The Elephant in the Room
The property podcast for the thinking person.

Episodes

Episode 440: Luke Metcalfe: Why AI Can’t Pick Property Winners

AI is rapidly becoming one of the most popular property research tools available to investors. 

From identifying growth suburbs to analysing housing market data, many people now turn to AI for faster insights, smarter decisions, and greater confidence in their property investment strategy. But as these tools become more common, an important question emerges: can AI predict property prices with enough accuracy to give investors a genuine advantage?

In this episode, Luke Metcalfe from Microburbs puts AI property market predictions to the test. Drawing on thousands of suburb forecasts generated by multiple GPT models and comparing them against real-world market performance, he examines how accurate AI is for property investing and whether these systems can reliably identify future growth opportunities. The findings challenge many assumptions about AI-driven property research and reveal why technology alone may not be enough to outperform the market.

Luke explores why AI fails at property predictions, particularly in a market as complex and dynamic as Australian real estate. While AI can process vast amounts of housing market data and uncover patterns quickly, it is ultimately trained on historical information rather than future events. This creates significant limitations when unexpected factors such as policy changes, economic shocks, shifting buyer sentiment, or supply constraints reshape market conditions. The discussion also highlights the risks of AI in real estate investing, including the tendency for different models to produce similar recommendations based on the same underlying datasets.

The conversation also examines some of the most common property investment mistakes investors make when relying too heavily on data. Luke explains why median price statistics can be misleading, how consensus-driven thinking can push buyers toward overcrowded opportunities, and why many property market trends are already reflected in prices by the time they become widely discussed. He also questions whether some AI-generated reports are providing genuine insight or simply repackaging information that investors could access elsewhere.

Beyond the limitations of AI, this episode focuses on what actually drives long-term capital growth property performance. Luke argues that successful property market predictions require a deeper understanding of owner-occupier demand, supply constraints, neighbourhood characteristics, and street-level dynamics. Rather than viewing property as a single market, he explains why investors should think of it as thousands of interconnected micro-markets, each influenced by unique local factors.

For investors looking to improve their property investment strategy, this conversation offers a practical framework for evaluating data more critically and avoiding costly assumptions. Whether you're curious about real estate AI tools, interested in better property data analysis, or simply trying to understand what drives long-term performance, this episode provides a timely look at where technology helps—and where human judgement still matters most.

Listen now to discover why following the data isn't always enough, and what investors should focus on when making decisions in an increasingly AI-driven property landscape.

Episode Highlights

01:27 – Meet Luke Metcalfe: Data Meets Property Reality

03:35 – AI vs Reality: Why Predictions Underperform

05:35 – AI Reports & Buyer Agents: Risky Shortcut?

13:52 – Selling at a Loss: A Hidden Growth Signal

16:18 – Humans vs AI: Who Gets Forecasting Right?

22:49 – Budget Shock: How Incentives Just Changed

26:52 – Supply, Scarcity & Why Tranquillity Wins

29:51 – Fast Food & Growth: The Signal No One Expects

31:44 – Rezoning Risks: Predicting What’s Coming

35:14 – Melbourne vs Sydney: A Tale of Two Markets

37:23 – Investor Incentives Are Shifting Fast

39:35 – Holding Property: What the Data Reveals

43:09 – Rental Crunch: What’s Really Driving It

49:33 – Airbnb & Second Homes: Supply Squeeze

52:24 – Yield vs Growth: What Actually Matters

55:15 – Model Accuracy: Where Data Goes Wrong

57:54 – Key Takeaways: What Investors Must Know

About the Guest

Luke Metcalfe is the founder of Microburbs, a property analytics platform focused on uncovering the underlying drivers of capital growth at a granular level. With a background in data science and machine learning, Luke has spent years analysing Australian property markets using large-scale, longitudinal datasets that go beyond traditional metrics like median prices.

His work centres on understanding property performance at the street and individual asset level, incorporating factors such as resale data, supply dynamics, and behavioural patterns. Known for challenging conventional wisdom in property investing, Luke combines technical expertise with practical insight to help investors move beyond trends, hype, and surface-level data.

Through his research and platform, he continues to push for a more evidence-based approach to property decision-making—one that prioritises real outcomes over popular narratives.

Connect with Luke

Resources:

Visit our website: https://www.theelephantintheroom.com.au

Enjoyed the podcast? Don't miss out on what's yet to come! Hit that subscription button, spread the word and join us for more insightful discussions in real estate. Your journey starts now!

Chris Bates