FirstTeam and Purlin on avoiding AI mistakes: ‘Some stuff actually creates more work’

FirstTeam Real Estate, a California powerhouse that reported $6.12 billion in 2025 volume across 5,978 transactions — good enough for national ranks of No. 34 and No. 67, respectively — has partnered with Purlin to deploy an artificial intelligence (AI)-powered operating system across the brokerage.

The integration brings together PurlinOS, Purlin Close and Purlin Offer & Negotiate into a single platform for contracts, negotiations, transactions and client communications. Agents can interact via voice, text, email and chatbot.

Lauren Henss, vice president of marketing and strategic initiatives at FirstTeam, and Tim Quirk, chief revenue officer at Purlin, sat down with HousingWire to discuss the state of broker AI adoption. They touched on what distinguishes successful integrations from mere software layers and how brokerages can avoid costly mistakes.

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

Jonathan Delozier: What evidence are both of you seeing across the industry that AI is producing measurable improvements that can actually be seen on an everyday basis?

Tim Quirk: I think it’s incremental depending on who you’re talking to. You’ve got individual real estate agents depending on where they are on the spectrum of AI fluency — [some] not doing anything, a lot dipping their toes in the water when it comes to using it for general content generation. But the real folks, when you take a look at who’s in the upper quadrant, are really able to understand where they can use this to start to automate some of their business process and streamline their operations.

With some of the things we’re doing with FirstTeam and with others, we’re starting to see folks streamlining compliance coordination between contract to close with clients. There’s others on the front-end side automating marketing and lead nurturing, especially when it comes to sphere-of-influence type of things.

But the opportunity more and more, you’ve got a ton of these small software providers that are solving a very small part of the overall transaction, maybe just one very small piece of the puzzle. They might have a great solution for that one tiny problem, but then how do you plug that in? Are you going to go plug in 100 of these to handle your workflow? Probably not.

A lot of the folks I’ve been talking to — if you’re out there looking as a broker, as an agent, on what tools to use — it’s really hard to understand the landscape and understand the good from the mediocre to the poor, simply because everyone has the same story. But the solutions are actually extremely different under the hood.

Lauren Henss: We expect at least 80% of brokerages to undergo digital transformation in the next few years. Digital transformation doesn’t end. To capitalize on Tim’s point, a lot of brokerages or agents have shiny object syndrome. They sit here, they look at something, and they’re like, “Ooh, I want it to do that.”

When you’re incorporating AI into anything, you look at the entire agent life cycle and you need to understand what you want this technology to do. What are the end goals of it? What do you hope to accomplish with it? For me, it’s to help agents close more deals in less time and service their clients better.

A lot of brokerages are sitting here and putting this product and that product together. But do you actually have an integrated platform nucleus that can serve as the hub, the ecosystem, for all of this? That overall experience to close deals in less time every single day that’s centralized — that’s what I feel is missing.

Delozier: What distinguishes a successful AI integration strategy from something that ends up being just another software layer that agents view as an additional step?

Henss: For me, obviously it’s education. This is not something that you can roll out and just expect them to listen to webinars. It is meeting agents where they’re at in terms of any software platform you are introducing, with communication training, on demand, in person, whatever it is.

The key that I have found is you need to use your agents as ambassadors. You need to find some raving fans first, who are part of the process before it even launches. They’re in the software platform. They’re trying to break it, so they feel like they have personal ownership of that success. You’re building it around them.

Based on the amount I’ve invested in this product, when is the agent going to see ROI? Is their productivity increasing? Are they closing more deals? Is the amount of that transaction increasing? I look at several different things across success in terms of what they’re doing, but also [net promoter score] and [customer satisfaction score].

Delozier: What parts of the transaction remain poor candidates for AI and what parts are going to stay that way — where no matter how good the tech gets, humans are going to be irreplaceable?

Henss: It doesn’t matter if you’re doing a transaction. Agents need to be the ones, overall, who are going to thrive, not just survive. We know that there’s a certain amount of the population of agents who do a majority of the deals, and there’s like 1.6 million agents out there right now. I do think that will shrink.

Regardless, if we’re talking about the transaction process, or the initial relationship process or marketing, agents need to be a true adviser to their clients. They need to not be order takers. They need to be five steps ahead. They need to understand their client’s business and what is important to them and offer solutions for that. That’s what’s going to have the cream rise to the top of who succeeds and who doesn’t.

Quirk: The technology is going to, over the next few years, really streamline the processes so that agents can focus more on what their real value is. If you’re going to go buy a property in a location that you’re not familiar with, you can get a lot of information online about it, but you still can’t get the particulars about that particular market, that particular neighborhood, what’s going on within that area. Along with, is it a good investment? Is it a good decision for you?

Most people that are buying and selling don’t do it every day. They need some confidence. They just need someone — a professional — to tell them you’re making the right decision for yourself. That’s literally something so simple that actually goes a long way and that’s not going to be replaced by technology. That’s going to be the human aspect.

With that knowledge, education and fiduciary responsibility of an agent to be able to provide that type of service, that can be supported by all this technology. Part of the technology stack that we provide is what we call Final Offer, which is an offer negotiation platform.

A lot of folks say there’s a lot of debate: “Oh, you can’t replace the agent with negotiation.” At the end of the day, what you can’t replace is what someone is willing to pay for a property if they love it, if they want to raise their children in that neighborhood. They might way overpay just to be able to get into that market. They know it’s a 20-year investment. They don’t really care what it is today, so no data can tell you what that’s going to be. But humans can help with that type of discussion and ultimately that rationalization.

Delozier: For any brokerage considering a major AI implementation, what governance and operational safeguards should be in place before deployment, particularly around data accuracy, privacy and compliance?

Quirk: If you look at the spectrum of where people are in terms of their AI fluency — if you’re using an OpenAI model and starting to either try and create contracts or upload things that have personal information, you’re going to get yourself into trouble very quickly.

With people that aren’t in real estate — if someone’s going to buy a property and they start relying on Claude or ChatGPT for guidance, which unfortunately is happening right now, those platforms aren’t specific. They don’t know about real estate, and so it’s pretty dangerous when it’s going to tell you what you want to hear. All of a sudden you think your house is worth X when it’s really worth Y, and now you’re not willing to take an offer because some system told you not to.

From a brokerage perspective, implementing an AI solution that fundamentally understands real estate rules, regulations, fair housing, and having all that in there is table stakes. You can run tests internally to understand how accurate this thing is so that you have the confidence in it. When it comes to contract compliance, as our clients adopt it and start to see it, all of a sudden they start to realize that it’s actually more accurate than the human when it comes to catching different issues and concerns.

If you can then create a consistent process within the brokerage or within the audit process itself, you can start to use that as a way to enhance the services you provide. It’s not about taking it away from the humans doing the work. It’s just about focusing them on the things they need to focus on. Ultimately, even insurance companies and [errors and omissions] are going to start to look at this stuff and be like, “Oh, you don’t have this type of solution deployed. We’re actually going to charge you more because the human error rate is going to ultimately be higher.”

Delozier: What metrics should brokerage leaders use to determine their return on investment with an an AI platform?

Henss: They need to look at time to value — that is a big thing. You need to look at what the overall adoption rates are, but I’m saying true adoption. I want utilization numbers. I want to know what areas of the AI that they’re using, like the average time per session.

Everybody normally goes and does measurements like, “Oh, what’s the monthly adoption?” I don’t care about the monthly adoption. I care about daily adoption and I also care about what the overall session length is. Where are they operating within the AI? What are the key areas that they’re using most? Because we know that those areas are working.

For example, in our platform, if they’re sitting here and they’re going to the marketing section of AI, they’re using that daily. I have 60% of agents in there daily with an average session time of 20 minutes per day. I would consider that a success because I’d be able to see what products they’re using, what they’re spending the most time on, and understand what’s working and what’s not.

But say they don’t go into their growth area, which shows their projections. They’re not checking their metrics. They’re not doing what they need to do in order to hit their overall goals. I then know if they’re not doing it — and this is with anything at the platform — I need to change that interface, that overall framework, so they do that.

Delozier: If you were advising an independent brokerage that hasn’t really jumped into AI yet, what should they prioritize over the next 12 to 24 months, and what are the most common mistakes they should avoid?

Quirk: Go to resources they trust and actually find someone to give them guidance in terms of where to start. Get an assessment on the current business and their technology ecosystem, what folks are using and what they aren’t, and understand what they’re trying to accomplish. AI for the sake of AI doesn’t do anything.

Don’t just jump in. You see something that looks cool and you just go with it without looking at what else is out there. There’s a lot of stuff that looks awesome right now, but then when you get under the hood, it actually doesn’t do anything. Some stuff actually creates more work. You might get something that does some analysis and sends an email, but then it doesn’t integrate with the rest of your technology stack, and all of a sudden you just created more work outside of your current systems that your team needs to do.

Henss: You need to do an audit of the business. What are your business goals now, but also what are your business goals over the next three years? Is that expansion? Is it agent recruitment? What is your overall retention rate now and why do agents leave your brokerage? You have to have a consultant or you need to have somebody in house who has led digital transformation like this before. Tim and I have worked a lot in digital transformation — so that has helped us significantly in terms of what to do and what not to do.

You need to understand in that audit what’s working and what’s not, because what I’ve learned is that the loudest people in the room are not always your adopters. Most often, they are not. They’re the ones that may not want to. So you really need to understand at a basic level, out of each platform you have, what is working well and what is not working well.

Lastly, you need to have leadership buy-in. You need to understand what the goals of the business are, and you need to have a team that is responsible on both sides — on the vendor side and partnership side, and internally — on measuring, identifying and being lockstep in what success is.