How BGO Uses AI to Outperform in Real Estate Investment

BGO Uses AI to Transform Investment Research Models

BGO, a global real estate investment manager overseeing $89 billion in assets, is applying artificial intelligence to reshape how it approaches investment research. Their AI-driven model revealed a key insight: the success or failure of investments hinges entirely on the choice of the local market. This breakthrough guided BGO’s decision to invest in an industrial development in Las Vegas alongside Northpoint Development. Today, investors hold more than 131,000 homes in the Las Vegas Valley.

Investment Strategy and AI

John Carrafiell, co-CEO of BGO, sits directly beside his chief data scientist, highlighting the firm’s commitment to data-driven decisions. While investment strategy has always involved research, AI has transformed traditional models that hadn't evolved much over decades.

Carrafiell, with nearly 40 years in real estate, noticed the sector’s research methods often recycled the same data leading to similar conclusions. His critical question was, “How do we really outperform?”

The answer came from analyzing two decades of BGO's past deals using a computer model that removed human bias. The model showed that investment results depend fully on local market fundamentals, not property pricing or national trends. This shifted BGO’s focus sharply onto local market dynamics when selecting future investments.

Building a Better Model

Existing research firms that rank local real estate markets often produce inconsistent results. BGO built its own model to backtest drivers behind their best and worst performance outcomes. This model incorporates numerous local data points like demographics and supply trends unique to each market. AI enhances the model by allowing faster analysis of larger data volumes.

Carrafiell explained, “We use thousands of data inputs, including free government data and paid sources like telecom providers. We found the key factors and confirmed accuracy through backtesting.”

Successful Investments

BGO’s data-driven approach led to an investment in a Las Vegas industrial development with Northpoint Development. While other models predicted mediocre returns, BGO’s AI model indicated strong potential.

The model forecasted rents around $5.88 per square foot, yet actual rents hit $9 per square foot. The analysis identified that rising costs in California’s Inland Empire were pushing companies to seek more affordable logistics hubs. Las Vegas offered lower rents, taxes, and labor costs, making it an attractive alternative.

Future Developments

BGO applied similar analytics to investments in Florida and the Rust Belt, achieving substantial returns. Carrafiell credits the model with materially improving performance but acknowledges that no model can predict unpredictable events with complete accuracy.

The investing team focuses on upside potential, while the lending team uses downside models to assess risk. Future versions of the model will incorporate asset allocation across commercial real estate sectors, aiming to recommend optimal portfolio mixes.

Carrafiell summed it up: “AI accelerates what we do, but it’s really data science at the core.”