300 agents in your market produce a FAQ page. 2 are actually useful. The rest are either a word cloud of questions nobody asks or a sales pitch masquerading as an answer.
We run every single one of those pages through Claude and ChatGPT and watch what they actually do with a real question. If you’re automating your FAQ page, here’s what each tool actually does with your real client questions — and why that difference matters when you’re trying to own your system instead of rent it.
If you’re a few years into this business — closing 6, 7, 8 deals a year, pushing for more lume, more consistency — you’ve probably already built or bought some kind of system to keep your documents and your data connected. A CRM, a follow-up sequence, a posting schedule, something like that.
And you’ve probably felt the frustration of wanting to ask that system a question — what’s my current pipeline look like, or which market is performing right now — and not being able to, because the question wasn’t phrased exactly right, or because you didn’t have the right report open at the time.
What It Actually Does
Here’s what both tools actually do when you ask them to draft a document like this, step by step.
First, neither tool generates anything truly new. They’re reading what’s already out there — articles, FAQs, agent guides — and sorting through it to find the questions people actually ask, the issues that come up over and over. Then they’re writing an answer that covers as many of those questions as possible, in a way that’s easy to scan.
Second, neither tool writes anything you couldn’t write yourself, with enough time and sources. The real work isn’t the typing — it’s getting the facts right, staying current, and covering everything that matters to your client right now. Both tools give you a starting point, but the final call on what’s accurate and complete is always yours.
Third, neither tool replaces a second pair of eyes on anything important. If you’re drafting something you know will actually go to a client, or that touches anything legally or financially significant, have it reviewed by a lawyer or a broker who practices in your area — the same way you would with a document you’d written yourself.
And one thing neither tool does that’s worth noting: neither is watching what you actually do with the draft afterward. It’s written, you have it, and then it’s up to you whether you use it, save it for later, or rewrite it entirely. There’s no notification, no follow-up, no rating system. The value comes from what you do with it, not from the tool checking back to see if you opened it.
How You Set It Up
Here’s the setup, straight and simple. We’re both ICON agents at eXp Realty — Al with eXp Realty in Eastern North Carolina, and Victoria with eXp Realty in Western North Carolina — and we run our businesses on Claude Code. This isn’t a sponsored video; we bought our own license and built our own automation. We’re just two agents who found a better way to get the same work done in a fraction of the time, and we’re showing you exactly how we did it, because we run this on screen, live, every single day.
If you’re new to us, the quick version is that we both run our businesses on a system called the Al and Victoria AI Real Estate Automation Stack — that’s the official name, but everyone just calls it the AV Stack for short. It’s a set of automation scripts that run on your own computer, on a schedule, completely free, no monthly charge. You just download it and run it. The AV Stack includes this automation plus several others that cover follow-up, content, and lead response. If you want to run the same automations on your own business, the free starter version is in the description, along with a link to book a free 15-minute call with us to see it running live before you decide.
Now, back to the main point of this video. We run two versions of this automation on our own machines — one on Al’s computer, and a second, slightly different version on Victoria’s — and that’s the key to understanding why we picked Claude over ChatGPT. Neither one of us is a developer. We didn’t build this from scratch; we just told Claude Code, in plain English, what we wanted it to do, and it wrote the automation for us. The reason we picked Claude over ChatGPT, for this job specifically, is that Claude gave us two things ChatGPT couldn’t touch.
The first was the ability to run the same automation on two different machines and have it stay exactly in sync, in real time, on both, without us doing anything extra to maintain that sync. Every time either one of us opens the automation and hits ‘refresh’ on our own screen, the exact same, up-to-date version runs on both of our machines, instantly, without us Emailing each other or merging anything by hand.
How It Gets You Found
Here’s the real test, beyond features: how easy is it for Google to understand and rank these pages, so a seller in your market typing ‘realtor near me’ finds you instead of the agent down the street who published last week.
Both tools generate targets that are, technically, rankable — a real, specific page on a real, specific topic, with a real, specific question asked and answered.
Where they differ is target freshness.
Every Monday, the ChatGPT setup publishes a completely new set of targets, wiping the slate clean and starting over from scratch — which means last week’s targets are already gone, replaced by this week’s, and there’s never an accumulation of older, stale targets sitting out there doing nothing.
The Claude setup, by contrast, adds targets one at a time, whenever the automation finds an opening — so there is an accumulation, with older targets still working alongside newer ones, and the whole pool stays alive instead of getting wiped every week.
In practice, that means the ChatGPT approach produces a sharper, more current target set every single run, while the Claude approach lets individual targets age naturally and build authority over time.
There’s no right answer — it depends on whether you want the fastest-moving target front, or the steady, ongoing value of a persistent pool.
Either way, both are significantly more purpose-built than the generic blog post or city guide most agents publish, and both sit in the exact same position to get found.
What It Means For You
For you specifically — the agent trying to get past 6 or 8 deals a year — this means you can stop letting a fear of technology hold you back from building real momentum. You now have two tools, side by side, that can write your listings, source your content, and answer your clients, and you get to choose which one runs which job. You are the one calling the shot, and that is the whole point.
If you want to try this yourself, start small. Pick one listing and write it with Claude Code. Then do the next one in ChatGPT and set up the comparison. You do not need to be a programmer to do this. You just need to be willing to type ‘listings + your city’ and see what comes back.
We are Al and Victoria Pinder, husband-and-wife ICON agents at eXp Realty, and we run this in our own business every single day. If you want to see it running live, the link to book a free 15-minute call is in the description.
What It Frees You Up To Do
For one thing, it frees up the real quick answer that used to be the first thing you’d lose an evening on — the client asks something simple, and you sit down and re-type the same search you’ve typed a dozen times before, hoping this time the answer’s at the top. That’s gone. It’s the same question, asked once, and the answer comes back before you’ve finished your next sentence.
More than that, it’s the same question answered the same way every time, which means you’re not quietly building up a collection of your own answers that contradict each other. That’s a real problem, especially in a team setting — everyone asks ChatGPT, everyone gets a different answer, and nobody knows which one to trust. This gives you one source, one standard, and everyone’s answer comes from the same place. That’s real compliance value, and it’s the same reason this beats a knowledge base in front of a client — nobody’s left watching you dig through your own site for the answer that used to be there.
And there’s a third thing, bigger than both of those. Neither of these is a tool you rent. You own the skill, you run it on your own machine, and that ownership is the real difference between asking a chatbot and asking something you built yourself. When you own it, you can open it up and look under the hood. You can see exactly what it’s searching and how it’s putting the answers together, and that understanding is what lets you trust it in front of a client. A chatbot answer is always just a suggestion, because you have no idea how it got there. Your own system, built from these skills, is the one you can actually put your name on.
What To Do First
If you’re building your own AI help systems — not buying a ready-made tool — the first decision is never about brands: it’s about your own sources. Before you call either Claude or ChatGPT, check what you already have. If you’ve already got a knowledge base or a collection of useful documents saved somewhere, start from that. An AI helper is just one way to connect to that information — the right one is whichever one can actually reach it. If the documents are on your own machine, or in a shared folder your team already uses, that’s almost always the easiest connection to make, and it’s usually the right one. Both Claude and ChatGPT can search a local folder or a shared drive, and they do it in pretty much the same way once you’ve set it up. So the first call is to your own team: what do we already have, and how can we connect to it directly instead of asking an AI to go find it.
If you want to run this on your own business, here’s where to start: pick one real question your clients ask, run it through both Claude and ChatGPT, and see which one gives you the answer you’d actually use. You don’t need to be technical: you just need to open Claude Code or ChatGPT, describe what you want your automation to do in plain English, and let it build it for you. You’re not committing to either one by trying it out. You’re committing to actually looking, which is the hardest part. And you’re not doing it alone — we offer a free 15-minute call. The link to book it is in the description, right next to the link to this on YouTube so you can watch it again later. We’re Al and Victoria Pinder, husband-and-wife ICON agents at eXp Realty, and we run this in our own business every single day. We’ll show you exactly what it looks like running on real data — yours, if you want.
Frequently Asked Questions
What’s the main difference between Claude and ChatGPT for real estate automation?
Claude lets you run the same automation on multiple machines and keep it synced in real time without extra work, while ChatGPT doesn’t offer that functionality. Neither tool writes anything you couldn’t write yourself, but Claude gives you better control if you’re running the system across your team.
Do I need to be a programmer to set up these automations?
No. You don’t need to be a developer—you just tell Claude Code or ChatGPT in plain English what you want it to do, and it writes the automation for you. Both tools can build what you need without any coding experience on your part.
Should I use this automation instead of having a lawyer review important documents?
No. If you’re drafting something that touches anything legally or financially significant, have it reviewed by a lawyer or a broker who practices in your area, the same way you would with a document you’d written yourself. The automation is a starting point, but the final call on accuracy and completeness is always yours.
There are two ways to get a high-quality image on a listing. One is to hire a photographer, which starts at $250 and goes up from there. The other is to sit down and build it yourself, on screen, with a computer program doing the actual work. This week, we walked away from the first call and did the second.
If you’re the relationship-first agent everyone in your community already knows — the one who shows up when a neighbor needs help, and the one people call when they decide to move — this is your play. You didn’t get into real estate to become a marketer, but you also know that being invisible online means losing deals to the agent who shows up in the search results. This is how you show up without becoming a marketer.
What It Actually Does
Here’s the part that actually matters, because neither of those features is the point.
When you run this job, you are not buying a piece of software you’ll need to install, or a monthly subscription you’ll need to renew. You are hiring a real service, the same way you’d hire a photographer or a designer — you set up what you want once, in plain English, and then you go back to your own work. The image comes back to you, finished, through the same channel it was set up on. You own the final image outright, with all rights released, and you can use it anywhere you like — your own website, your own marketing, your own brand, with no license or attribution string attached.
That ownership is built into the job on purpose, because the real value is not the image itself. It’s that you now have a repeatable, consistent way to produce your own high-quality listing images, without depending on a third-party platform or a separate vendor, and without adding another recurring charge to your business. That consistency is what gets noticed, both by the agent who’s reviewing your listing and by the AI tools people are now starting to ask when they’re looking for an agent.
If you’re building a brand, this is how you ensure that every single one of your listings looks like it belongs to the same professional operation — the same look, the same quality, automatically, every time, without you ever having to think about it. If you’re experimenting with a new market or a new niche, this is how you get the matching visuals for that market in front of you on day one, without committing to a whole new website or a new content calendar. You set up the job once, in plain English, and from then on it’s running on the backend of your business, quietly, doing a job that used to eat up a full afternoon. That’s real leverage, because it means you get the same high-quality image work done on a $300 listing as you do on a $3,000,000 one — the same automatic system, running the same way, every single time.
How You Set It Up
Here’s the actual setup. We run this on a Mac, and the free version of iTerm — just a terminal window, no interface to open. You don’t need to be a programmer for any of this. You open the terminal and type the first command, and it opens a folder on your own computer named after the listing, with today’s date built in. Every time you run it, it goes into that exact same folder — so last week’s image is still sitting there if you want to go back and look at it.
The first command is the only one that actually matters. It tells Claude Code which listing you’re working on, pulls the right property file, and generates the image in one go — you don’t need to open anything or switch windows. You just type ‘generate-listing-image’ followed by the listing’s own ID, the one that’s already in your MLS system, and it knows exactly which property you mean. Ours looked something like ‘generate-listing-image 123456’ — replace that with your own listing ID, and run it the same way you’d run any command.
The image shows up in the folder automatically, named with the address, the date, and the word ‘image’ — so you can search for it and find it, even if you have 100 of these on your hard drive. And because the folder is named with today’s date, next week’s image will go into a completely different folder on its own, so nothing gets mixed up.
The only other thing worth understanding is the size it generates. It’s set up to make a landscape image — so the width is 1920 pixels, and the height comes out automatically whatever it needs to be to match that aspect ratio. That’s the right size for just about any listing platform, and it’s what we’d recommend.
If you wanted something taller instead — a portrait, for an Instagram post for example — you could change that one setting yourself, without needing us to touch anything. You just tell it ‘generate-listing-image –aspect 9-16’ and the height becomes the dominant number instead. There are a few other aspects you can pick from if you want something specific, or you can give it your own custom aspect ratio if you’ve got a particular use for it.
How It Gets You Found
Here’s what actually matters, and what almost no one tells you: whether it’s Google, or ChatGPT, or Gemini, or Perplexity, an AI tool answers a question about a property by looking at what’s been written about that property online, and picking the best-sourced, most recent, most complete answer it can find.
That’s it. That’s the entire contest. You don’t need to build the best website in the world, or the biggest one, or the fastest-loading one, or the most secure one — you just need to be the one who put out the best, most recent, most complete information about the property, in a form that the AI can actually understand and hand to the person asking the question.
And that’s exactly what this does. It publishes the listing, with the full description, the photos, the features, the price, the location, the MLS number, the walk-score, the crime-score, the school-score, the weather, the local points of interest, the drive times to nearby cities — all of it, pulled together into one place, in a format that’s easy for an AI to pick up and hand to a person.
It publishes that to a page that’s automatically named after the address, so the address itself, typed into Google or ChatGPT, is the exact query that pulls it up. It publishes it to a real, live, indexable web page — not a PDF, not a JPEG, not a slideshow, not a locked-down portal page you can only see if you’re signed in. A page that you, the agent, can link to from your own site, your own email, your own social posts, and have that link actually work, forever, without depending on anyone else’s system staying up.
And here’s the part that actually matters more than anything else: it publishes that page the very first moment the listing goes active, the second the status changes from ‘pending’ to ‘active.’ Not tomorrow, not next week, not after you’ve remembered to tell a marketing coordinator to put it on a list. Right now. Because the source connection is live, running, watching the listing in real time, and the publish trigger is set up exactly right — and that’s the whole reason we built this, because we watched agents losing ground to agents who just published their listings faster.
If you’re the agent who shows up first when a seller asks Google, or when a buyer asks ChatGPT, you’re the one who gets the call.
What It Means For You
If you’re building a brand on visuals — a real look at a real place, not a generic stock photo — this is what a consistent aesthetic does: it’s how people start recognizing your listings in a feed full of look-alike posts. If you’re chasing more volume, it’s leverage: the same hours produce more finished, on-brand content than anyone can create by hand. And if you’re protecting a reputation, it’s compliance built into the workflow — every image defensible in front of a client or a council.
What It Frees You Up To Do
It’s compounding.
Every week you generate one more image you never have to remake, one more look you never have to describe from scratch.
Four weeks in, that’s 4 different styles you can pull from, without ever opening a design app.
Twelve weeks in, it’s 12, and those 12 are quietly building a consistent look for your content, a look your audience starts to recognize before they even read a headline.
That consistency is what makes you easy to follow, and being easy to follow is what turns a channel into a relationship.
So this isn’t about getting faster at one task.
It’s about building a body of consistent, recognizable work, one quick session at a time, without ever opening a computer.
What To Do First
If you’re building your own AI-powered real estate business — listing copy, images, videos, and more — generated automatically, in a few minutes, from a single question — this is the one you ask first.
Not ‘generate an image for my listing’, ‘write my listing copy’, or ‘make a video about my listing’. None of those.
A file named ‘listing-image.jpg’ appears on your computer in just a few minutes. It’s the right size, the right aspect ratio, for your blog’s current layout. And the background is removed — a house on a clear, simple white background — so it looks good on your blog even if the image was shot on a cloudy day with grass and sky in it when you took it.
You were never told a piece of software’s name, you never opened a program, you never watched a tutorial, and you never learned anything technical. You just asked two questions, in plain English, and a high-quality image you can use this week is on your computer.
If you want to see exactly how we set this up, the free starter skills are open source — the repository link’s in the description, along with a free 15-minute call if you’d like to talk through your own use case. We’re Al and Victoria Pinder, ICON agents at eXp Realty.
Where to start
If you want help setting up something like this for your own business — not just watching us do it on YouTube — we’d love to talk. The free call link is right in the description below, and we’ll look at what you’re already doing and tell you honestly what fits and what doesn’t. No cost, no obligation, no pitch. We’re Al and Victoria Pinder, husband-wife ICON agents at eXp Realty, and we run every piece of this stack in our own business every single day.
How much does it cost to hire a photographer for listing images?
Hiring a photographer starts at $250 and goes up from there, depending on the photographer and the scope of work.
What image size does this system generate?
The system generates landscape images that are 1920 pixels wide, with the height automatically adjusted to match that aspect ratio. You can change this to a portrait orientation (like 9-16 for Instagram) or create a custom aspect ratio if needed.
When does the listing get published online after it goes active?
The listing is published to a live, indexable web page the moment it goes active in your MLS system, not days or weeks later. This happens automatically because the source connection is live and watching the listing in real time.
You’re 2 to 3 deals into the year, closing strong, and the one thing you still can’t get exactly right is which assistant question actually triggers the right saved run — you either ask it the wrong way and get nothing, or you remember the exact phrase by accident and it works, and you never know why either way.
If you’re a few years into this business, closing 6, 7, 8 deals a year, pushing for more volume without adding hours to your day — this is your play.
You didn’t get into real estate to become a part-time IT specialist. You didn’t sign up to spend Sunday nights building out a follow-up sequence that doesn’t exist yet.
But you also know the agents who scale are the ones who stop doing everything by hand.
So you’re not opposed to technology. You’re just short on time to experiment with it.
That’s exactly the gap these 2 systems are built to fill. You describe what you want in plain English, and both will build it for you — Claude in a few minutes, ChatGPT in a few hours, sometimes a full day.
If you’re chasing volume and you’re technically inclined, the Claude version is the one you’ll want to sit down and tinker with yourself. It’s open source, it runs on your own machine, and you can watch exactly how every piece fits together.
If you’d rather just describe what you want in an email and come back to a working version already sitting in your inbox, the ChatGPT version is the one we’d set up with you. Same underlying engine, same capabilities — it just runs on our end instead of yours.
What It Actually Does
Here’s what both actually do, step by step, when you give them the same job.
The first few moves are the same for both. They read your instruction, they open the right connected account, and they load your current list into a working table — a spreadsheet, essentially, that they can read and write from. That’s the setup, and it’s the same whether you’re talking to Claude or ChatGPT.
Where they start to diverge is the first real question they answer against that table: what is actually active versus what is paused or pending. Both will look through your list and tell you, correctly, which properties are currently live and which are not — but Claude does that in one pass, reading the table exactly once, and writes the answer directly to a file so it’s saved for next time. ChatGPT, on the other hand, checks each property one by one, which means it’s re-reading your whole list for every single question, and the answer disappears into a chat window with no record of what it found last Tuesday.
That’s the first real difference, and it’s about whether the tool re-does the same work every time you ask or remembers what it found last time and writes it down so the next question starts from the same current picture. Claude builds a real working file.
How You Set It Up
Here’s the setup, plain and simple. We’re comparing two versions of the same workflow — the same daily check-in, the same weekly review, the same monthly deep dive — run on two different AI engines. One version runs on ChatGPT, in plain English, by typing a few sentences and pressing enter. No coding, no API key, no monthly seat cost. The other version runs on Claude Code, which is Claude with extra horsepower for running automations, and it’s set up as a scheduled job that runs on its own, without us at the keyboard. It checks in with the right team, pulls the right data, and posts the right digest — automatically, on a clock, while we’re out at a showing. Both versions start from the same base — the same instructions, the same data sources, the same schedule. The only thing that changes is the engine running it.
The instructions live in our public starter repository on GitHub — the one you can clone and run on your own machine right now, free, no sign-up. If you haven’t already, do that. Clone it, open the readme, and follow the instructions. Get it running on your own computer before you ever talk to us. That’s the whole point — you own it, you touch it, you understand it, before you ever let anyone add a layer on top.
Now, what we’re looking at today is the difference between telling an AI ‘run this check in the background every morning’ and actually putting that job on a schedule so it runs without you remembering to open a window and type anything. Both versions start from the same instructions. You open the readme and section 6 says, in plain English, ‘every morning, check in with the team and post a digest of what landed overnight.’ That’s it. That’s the instruction. On ChatGPT, you just type ‘run the morning check’ and it does that one check that one time, live, while you’re watching. On Claude Code, you go into the scheduler and set up a recurring job — day, time, what you want it to run — and from then on it runs on its own, unattended, while you’re doing something else. Same instruction, two different ways to fulfill it. One is a live question you type when you remember, and the other is a scheduled job that runs whether you remember or not.
How It Gets You Found
Here’s the actual question Google gets asked, over and over, by agents trying to get found: ‘How does Google decide which website to show when somebody searches for a real estate agent near me?’
Both of these tools can answer that, in a way that a client could actually understand.
A ChatGPT chat window can pull up a simple, plain-English explanation of how Google’s search algorithm works, what it’s looking for when it ranks a local agent website, and how you can actually influence where you show up.
A Claude Code chat window can go a step further and actually show you, line by line, what a real, working local SEO system looks like — the way the website is structured, the way the content is targeted, the way the citations are managed, and how all of it ties back into the way you run your business every single day.
That’s not a feature either tool is specifically promoting. It’s just what they’re capable of, when you ask the right question.
If you’re building a real estate brand from scratch, this is how you stop chasing the next marketing tactic and start actually understanding the one thing that actually matters: how you get found.
What It Means For You
If you’re pushing for volume, this points you to ChatGPT for the pipeline and Claude for the complexity. If your business runs on relationships and referrals, the Claude version is the one that protects that asset — the same questions, asked differently, yield a check instead of a repeat client. Either way, it’s the same two engines, doing the same job, and you pick the one that runs on your data instead of a generic answer pulled from the internet.
We’re Al and Victoria Pinder, husband-wife ICON agents at eXp Realty. If you want to run this on your own business instead of watching us do it on YouTube, the free starter skills are in the description — everything in this video is yours to build, whether we know you or not. If you’d rather talk it through, we do a free 15-minute call. That link’s in the description too. Subscribe so you don’t miss the next live session — we build one of these on screen, start to finish, every month. We’re Al and Victoria Pinder, and we run our business on the system we just showed you.
What It Frees You Up To Do
The real recording: a side-by-side comparison, on screen, of both systems running the same workflow — the same questions, the same data, the same instructions — while we watch, live.
What To Do First
If you’re watching this, you’ve probably already built or bought a few automations — a follow-up sequence, a report, a posting schedule — and you’ve felt the frustration of wanting to change one instruction and having to open up 3 or 4 different tools to do it.
This is about getting rid of that frustration. Before we get into which engine actually runs the automation — Claude or ChatGPT — there’s a decision you need to make yourself, and it has nothing to do with AI.
The first thing to ask is: what actually needs to run automatically, and what is better answered by a person?
Almost every agent says the same thing when we ask them straight out: ‘My brand is about relationships. I’m not handing that over to a robot.’
Nobody is suggesting you do that. The brand voice, the judgment call on a specific post, the real conversation — that’s you, in the room, every time.
What you can hand off are the repeatable jobs that eat up your time and don’t depend on a client knowing your name: the daily check, the weekly email, the monthly report, the occasional reminder.
So the first move is not downloading anything. It’s opening a notebook — an actual notebook, or a document named ‘Automation Candidates’ — and listing out, in plain language, everything you or your team currently does by hand that fits that description.
You don’t need to build a map of your business before you start. Write down the first 5 or 6 items you can think of, and then start checking them one by one.
Where to start
If you want help setting up something like this in your own business — your leads, your content, your brand — we do 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. Clone it, try it, own it.
We’re Al and Victoria Pinder, husband-and-wife ICON agents at eXp Realty here in Eastern North Carolina, and we run every one of these workflows every single day, in our own business, for our own clients. We’ll tell you what works and what doesn’t, honestly, because we’re building this on the job, the same way you are.
Frequently Asked Questions
What’s the main difference between Claude and ChatGPT for real estate automation?
Claude reads your list once, saves the answer to a file, and remembers what it found for next time. ChatGPT checks each property one by one and re-reads your entire list every time you ask a question, with answers disappearing into a chat window instead of being saved.
Do I need coding skills to set this up?
No. ChatGPT requires no coding, API key, or monthly seat cost — you just type a few sentences in plain English and press enter. Claude Code runs on your own machine as a scheduled job, but the instructions live in a free, open-source GitHub repository that you can clone and run yourself.
Should I automate everything in my business?
No. The brand voice, judgment calls on specific posts, and real conversations with clients should always be you. What you can hand off are repeatable jobs that don’t depend on a client knowing your name — like daily checks, weekly emails, monthly reports, and occasional reminders.
