Can AI Really Put A Price On Yorkville Condos?

Where Artificial Intelligence works and where it falls short with Toronto Real Estate

More and more people are using artificial intelligence for all sorts of tasks including evaluating condos in Toronto.  But before you let AI seduce you from your human reasoning here are some things to consider.

Publically available AI models such as ChatGPT, Claude and Gemini are only as good as the websites they source.  These Large Language Models pull the sales data from the MLS (Multiple Listing Service).  For the public the MLS may be just another website to peruse properties for sale but it is actually the primary website from where property listings originate.  All other real estate websites get their feed of condo listings from the MLS.  If one provides their email address they may even have access to sold prices but viewers are only allowed to see info made for public display.  

However Toronto real estate agents' access to the MLS is more robust and comprehensive.  Realtors acquire much more information including pricing history and notes on the condo unit that are for agents only known as Broker-only remarks.  This info never makes it on to public real estate websites and therefore is never analysed by publically available AI.

The sold price of the condo only tells part of the story.  Just working with sold prices alone creates an incomplete picture without context. Real estate agents use the MLS to get live MLS data, compare condos and evaluate most recent solds.  Realtors then dig deeper identifying off-market sales (condos that sold but are not posted on the MLS) and failed listings. But real estate agents expertise also goes beyond their upper-tiered access to the MLS. 

AI is purely data driven but buying and selling Toronto condos extends beyond the data. Its also deeply human and more sensory driven. Unlike a human, AI cannot experience a condo unit in person.  It can’t physically walk through a unit to get an overall feel for the suite. It can't detect the smell of cigarette smoke nor step out on the balcony to hear the traffic noise. It can’t see the view overlooking the park or the garbage dumpster.  It is these detailed tactile experiences where AI falls short. The most important features of condos are often the least quantifiable.  AI cannot measure these nuances of how a condo feels nor how a buyer responds.

AI can be a worthwhile tool at the beginning stages of buying or selling.  For condo buyers, the initial research of identifying locations that feature must-haves can assist in narrowing down the search to the right neighbourhoods and condo buildings offering preferred amenities. It can point you in the right direction as to where to start looking. For sellers one can get access to general information and prices and common market conditions but without the the why.  Why one unit with the exact same floor plan sells for more than another.  

Unlike traditional search engines such as Google,  LLMs are advanced prediction models.  When you ask AI a question it isnt getting the answer by searching webpages and databases across the internet, but rather its answer comes from predicting the most likely next word based on patterns it learned during training.  AI learns from the data rather than storing it.  If the training is weak it may fill in the gaps and fabricate answers which may be plausible but untrue and this is known as hallucinating.  

I was recently showing a buyer client a Yorkville penthouse and before we entered the suite he asked me why the unit was up for sale again when it didnt sell last time and now the Seller was asking a higher price.  He said the Seller had tried to sell his penthouse a few years back at a lower price and that was when the condo market was booming.  After learning he got his info from AI, I revealed to him the correct sales history for the suite.  The penthouse was indeed up for sale a few years back like my client said but the unit had actually also sold at that time.  The truth was the new owner was now selling it at a lower price than he had purchased it for.

AI models often present information with certainty and authority. Its hard not to be lured in as AI seems so helpful and agreeable that you stop questioning it. But over reliance on AI can lead to overlooked property flaws, missed opportunities and mostly importantly costly financial missteps.

AI and its predictions create an incomplete picture.  An experienced real estate agent adds the context by having the expertise and knowledge to interpret the data and understand the nuance based on years of experience physically walking the neighbourhoods, touring the condo buildings, and actually transacting and negotiating the sale of condominium suites.

Nov 16, 2026