AI Is a Research Tool in Real Estate, Not a Valuation Authority

Written by:

Heather Hook

Date:

03 Aug 2026

KeyCrew Studio

Share

Share via Facebook

Share via Linkedin

Share via X

Copy to clipboard

Automated valuation tools have improved dramatically over the past decade, but they still cannot distinguish between a home with modern finishes and one that hasn’t been updated in 35 years. That gap between what algorithms can see and what they miss defines the current state of AI in residential real estate – useful for preliminary research, unreliable as a final answer. Scott Spelker, a real estate professional with The Spelker Team in Morris County, New Jersey, uses AI daily in his practice but treats every output as a starting point that requires local knowledge to interpret.

Spelker’s Early Adopter Framework

Spelker compares current AI adoption to the mid-1990s internet – a period when the technology’s potential was evident but its applications were still being discovered through trial and error. Five years after dial-up connections became common, Amazon and eBay had changed commerce. Twenty-six years later, the internet’s uses are too numerous to catalog.

“It’s probably bigger than even the internet coming on,” Spelker says. He describes AI as being “in the first inning,” with the most significant real estate applications likely still undiscovered. For agents, that means dismissing AI as overhyped carries the same risk as dismissing the internet in 1995 – but treating current capabilities as the finished product is equally mistaken.

How The Spelker Team Uses AI Now

The most straightforward application is listing description generation. Writing polished MLS copy for syndication to Zillow and Homes.com requires evocative prose that Spelker says does not come naturally to him. He feeds detailed talking points into ChatGPT and receives a draft that captures the property’s features in publishable language. “I can just plug in an enormous amount of different talking points to AI and ChatGPT or whatever you use, and it’ll come back with a pretty nice write-up on the house,” he says.

More contested within the industry is Spelker’s use of AI for preliminary valuation research. He walks buyers – particularly those relocating from out of state with no frame of reference for local pricing – through a ChatGPT-driven process that pulls tax records, square footage, and comparable sales on the same street, then adjusts for size differences between properties.

The interactive nature of the process is what Spelker finds most valuable. Rather than producing a static report, the tool allows buyers to follow a line of inquiry in real time. “It would say, ‘Would you like me to show you other similar properties that sold on Maple Avenue?’ Well, yes, I would,” Spelker explains. “And then it’ll say, ‘Well, this is a really good comparison, but notice their house is 400 square feet larger than the one you’re looking at.'”

For buyers who have just met an agent and want to verify pricing independently, this gives them a way to do preliminary research on their own terms. Spelker frames it as additive rather than competitive with agent expertise – one more input in a decision that ultimately requires human judgment.

Where AI Cannot Follow

Spelker draws a firm line between AI as a research aid and AI as a valuation authority. The limitation is that AI cannot account for qualitative factors that experienced local agents carry: which streets command a premium, which neighborhoods have shifted in desirability, which homes have deferred maintenance that tax records do not reveal. “It’s not a science, it’s an art,” Spelker says. “There’s no magical formula.”

He points to automated valuation tools more broadly – including Zillow’s Zestimate – as an illustration. Those estimates have improved over the past decade, but they still fail when a home’s interior condition diverges from what the algorithm assumes. A seller whose home hasn’t been renovated in decades may see an estimate based on the assumption that finishes are current. “AI doesn’t know that your house hasn’t been touched on the inside in 35 years, and they’re assuming that everything is updated,” Spelker says.

For buyers, the practical implication is that AI-generated price estimates are useful for establishing a range but cannot substitute for an agent who has been inside the comparable properties and understands how specific conditions – a dated kitchen, a premium lot, a less desirable street – affect what the market will actually pay.

Ongoing Experimentation

Rather than adopting a fixed AI workflow, Spelker describes a process of continuous experimentation – identifying problems in his practice and testing whether AI offers a useful approach. One area is prospecting: using AI to develop strategies for reaching potential sellers who may be approaching a decision point. “I ask AI, ‘What are ways that I can reach potential sellers that may want to sell their houses, to maybe get them to cross over the line?'” he says. “We’re just scratching the surface.”

His approach treats AI as a supplementary research layer rather than a replacement for professional judgment. The technology handles tasks that are time-consuming but formulaic – generating listing copy, assembling preliminary comparable sales data, brainstorming outreach strategies – while decisions that require interpreting local market conditions remain with the agent.

For buyers and sellers working with agents who have spent the past few years learning AI’s capabilities and limits, the benefit is practical: faster preliminary research, better-written marketing materials, and more data points entering the conversation – filtered through someone who knows which data points actually matter in a specific Morris County, New Jersey neighborhood.

About The Expert: The Spelker Team, Scott and Amy Spelker, are real estate agents at Coldwell Banker Realty in Madison, NJ, consistently ranked among the top producers in their office. Scott is also Madison’s Town Historian.

This article is based on information provided by the expert source cited above. It is intended for general informational purposes only and does not constitute legal, financial, or real estate advice. Readers should conduct their own research and consult qualified professionals before making any real estate or financial decisions.

Disclosure: Individuals or companies mentioned may have a commercial relationship with KeyCrew.