Let me set the scene. A buyer comes to me, preapproval in hand, Pinterest board fully loaded, absolutely convinced they've done their homework. They've been chatting with AI for two weeks. They know the neighborhood, they know the market, they even know what their dream home is worth.
Except they don't. Because they asked an AI — and the AI made most of it up with complete confidence and zero apology.
This is the part nobody tells you about the AI real estate revolution: it's genuinely impressive until it isn't, and the gap between those two states is invisible until you're standing in the wrong house with the wrong number in your head.
The Good Stuff — And It Is Actually Good
I'm not here to trash AI. I use it. A lot. It's changed how I work, how I write, how I analyze market data. For buyers starting a home search, it's a legitimately useful first tool.
AI can comb through hundreds of listings while you're still deciding what to eat for breakfast. It can filter for your specific criteria — three bedrooms, single story, home office, south-facing backyard, somewhere in Rio Rancho that isn't next to a gas station — and surface properties you'd have missed scrolling on your own. Virtual tours have gotten genuinely immersive. Some listing videos now look like they were produced by someone who really wanted to be in the film industry but settled for real estate marketing. Honestly? Not complaining.
AI is a research accelerator, not a decision-maker. It can narrow your options from 400 to 12 faster than any human. What it cannot do is tell you whether that house actually smells like cat, whether the HOA president is unhinged, or whether the "mountain views" are real or imaginary.
Used in that lane — as a search tool, a filter, a way to organize your thinking — AI is genuinely useful. The problems start when people ask it to go further than it's equipped to go.
What My Home Is Worth, According to Absolutely Nobody
New Mexico is a non-disclosure state. That means sale prices are not public record. Zillow doesn't know what your neighbor's house sold for. Redfin doesn't know. The county assessor's website doesn't know. Nobody knows except the buyers, the sellers, and the agents who handled the deal.
AI knows even less than that.
When you ask an AI chatbot "what's my house worth in Albuquerque," it's doing one of two things: either pulling from publicly available listing prices (which are asking prices, not sale prices, and often wrong) or it's just generating a number that sounds plausible based on general patterns in its training data. It has no access to what homes actually closed for on your street last month. It cannot.
AI-generated home valuations in New Mexico are unreliable by design — not because AI is broken, but because the underlying sale data isn't public. If you're making a pricing decision based on a number an AI gave you, you're pricing your home based on a guess dressed up in a confident font. Call an agent who pulls real comps from MLS.
I've had sellers come to me with a number they got from an online estimator or an AI tool. Sometimes it's close. Sometimes it's $80,000 off. The variance isn't because the sellers are bad at research — it's because the tools are working with incomplete information and nobody told them that.
I Asked AI to Find Me a Great REALTOR® in Rio Rancho
This is where it gets fun.
A few months ago, just to see what would happen, I asked a popular AI assistant to recommend a good buyer's agent in Rio Rancho. It gave me three names with brief, confident bios. One of them retired years ago. One had the right name but the wrong market — she works in Taos. The third one I couldn't verify at all. No license, no brokerage, no trace on the NMREC database. The AI had, with complete self-assurance, partially invented a person.
To its credit, it didn't recommend anyone terrible. It just recommended people who were unavailable, misidentified, or possibly fictional.
This is what's called a "hallucination" in AI terms. The model generates something that sounds correct and coherent but is factually wrong. It's not lying — it doesn't know the difference. It's pattern-matching toward a plausible answer, and "plausible" and "accurate" are not the same thing when you're trying to find someone to negotiate the biggest purchase of your life.
If you're going to search for a REALTOR® using AI, treat the output like a starting point, not a referral. Look up every name on the New Mexico Real Estate Commission license lookup. Verify they're active. Verify they work in your market. Call them and see if they pick up.
The Bathroom Mirror Incident
I'd be leaving out the best part if I didn't mention the AI-generated listing photos.
There's a now-legendary listing — circulated widely in real estate circles as a cautionary tale — where AI-enhanced photos showed a creature lurking in the bathroom mirror. Not a person. A thing. The kind of thing that makes you question your decision to browse Zillow after dark.
Other documented highlights from the AI listing photo genre: a perfectly normal living room with an extra window that didn't exist, overlooking what appeared to be a post-apocalyptic landscape. A kitchen where the cabinet handles were placed somewhere between art installation and fever dream. A backyard with a pool that the actual property did not have.
If listing photos look unusually perfect — and especially if architectural details seem slightly off or don't match each other — look closely. AI photo enhancement is increasingly common in listings. A good agent will always schedule an in-person showing before you fall in love with something that may have been digitally improved into a different house.
I shoot my own listing photography. Real photos. What you see is what's actually there, for better or worse. No digital pools, no phantom windows, no bathroom entities. You're welcome.
The Chatbot That Explained Waterfowl Migration Instead of School Districts
AI chatbots embedded in real estate sites are getting better, but they're still operating on vibes in a lot of situations. Ask a generic chatbot about a school district and you might get accurate information, outdated information, or — genuinely — information about something adjacent that the model decided was related.
I've seen clients get answers that were technically correct for a different city. I've seen questions about HOA fees answered with information about a different subdivision entirely. The chatbot wasn't wrong about something — it just wasn't right about the specific thing you asked.
This matters in Albuquerque and Rio Rancho because our market has enough micro-neighborhoods, unique covenants, and hyper-local quirks that generic information is often worse than no information. "Generally speaking in the Southwest" is not the same as "specifically in Corrales where the irrigation ditches affect your fence placement."
So What Is AI Actually Good For?
Here's the honest version, having watched clients use these tools for a couple of years now.
The takeaway isn't "avoid AI." It's "use it for what it's actually good at, and don't let it make the decisions that require judgment, local knowledge, and someone who has personally driven every street in the Rio Rancho High Desert subdivision at least twice."
That last one's me, by the way. I have opinions about every street in the Rio Rancho High Desert subdivision.
Want a Real Answer About Your Home's Value?
No AI estimates. No guesswork. Just actual MLS comps pulled for your address, from someone who knows this market. Takes about 20 minutes and it's free.