One of the biggest mistakes I see agents make with AI is assuming the prompt is the most important part of the process.
It isn’t. The quality of your listing description is almost always determined by the quality of the information you provide. Most agents are feeding AI a property profile, a few photos, and hoping for the best.
I build a complete property file before I ever ask AI to write a single word. The goal is simple: teach AI the property before asking it to describe it.
Here are the seven sources of information I use for nearly every listing.
1. Seller Walkthrough
Before the seller leaves the property, I ask them to walk me through the home.
This is where I gather information that doesn’t exist in public records, the MLS, or property photos.
- Recent upgrades
- Replacement dates
- Custom features
- Improvements they’ve made over the years
- Details only the owner would know
Rather than trying to remember everything later, I use Breezy‘s note taker to create a transcript of the entire conversation.
This becomes the first piece of the property’s knowledge base.
2. Agent Walkthrough
Once the seller leaves and the photographer begins working, I complete a second walkthrough by myself.
This time, I’m documenting everything I can physically observe.
- Countertop materials
- Cabinet construction
- Pantry storage
- Smart home features
- Lighting upgrades
- Built-in storage solutions
- EV charging capabilities
- Room relationships and flow
I again use Breezy’s note taker so I can create a complete transcript of my observations without slowing down the photography process.
3. Tax Assessor and Public Records
Next, I gather information from tax records and public property data.
This helps verify:
- Square footage
- Lot size
- Bedroom and bathroom count
- Year built
- APN information
- Legal additions
- Other important details
Even when I already know the information, I want AI working from verified data.
4. Previous Listing Data
If the property has sold before, I review previous listing descriptions and historical MLS information. I’m not looking to copy old marketing. I’m looking for facts.
Often, previous agents uncover details that may not be immediately obvious years later.
Reviewing historical information helps ensure nothing important gets overlooked.
5. Property Photos
Once photography is complete, I upload the images into ChatGPT. You could do this with photos you take on your iPhone, but I generally find myself writing the description around the time I get pictures back.
The photos help provide context around design, finishes, condition, style, lighting, views, and overall presentation. Rather than relying solely on written descriptions, AI can now see the property itself.
6. 360 Tour Screenshots
This is one of the newest additions to my workflow.
Traditional listing photos show individual rooms. A 360 tour helps explain how the home actually lives. I capture screenshots throughout the tour and upload them into ChatGPT so it can better understand room relationships, layout, and the overall flow of the property.
This often leads to more accurate descriptions because the AI can understand how spaces connect to one another rather than viewing each room in isolation.
7. Neighborhood and Lifestyle Information
The property itself is only part of the story. Buyers are also purchasing a lifestyle.
For every listing, I include nearby amenities, recreation, shopping, transportation access, beaches, parks, golf courses, dining, and other location-specific benefits that make the property attractive.
A great listing description should help buyers understand not only the home, but what life might look like after moving in.
Bringing It All Together
Once all seven sources have been collected, I place them into ChatGPT. At that point, I’m not asking AI to write a listing description from scratch.
I’m giving it a complete understanding of the property and asking it to organize that information into a compelling narrative. The result is a listing description that is more detailed, more accurate, and more likely to highlight the features buyers are actually searching for online.
AI didn’t make my listing descriptions better. Feeding AI richer, more detailed information made my listing descriptions better.
The technology simply helps me organize and communicate information that I already collected through a more intentional process.
