A story out of Miami is making the rounds about a homeowner who used ChatGPT to sell his house instead of hiring a real estate agent. The story notes that the seller says the process “exceeded expectations” and went smoothly. But the agent on the opposite side of the transaction, and her broker, noticed that some key details were missing from the narrative.
The buyer’s agent, Elinor Solomonoviz, and her broker, Ines Hegedus-Garcia, noted that AI missed key market nuances that likely cost the seller money. Solomonoviz, who wasn’t representing the seller, then became the person guiding the transaction forward for both sides. Solomonoviz is an agent and Hegedus-Garcia is managing partner and branch leader with Avanti Way Realty, with multiple offices in the Miami area.
What AI Didn’t See
Solomonoviz recalls viewing the listing for the home and telling her buyer that some upgrades were likely needed but that the property checked off most of the buyer’s needs. Solomonoviz worked with the seller to set up a showing, but when she and her buyer showed up, Solomonoviz immediately noticed the photos from the listing didn’t match the property.
“We were in shock. The property looks much better,” she says.
The listing photos didn’t capture much of the property’s differentiating features: a large backyard, an oversized pool, upgrades and design elements that changed how the home felt in person. Those details don’t always translate through automated systems, and they directly impact perceived value. Solomonoviz knew that because she knows the market intimately.
“I’ve been looking at this market for the past few years,” she says. “I go to see most of the homes when they’re listed. So I know most of the floor plans, what’s good in the house, what’s not good in the house.”
Hegedus-Garcia noted that whenever a potential client wants a home in Cooper City, Solomonoviz is the area expert she refers them to.
That kind of on-the-ground knowledge isn’t something AI can replicate. It’s built over time, property by property, transaction by transaction.
Pricing Isn’t Just Math
The biggest gap, according to Hegedus-Garcia, came down to valuation. Her agents are trained to pull comps and use automated tools like RPR to come up with valuations, but then, she encourages her agents to run through those valuations by hand.
“I like to do three to four different models to triple check those numbers because at the end of the day, it's difficult to arrive at a number, and an automated value model is never going to know the nuances,” Hegedus-Garcia says. “Like Elinor told us that the bigger yard, the views, the trees or the cul-de-sac, those things are not [calculated] in the AVM.”
After her analysis, Hegedus-Garcia estimates the seller left significant money on the table, potentially far more than the commission he set out to save.
The seller notes in the news story that he received five offers on the home within a 72-hour period, and working with an agent could have helped him leverage the interest in the property to maximize profits.
The Transaction Itself
Despite choosing not to hire a listing agent, the seller still needed help navigating contracts, disclosures and next steps. When Solomonoviz and her client were ready to make an offer on the home, she sent over the standard Florida real estate contract, to which the seller responded saying he needed to go over it with his lawyer. But the lawyer didn’t always relay answers in a timely fashion, and the seller still had many questions. Since he didn’t have an agent to rely on, he directed them to Solomonoviz, she says. She stepped in, informally, and answered all of his questions.
“We ended up being on the phone [one night] until 11 p.m. to go over everything,” she says.
After she thoroughly explained everything in the contract, the seller accepted the offer from Solomonoviz’s buyer, but the questions kept coming, she says.
“Seller’s disclosures, HOA addendums, inspections—he had no clue,” she recalls. “Every single day I would take his phone calls from sometimes 7:30 in the morning to 8 or 9 p.m.,” she says. During the home inspection, she was present for the entire three hours, and the seller talked with her and asked her questions about what came next and how to move forward.
“If it was somebody else on the buyer side, someone who didn’t want to answer his questions, this deal might have not closed,” Solomonoviz says.
The seller’s decision to use AI wasn’t arbitrary. He told Solomonoviz that it stemmed from a bad experience with an agent about a decade ago. But by the end of the transaction, his perspective had shifted.
“He told me how I had changed his mind on real estate agents and that next time he needs to hire me,” she says.
Her experience and poise—specifically her willingness to answer questions, interpret contracts and guide the process in real time—made the difference.
The Bigger Picture
Hegedus-Garcia doesn’t dismiss AI. She sees it as a tool, one that consumers are already using and that she uses in her office. But tools don’t replace market knowledge or judgment, and that’s where real estate professionals hold value and shine during the transaction. Timing, negotiation and clarity matter, and in this case, Solomonoviz’s expertise highlighted AI’s limitations.
What might be most important was the level-headedness and reassurance she exuded, often in moments where the seller didn’t know what to do next, Hegedus-Garcia says.
“There’s nothing like answering a phone call and calming someone down when that person thinks something is going wrong,” Hegedus-Garcia says. “I don’t think any AI can do that.”
