6 out of 10 buyers start their search by asking Google something like ‘what’s for sale in Greensboro’ or ‘homes in Charlotte’. And 6 out of 10 of those answers are now AI-generated.
The old way of getting found there was paying Realtor.com for a lead package, or posting your listings to Zillow and hoping for the best. Either way, you were renting access to your own future clients from someone else’s platform.
This month, we stopped renting and we built our own. A chat window that takes a question like ‘what’s for sale in Greensboro’ and answers it with our own fresh, filtered, real MLS data — the same way Google answers a question now, but it’s our data and our conversation, working for us 24/7, automatically, while we sleep.
And here’s the part nobody tells agents: you don’t need to be a developer to build it. You just need to know the right question to ask, and the right way to ask it. Today we’re going to show you exactly what that question is, what it’s actually doing behind the scenes, and how you can put the same thing on your own website, working for you, in 45 days or less.
Related: Claude Code for Real Estate Agents — Our Stack
Related: Claude Code for Real Estate Agents — Our Stack
Who This Is For
If you’re a few years into this business — closing 6, 7, 8 deals a year, pushing for more volume, more consistency — you’ve probably already built or bought some kind of helper: a follow-up sequence, a report, a posting schedule.
The question then becomes: does that helper actually know anything about my market, or is it just a generic script I pasted in?
Either way, it’s sitting on my machine. If it breaks, I’m the one who has to open it and fix it.
We’re going to show you a different way to build those helpers — they run on a cloud computer that’s already paid for, and they stay connected to your real MLS data every single day, automatically.
If you’re closing deals and you want to build something like this yourself, watch the whole video — we’re giving away everything we made, including the exact setup we use in our own business.
What It Actually Does
Here’s the mistake we see agents make: they ask ChatGPT or Claude something like, what’s a good lead follow-up email, and the AI writes back a generic, let me tell you about our company, message that it found by looking at how other agents phrase their emails. It’s not actually using your MLS data to say, this buyer looked at 4-bedroom homes, so here’s what you should emphasize in your reply. And it’s not the same as asking it to draft a follow-up on a specific client you’re already working with — that it can do, because that’s a real, concrete question with a real, concrete answer.
Where it gets confusing is that both tools give you the same kind of answer when you ask the same vague question. It’s when you start asking your own MLS, your own leads, your own numbers, that they start handling it differently.
With ChatGPT, you get what you put in. If you give it a plain list of MLS numbers and ask what you should lead with, it will give you a plain list of features to highlight — bigger is better, more bedrooms, more baths, lower price per square foot. It’s doing the work of reading through a set of listings and pulling out the obvious signals, but it’s not doing anything fancy with them. It’s not looking at what’s actually selling in your market this week, or how your own past clients have responded to different listings. It’s just giving you what you gave it, formatted as a bullet list.
Claude, on the other hand, is set up so you can throw in a market snapshot — this is what’s selling, this is what’s sitting, this is what people are asking about — and ask it to read that and tell you what that actually means for your business. What’s the real opportunity here, not just what’s the biggest house on the list. And it can do that with your own MLS data, your own leads, your own numbers — you’re not depending on it having a pre-built understanding of your market, you’re showing it what’s actually happening and asking it to tell you what to do next.
So the real difference isn’t that one tool is smarter than the other. It’s that ChatGPT gives you back whatever you put in, while Claude is set up so you can actually ask it to make sense of your own business, instead of just regurgitating what it’s been told before.
How You Set It Up
Here’s the honest setup. We run both Claude and ChatGPT on our own machines, on our own schedules, and we built the connection to Realtor.com’s MLS data the same way with both. Then we did the exact same thing with ChatGPT, in plain English, with the exact same instruction. Same job, same description, given to both, in our own words, with no developer in the loop.
How It Gets You Found
Here’s the real question, then, not which one’s smarter, but which one gets you found — and the answer is quiet, because it’s not about what it says, it’s about what it writes every single day, automatically, on the 3 to 5 websites most agents share leads from.
Both can write a clean, correct, human-sounding description from a set of MLS facts — title, bed-bath, square footage, lot size, year built, list of features, right image on it, no spam, no license numbers, no URLs.
The difference shows up in the 3 to 5 places most leads come from today, not in a chat window.
On Realtor.com, both get you found — both check boxes get filled, both show up in the same way, with the same information, to the same user, at the same time, side by side, swapping places depending on which town the search is in.
On a branded agent website, built with one of the popular systems, both show up the same way — a clean, correctly-formatted listing, same as the MLS feed, checking the same box as the paid lead platform, doing the same job as the $500-a-month subscription service you’ve already paid for.
Where it matters, they both get you found.
Where it doesn’t matter, they handle it differently.
ChatGPT will write you a blog post, every day, on a schedule, on whatever topic you tell it to write about — this week’s new listing, last week’s market update, tomorrow’s neighborhood guide, next Monday’s FAQ.
The post comes out ready to publish, with a headline, an intro, a body, a closing, and an image description, and it goes out through whatever publishing channel you tell it to use — your website’s blog, a specific page on your site, even a shared Google Doc you’ve told it to write into.
How to Verify It’s Working
If you’re building a brand around a specific market, this means you can ask both engines the same question — which source gave you the right, current MLS data — and watch them arrive at the same answer independently, or watch one hand you a correct answer while the other reads back a old, incorrect version.
If your business runs on relationships — repeat clients, referrals, the people who already know you — this is how you verify that the tool you’ve built to stay top-of-mind isn’t posting outdated or incorrect information that erodes trust.
And if you’re scaling, adding agents or building a team, this is how you set a baseline for what a local market guide actually is, so everyone on your team is building from the same map instead of 5 different, slightly-off versions of the same county.
What It Frees You Up To Do
You’re essentially building a private, brand-specific version of ChatGPT — own the AI, own the brand, own the relationship.
If you’re pushing for volume, every repeat question becomes a new training example, and the same query gets the same right answer next time, instead of a suggestion that needs a human check.
If your business runs on relationships, every client or agent becomes a separate knowledge base, and the same person query always pulls from the right history, instead of a generic suggestion that doesn’t know who they are.
Either way, it’s the same core idea: you stop asking a chatbot and you start telling one, and the difference is real because the data never leaves your own machine.
If you want to see this running live — the actual Claude code, on screen, talking to an MLS feed — book a free 15-minute call with us.
The link is in the description. And there’s a free starter guide in the description too, so you can see the first steps before you ever get on the phone with us.
Subscribe so you don’t miss the next live session — we build one of these, start to finish, on screen every single month.
We’re Al and Victoria Pinder, ICON agents at eXp Realty.
What To Do First
If you’re watching this on YouTube, subscribe so you don’t miss the next live session — we build these on screen, start to finish, and take questions while we code. And if you’d rather read than watch, the free, open-source version of this is on GitHub — the link’s in the description, and it’s the same exact code we’re using here, right now, on screen.
Before we open either engine and type anything, let’s be clear on what we’re asking. We want to pull MLS data — listings, sales, prices — and use it in a real, production application, not just print it to a screen. That means the tool we pick has to connect to whatever backend we’re building on, fetch the right data every time, and do that reliably, on a schedule, without us watching.
We’re also clear on what we’re not asking. We’re not asking which AI is ‘the best’ at general conversation, or who wins a typing contest. This is an actual client question, not a debate-club topic — it has a real answer, and we’re going to build it on screen and show you exactly how it works.
With that background out of the way, here’s the decision framework, and we promise to build both options, live, on screen, before you decide which to run in your own business.
Step one, and the whole reason you open either engine, is that you already have a working version of this question, running in your own business, somewhere. Almost every agent or broker who’s been around for a few years has built some kind of MLS data pull — a saved search, a shared sheet, a dashboard — and it’s probably running on a schedule, not just sitting there waiting for you to open it. So before you ask either AI for help, find that pull yourself, the old-fashioned way, by looking at what’s already running in your business. Open your CRM, check your shared drives, and ask your assistant, if you have one, what’s on the schedule this week. The vast majority of the time, one of two things is true: either the pull you’re looking to replace is already running on a schedule and just needs the source updated, or it’s one of those old saved searches that’s been sitting there for years doing nothing. Either way, find it before you ask an AI to build it from scratch.
Where to start
If you want to run this on YOUR business — not just watch us do it on YouTube — here’s your next step. We offer a free 15-minute call. It’s just us, Al and Victoria, and we’ll look at what you’re already doing and tell you honestly what fits and what doesn’t. The link to book it is right in the description below, and our number’s there too. And if you’d rather start tinkering yourself, the free starter skills are open-source on GitHub — that link’s in the description as well. Find out what Claude and ChatGPT can actually do in YOUR business, not just in a demo. We’ll see you in the next one.
Frequently Asked Questions
What’s the difference between Claude and ChatGPT for real estate?
ChatGPT gives you back whatever you put in, formatted nicely, while Claude is set up so you can show it a market snapshot—what’s selling, what’s sitting, what people are asking about—and ask it to tell you what that actually means for your business. With your own MLS data, Claude can analyze real opportunities instead of just regurgitating generic advice.
How do I set up an AI chat window that answers buyer questions with my MLS data?
You build a chat window connected to your real MLS data that answers questions like “what’s for sale in Greensboro” with fresh, filtered data from your market. You don’t need to be a developer—you just need to know the right question to ask and the right way to ask it, and you can have it running on your website in 45 days or less.
How do I verify the AI is giving me correct and current information?
Ask both Claude and ChatGPT the same question about your MLS data and watch which one gives you the right, current information versus outdated listings. This matters most if your business runs on relationships and repeat clients, since posting outdated information erodes trust with your market.