AI Replaced Your Real Estate Agent? NOT SO FAST

AI has gotten remarkably good at parts of the job — and remarkably risky in others. A New York Times journalist sold his Hudson Valley home using a chatbot and pocketed real savings. Meanwhile, deepfake scams targeting real estate transactions are up 40% year over year. Christina Chorna, CT Realtor, read the headlines, dug into the data, and is giving Connecticut buyers and sellers the answer most agents won't: some of this is genuinely impressive, some of it is genuinely dangerous, and knowing the difference matters more than ever.

The Headline That Made Every Realtor's Group Chat Explode 📱

In May 2026, New York Times technology journalist Stuart Thompson published an article titled "I Tried to Sell My House With a Chatbot" — and it did exactly what a good headline is supposed to do. It traveled. Thompson used Google's Gemini to handle nearly every function of a real estate agent for his Hudson Valley home: writing the listing, finding local photographers, coordinating staging, scheduling showings, and drafting negotiation language. The home sold above asking price. Thompson estimated he saved somewhere in the range of $36,000 to $90,000 in commission, depending on how the math was framed.

He was not alone. Florida homeowner Robert Levine used ChatGPT to sell his Cooper City home — pricing, marketing, and negotiating almost entirely through prompts — and closed in five days with five offers in the first 72 hours, at a price $100,000 above what human agents had estimated. He pocketed roughly $28,000 in commission savings and told reporters: "I'd recommend it to everyone."

Christina Chorna read both articles the week they came out. Her reaction was not defensiveness. It was genuine professional curiosity — followed by a much more nuanced conclusion than most agents in her position would be comfortable sharing.

🤖 What AI Is Actually Good At in Real Estate

Let's give credit where it's due, because pretending AI can't do meaningful real estate work is not honest, and Christina Chorna does not do the dishonest version of this conversation.

• Writing listing descriptions: AI is genuinely excellent at producing clean, well-organized property copy quickly. Thompson's buyer's agent reportedly told him she assumed a professional realtor had written his listing — she was fooled by the AI's language and organization.
• Pricing research and market data synthesis: Both Thompson and Levine reported that AI gave them more pricing confidence than the human agents they initially consulted, because chatbots pulled comparable sales data and market trend information quickly and presented it clearly.
• Scheduling and logistics coordination: Finding photographers, drafting emails, organizing showing calendars — the administrative backbone of a sale is exactly the kind of structured task AI handles well.
• First-draft negotiation language: AI can suggest counter-offer phrasing, explain typical negotiation tactics, and help a seller think through leverage points before a conversation.
• Removing a structural conflict of interest: One widely shared observation about the NYT piece deserves acknowledgment: a traditional agent's commission is technically tied to the sale happening — creating pressure toward speed over maximum price. An AI tool has no such incentive. That's a legitimate point, and Christina Chorna won't pretend otherwise.

Christina's Honest Take: None of this threatens Christina Chorna's job the way the headlines suggest — but it should absolutely change what buyers and sellers expect from their agent. If AI can write a listing and pull comps, the agent's real value has to be everything AI genuinely cannot do. More on that below.

🏠 What AI Cannot Do — And Why It Matters More in Connecticut

Here is what got quietly buried in most of the viral coverage: AI did not actually sell either home alone. A human still had to physically show the property, move furniture, meet strangers in an empty house, negotiate face-to-face with a buyer's agent, and navigate state-specific legal requirements that a general-purpose chatbot simply does not reliably know.

In fact, Levine's own buyer's agent caught ChatGPT giving him legally incorrect guidance during negotiations specific to Florida law. As Miami broker Josh Hegedus-Garcia observed: "He would run it by ChatGPT, come back to her and say, 'How about this?' And she's like, no, that doesn't apply in Florida."

That distinction — general AI knowledge versus state-specific, hyper-local expertise — is where Christina Chorna's actual value lives in the Connecticut real estate market:

• Connecticut-specific contract law and disclosure requirements: A chatbot trained on general internet data does not automatically know Connecticut's specific seller disclosure obligations, attorney-involvement customs, or municipal-level requirements that vary between Milford, Norwalk, and New Haven.
• Physical presence and in-person judgment: Someone still has to walk the property, notice the water stain in the basement, sense when a buyer is genuinely serious versus casually browsing, and stand in the room during a home inspection asking the right follow-up questions.
• Hyper-local pricing nuance: AI can pull comparable sales. It cannot tell you that a specific Milford street commands a premium because of school bus routing, or that a particular Norwalk block has flooded twice in the past decade in ways that don't show up cleanly in public records.
• Emotional navigation: Buying or selling a home is, for most people, the single largest financial and emotional decision of their lives. A chatbot does not sit across the table during a difficult conversation about a lowball offer, a family conflict over an estate sale, or the anxiety of a first-time purchase.
• Negotiation in real time, face-to-face: Reading a room, knowing when to hold firm and when to concede, and adjusting strategy mid-negotiation based on another human's tone and body language remains squarely outside what current AI tools do.

Christina Chorna's honest read: the sellers in these viral stories succeeded largely because they were technology journalists and tech-savvy early adopters, willing to do significant manual legwork themselves, in relatively straightforward transactions, in states without Connecticut's specific complexities. That is a real outcome for a specific type of seller — not a universal replacement for professional representation.

💼 What This Means for the Real Estate Career Path — Not Just the Transaction

There's a broader story unfolding here that goes beyond any single home sale. According to the World Economic Forum, entry-level job postings across industries have fallen 35% in the United States over the past 18 months, largely due to AI absorbing foundational task execution. This isn't unique to real estate — but real estate is absolutely part of the pattern.

What's interesting is what the research says should happen next, rather than what actually happens by default. Per Cognizant Chief People Officer Kathryn Diaz: "The work assumed to be done by AI in early-career roles is simply being pushed upward — leaving middle management and senior talent overextended, burned out and increasingly disengaged as they absorb junior tasks."

The forward-looking framing: entry-level roles aren't disappearing so much as being redefined around judgment rather than task execution — reviewing AI outputs, flagging complex cases for escalation, and connecting AI tools to real business context. That is exactly the shift happening in real estate. The tasks are automatable. The judgment isn't. Yet.

Christina's Career Reflection: This research changed how Christina Chorna thinks about her own role. The parts of real estate that were always administrative — writing copy, pulling comps, scheduling — are genuinely being automated, and pretending otherwise would be dishonest. What remains irreplaceable is judgment: knowing which comp actually matters, reading a negotiation in real time, and standing in a Connecticut basement noticing what a photo never would have shown.

🚨 The Part of the AI Story Nobody's Talking About Enough: Real Estate Fraud

Here is the AI real estate story that deserves far more attention than viral chatbot success stories: deepfake and AI-enabled real estate fraud is exploding, and Connecticut buyers and sellers need to understand it before it costs them their down payment.

According to the National Association of REALTORS®, deepfake scams increased 40% year-over-year according to security firm Entrust's 2026 Identity Fraud Report. AI-enabled fraud attempts surged 456% between May 2024 and April 2025 according to Chainabuse data. Americans lost nearly $893 million to AI-powered crime in 2025 alone, per the FBI's Internet Crime Complaint Center — its first year specifically tracking AI-related scams.

Real estate is a prime target for a simple reason NAR's own experts spell out: large sums of money change hands, multiple parties communicate remotely, transactions operate under tight deadlines, and buyers and sellers often never meet each other in person. As Sharon Love-Bates, NAR's director of emerging technology, put it during a recent industry webinar: "Trust has always been a foundation of real estate. Now, AI can manipulate what we see, hear and read."

How the Scams Actually Work

• Deepfake video meetings: Fraudsters impersonate sellers or buyers during virtual meetings or closings using AI-generated video and voice — sometimes convincingly enough to fool people who know the real person well.
• Wire fraud with AI-cloned voices: Criminals impersonate title companies, lenders, or real estate agents and pressure buyers into redirecting closing funds to fraudulent accounts. In one widely reported case, a Hong Kong-based financier wired $25 million after being fooled by a deepfake video call impersonating his own colleagues.
• Fabricated listings and manipulated photos: AI can generate convincing images or virtual tours of properties that don't exist, or manipulate real listing photos to hide defects or exaggerate features.
• Seller impersonation on vacant or remote properties: A Maryland vacant land transaction narrowly avoided roughly $100,000 in fraud when detection software flagged an AI-generated impersonation of the actual property owner during a live video notarization.

How Christina Chorna Protects Her Clients

• Verified wire instructions, always by phone: Never trust wire instructions received solely by email or video call — Christina confirms transfer details by phone using a previously verified number, every single time, no exceptions.
• In-person meetings whenever possible: Per NAR's guidance, transactions involving a buyer or seller who communicates only electronically deserve extra scrutiny. Christina Chorna prioritizes meeting clients face-to-face specifically because it closes the door on remote impersonation risk.
• Independent title and ownership verification: Working with trusted title companies to independently verify property ownership and documents — never relying solely on digital copies that could be altered — protects every Connecticut transaction Christina handles.
• Watching for red flags in real time: Mismatched lip movement, unnatural blinking, robotic vocal tone, sudden urgency or pressure to "act now" — these are the same warning signs NAR, the FBI, and cybersecurity firms all point to. Christina Chorna trains her eye on these details specifically because AI has made them necessary.

Real Talk for CT Buyers & Sellers: If anyone involved in your transaction — a 'lender,' a 'title company,' even someone claiming to be Christina Chorna herself — asks you to urgently change wire instructions over email, video, or phone without independent verification, stop and call a known, saved number immediately. This single habit prevents the overwhelming majority of real estate deepfake fraud.

The Bottom Line: AI Is a Tool. It Is Not (Yet) a Trusted Advisor. 🎯

Here is Christina Chorna's honest, unhedged answer to the question in the headline: AI cannot replace a real estate agent in Connecticut today — but it is absolutely changing what a good agent needs to bring to the table

The administrative tasks — writing copy, pulling comps, scheduling logistics — are genuinely automatable, and pretending they aren't helps no one. What AI cannot yet do is stand in a Milford basement and notice the subtle water stain a listing photo would never show. It cannot read a Norwalk negotiation in real time and sense when a buyer is bluffing. It cannot navigate Connecticut's specific disclosure requirements with the confidence of someone who has closed hundreds of transactions in New Haven and Fairfield County. And it absolutely cannot protect a client from a sophisticated deepfake wire fraud attempt the way a suspicious, experienced human professional can.

The sellers who succeeded with AI chatbots were tech journalists and tech-savvy early adopters doing significant manual work themselves in relatively straightforward transactions. That's a real, legitimate path for the right seller. It is not, however, the same thing as professional representation — and in a market where Connecticut homes are selling above asking price in competitive multi-offer situations, the difference between AI-assisted guesswork and genuine local expertise can be worth far more than any commission saved.

Christina Chorna uses AI tools herself, every single week — for research, for drafting, for market analysis. What she doesn't outsource is judgment, protection, and the kind of local expertise that only comes from actually knowing these Connecticut streets, these Connecticut buyers, and these Connecticut risks.

📞 Thinking About Buying or Selling in Connecticut?

Whether AI is part of the research process or not, having real, local Connecticut expertise on your side protects you from the risks a chatbot cannot see — and captures value a chatbot cannot negotiate. Christina Chorna combines the efficiency of modern tools with the judgment, protection, and local knowledge that only come from experience.

Explore homes currently available across Milford, Norwalk and New Haven, or find out what your Connecticut home is worth right now with a real, human, locally-informed valuation.

And if you're curious about what's happening around Connecticut this month while you think it over — that's a great use of a chatbot. The house part? That's what Christina is for.

🌐 Connect at: www.ctrealtorchristina.com