the foundation to be able to become an AI-enabled organization. There are a lot of fly-by-night ChatGPT tools out there right now. We’ve turned our head to all of that, and we focus on setting up good data, taking in all of our data sources and making sure we have clean data to layer AI into. 

It’s really helped the company out in things like data and business intelligence, being able to audit workflows and come up with more efficient solutions. That will lead to more efficient lead gen, higher ROI, [and more] but right now the biggest effect has been operational efficiency and being able to do more with less.

Delozier: Beyond the obvious — face-to-face relationships — what’s a less obvious area where humans are still irreplaceable compared to AI?

Burnett: Anything creative is probably my first answer. There’s a quote out there. Ben Affleck is a really vocal voice in Hollywood around AI. He said something that really resonated with me; “Somebody who can do something is a craftsman. Somebody who knows when to stop is an artist.” That really applies for a lot of our creative. We have a brilliant and creative marketing and design team, and while some of us might be using AI to brainstorm, it’s never in place of [our] creative eye.

Something a little less obvious that we’ve run into is as we’ve layered AI into some of these operational efficiencies, we’ve learned that AI is horrible at identifying long strings of numbers and text. An example would be we have some inbound line set up where people can call in and verify things like 10-digit numbers or long strings of characters, and AI starts to get lost in the sauce when you start rattling off digits and character strings.

And of course, compliance — there’s a big push right now for AI to solve transaction management compliance, but it’s never been 100%. It always requires somebody to have eyes on that.

Delozier: Looking at the next three to five years, what are some AI-related skills that agents are going to need to pick up that maybe aren’t jumping out to them right now?

Burnett: I have an interesting response to that, because right now I think every agent is inundated with, “You need to use AI.” There’s a gap between the desire for agents to use AI and the mandate to use AI and the ability to actually execute on that.

That’s largely because agents are too busy to figure out how to go in and connect Claude to a [model context protocol] server and write custom skills. They want something that does it for them — and that makes sense because agents are busy.

I’m going to be a little bit contrarian and say that I don’t know that agents will need to necessarily learn more, because in three to five years the tools will have evolved to cater to the agent’s capability and bandwidth. 

We’re not too far away from custom-built AI solutions that have their own pre-built knowledge base of skills that are there for the agents to just say, “Hey, do this.” The gap between desire and execution and ability to execute keeps shrinking every month or two. It’s not necessarily an increase in an agent’s ability to effectively use AI — it’s an adaptation of the people providing AI tools to work with the agent’s current bandwidth and capabilities.

Delozier: What could fundamentally change about the real estate transaction in that same time period with AI?

Burnett: Let me think about how to say this, because we’re in the business of helping people buy and sell real estate. I think there are some people out there that think you can TurboTax the real estate transaction, and there’s an obvious element of face-to-face required and relationship building that AI and any piece of tech just won’t replace.

I don’t know that there’s going to be huge upheaval in the real estate transaction as we know it. I think there’s going to be an upheaval in people’s perception of what their specific agent is bringing to them in terms of value. The way we counteract that is the same way we’ve been telling everybody for years, which is be an expert at your job.

This is the single largest transaction in a person’s life. As much as that’s repeated, I don’t think it can be overstated how important it is to have somebody there to hold their hand throughout that process. An agent’s relationship-building skills, ability to be an expert on the transaction itself and expert in the industry — that need is never going to go away.

Delozier: What AI initiatives or tools have met your expectations and delivered day-to-day value and which ones haven’t been as useful?

Burnett: Answering the second part of that question first — anything related to compliance and transaction management is touch and go. Every state is different in terms of what their compliance and transaction management needs are. Compliance and transaction management is such a nuanced workflow. I would advise people to be wary of that.

Where I’ve seen a lot of [promise] is any tool for agents who want to run ads or manage their social media. There’s a lot of good brand-building AI solutions out there that are tailored to the individual that have low price points.

On the other end of the spectrum, for business owners, franchise owners or even big box brokerages dealing with massive amounts of data, there are solutions that can handle financial data and help with forecasting, running [profit and loss statements], finding inefficiencies and forecasting market conditions. The same goes for identifying recruiting and retention risks by analyzing the data that brokers have.

Brokers have been hearing for years that they’re sitting on a treasure trove of data.We went through a period where some people were talking about data as the new oil. We’re catching up to ourselves here. Data was and has been one of the single largest assets on the organizational level that brokers have had.

For the last few years, they’ve been trying to figure out what to do with it — like how today they’re hearing about AI and trying to figure out what to do with it. Well, we’ve hit a convergence where this treasure trove of data is now accessible and easy to extrapolate and interpret because we’re in this AI age. We see a lot of brokers do interesting things with the data they’ve been sitting on for years.

Delozier: You mentioned building solutions in-house now. How has that changed the calculus for brokerages?

Burnett: Right now, we’re building stuff in house that traditionally would have required a room full of [developers]. It used to be, “Do we go out and find a solution or do we build it?

If we go out and find something, then we have to decide what we’re willing to bend on in terms of our use cases. If we decide to build something, then there’s an obvious cost to that. 

Right now, we’re saying we’re going to build it, and it’s costing us 90% less to build a very custom solution than it would have two years ago. That’s the most exciting part of it, and that translates directly into agent empowerment and our agents’ ability to do business, as well as our staff.