In Houston’s Memorial neighborhood, a “For Sale” sign stands in front of a brick two-story home just steps from a busy road. Before her agent finishes the walkthrough, a prospective buyer is already on her phone, running the address through ChatGPT to see if the asking price checks out.
A growing number of buyers and sellers are cross-checking agent recommendations against AI tools, creating friction that some Houston realtors say is complicating pricing conversations and slowing deal momentum.
AI as Second Opinion
Real estate agents have always had to manage client skepticism. But according to Mary Giovinale, Team Leader and Realtor at The Giovinale Team at Compass RE Texas, LLC, something has changed in the past year. Clients are no longer just asking questions; they are arriving at consultations with AI-generated answers already in hand. In some cases, they appear more inclined to trust those outputs than the agent sitting across from them.
“Many clients right now, they just want to believe AI more than me,” Giovinale says. She uses AI herself for planning, marketing, and analytics. But she draws a clear line between AI as a productivity tool for professionals and AI as a substitute for local market expertise in the hands of consumers who may not know how to evaluate what it produces.
The specific behavior she describes is clients going to ChatGPT and asking for property valuations. The problem, in her view, is not that AI lacks data. It is that the models cannot account for the invisible factors, specific to one street or neighborhood, that determine what a specific property will actually trade for in a specific market at a specific moment.
What AI Misses
The gap between computer-generated estimates and real-world pricing shows up most clearly in homes that look strong on paper but carry drawbacks only visible at street level. Giovinale points to a current listing in Houston’s Memorial area as an example. The property is in a desirable neighborhood with strong school district ratings, factors that would push any model toward a higher valuation. But the home backs up to a major street, a drawback that is causing buyers to hesitate despite their interest in the location.
“People love the house, people love the school, but unfortunately the house has a main street on the back, so that’s a draw factor for many people,” she says. An AI chatbot queried for a valuation would likely weigh the neighborhood positively and miss the street-facing issue entirely, or assign a generic discount that does not reflect how actual buyers in that neighborhood are responding.
This is the category of judgment Giovinale argues AI cannot replicate: the accumulated, street-level knowledge of how buyers in a given neighborhood react to specific property characteristics, what comparable sales actually looked like in person versus on paper, and how current inventory levels are affecting how much room buyers and sellers have to negotiate in real time. “The human eye is there to guide you,” she says.
A Credibility Problem
The AI valuation issue connects to a larger pattern of client skepticism that has intensified alongside broader economic uncertainty. Buyers and sellers who are already anxious about making large financial decisions in an unstable environment are more likely to seek out additional information sources, and more likely to treat those sources as checks on their agent rather than supplements to the agent’s advice.
This creates a dynamic that goes beyond any single technology. When clients arrive with AI outputs, Zillow estimates, and social media takes all pointing in different directions, the agent’s job becomes as much about figuring out which sources to trust as it is about closing the sale. Giovinale says the response has to be education, helping clients understand not just what the right answer is, but why certain information sources are more reliable than others for specific questions.
“You have to take it with a grain of salt. You have to understand and analyze that information, so we have to be very careful about how we educate our clients,” she says. That education burden now falls disproportionately on agents who want to maintain credibility in a market where clients have more information, and more misinformation, than ever before.
The risk for the broader industry is that agents who cannot make a compelling case for their expertise over AI outputs may find their role as a trusted guide gradually diminished. If clients believe a free chatbot can value a home as accurately as a licensed professional with years of local experience, the perceived value of that professional relationship weakens, even if the belief is not well-founded.
The AI Conversation
Rather than dismissing the tools her clients are using, Giovinale says she leans into transparency. She acknowledges AI’s utility openly, which she believes makes clients more receptive when she explains its limitations in specific contexts.
Her approach involves walking clients through the factors that a model cannot see, the street-backing issue, the exact lines that decide which school a home is zoned for, the way a particular floor plan is landing with buyers right now, and showing them how those factors translate into actual offer behavior. The goal is not to discredit AI but to demonstrate where professional judgment adds value that the technology cannot replicate.
“I love AI. I definitely think that it’s a great tool for us to improve and be more efficient,” she says. “But you have to take it with a grain of salt.” That framing, embracing the tool while contextualizing its limits, may represent the most viable posture for agents navigating a client base that is increasingly good at finding information but not always able to judge whether it’s reliable.
As AI tools become more sophisticated and more widely used, the gap between what they can approximate and what experienced agents actually know may narrow in some areas. But the neighborhood-specific, relationship-driven dimensions of real estate transactions, how a specific buyer pool reacts to a specific property flaw, how negotiating leverage shifts week to week based on local inventory, remain difficult to model. For now, the agents most likely to maintain their role as a trusted guide are those who can precisely articulate where their knowledge begins, and the algorithm’s usefulness ends.
About the Expert: Mary Giovinale is team leader at The Giovinale Team at Compass RE Texas, serving the Houston market with a team of four and a specialization in domestic and international relocation clients.
This article is intended for informational purposes only and does not constitute legal, financial, or investment advice. The views and opinions expressed herein reflect those of the individuals quoted and do not represent an endorsement of any company, product, or service mentioned. Readers should conduct their own due diligence and consult qualified professionals before making any investment decisions.
