The explosive growth of generative artificial intelligence (AI) has affected the real estate industry through various applications. ChatGPT, DALL-E 2, and other large language models (LLMs) have captivated society and brought generative AI to the forefront of innovative technology. 

According to JLL’s 2020 Global Real Estate Technology Survey, AI and generative AI were ranked among the top three technologies expected to have the greatest impact on real estate.¹ The McKinsey Global Institute suggests generative AI could generate between US$110 billion and US$180 billion (or more) in value for the real estate industry.² 

Traditional AI vs. generative AI: What’s the difference? 

Traditional AI, also known as weak AI, focuses on performing a specific task based on predetermined algorithms and rules. Examples of traditional AI include voice assistants like Alexa. Traditional AI cannot create anything new and is trained to follow specific rules to perform a particular task. 

Generative AI can produce text, images, video, code, and other types of content. Generative AI advances traditional AI by using machine learning algorithms to generate output based on a training model. 

Generative AI has revolutionized the process-oriented real estate industry by enhancing decision-making capabilities and operational efficiency. AI tools have already been implemented to better manage property operations, such as energy management, rent collection, lease tracking, and maintenance requests. 

Specific uses of generative AI in commercial real estate 

Generative AI automates the acquisition process through information gathering and filtering, such as summarizing research reports, presentations, and government policies. These manual processes usually take real estate professionals a significant amount of time. Due diligence efforts become less laborious and more reliable and efficient with generative AI. 

Real estate investment decisions are often made based on individual analysis and qualitative opinions. Generative AI can analyze large data sets, identify trends, patterns and anomalies, and predict future real estate market trends. Use cases for generative AI include using predictive analytics to provide greater precision in cash flow forecasting, property valuation, yield, and other factors that traditional analysis methods may overlook. This approach can increase accuracy and reduce the likelihood of human error and subjectivity in financial modelling. 

Real-time valuations 

Traditional cash flow management often involves static reports of the financial landscape, while generative AI enables real-time cash flow, valuation, and rent monitoring. For example, generative AI can analyze power usage by performing real-time analysis of power monitoring and generating real-time remaining power predictions based on actual conditions. This enables real estate organizations to manage rents, fees, and revenue better. In addition, generative AI can run multiple valuation models and produce outcomes based on different parameters and assumptions. For rent monitoring, generative AI can analyze rental market trends, predict rental changes, and generate recommendations based on market demand and asset characteristics. This also helps landlords of shopping centres, grocers, offices, industrial parks, manufacturing facilities, and more make more informed rental decisions. 

Risk management 

Traditional risk assessment models can be enhanced by predicting potential downturns, market volatility, geopolitical events, and economic indicators. Generative AI can help identify tenants at risk of defaulting on rent payments, reduce costs associated with accounts payable, and analyze patterns for liquidity needs. 

Enhancing the customer experience 

Generative AI can enhance customer relationship management, including investor relations. It can be used to create marketing materials and investor presentations to target potential investors and maintain ongoing relationships. 

On the property management side, using AI-powered chatbots and virtual assistants can support real estate organizations in answering customer questions promptly, managing tenant requests, and executing lease negotiations. 

Limitations and ethical considerations of using generative AI 

Generative AI is not viewed positively by all stakeholders. A major concern of using generative AI is the accuracy and reliability of the data used to train the models. Generative AI’s capabilities are limited by the size of its network and the amount of data it trains on. The phrase “garbage in, garbage out” is commonly used in financial modelling. This means that the output’s quality depends on the input’s quality. The same rationale applies to developing a large language model. In generative AI, “AI hallucinations” occur when large language models generate inaccurate or misleading information and produce outputs that are not based on training data. If real estate organizations do not train AI systems appropriately, they can produce inaccurate output. 

AI hallucinations can be prevented by using high-quality training data that is diverse, representative, and bias-free. Data collection should include a wide range of scenarios, factors, variations, and contexts to help reduce the chance of AI hallucinations. Additional training techniques, such as reinforcement learning from human feedback (RLHF), a machine learning technique that uses human feedback to improve language models’ learning, could also be applied on top of the initial training while building the AI-powered chatbots to reduce misinformation in the language model. “Red teaming” is another method of testing an AI system. In red teaming, a group of people challenge a system by thinking like attackers and injecting unexpected prompts to test and evaluate model vulnerabilities for undesirable behaviours. 

Data privacy and security of personal data are critical elements for organizations to consider when implementing generative AI. Generative AI systems often require access to large amounts of data for training and operation. Real estate organizations have tenants’ and clients’ sensitive data and banking information. If a generative AI system is not secure, it could be vulnerable to data breaches and cyberattacks, resulting in the theft of sensitive data. 

Organizations must also ensure that their language models comply with changing AI ethics guidelines and regulations. For example, marketing material generated by AI may require additional checks for branding and regulatory compliance. 

For organizations developing large language models to enhance their operational processes, transparency practices must be implemented to prevent “black box” scenarios where the system’s internal workings are hidden. By implementing “explainable AI,” organizations can clearly express why an AI system reached a particular decision or prediction and build stakeholder trust. 

Conclusion 

The world of real estate is on the brink of transformation, as generative AI emerges with the power to revolutionize the industry. Generative AI can be a powerful tool for automating mundane tasks, providing sophisticated analytics for investment decisions, enhancing customer experiences, and offering real-time valuations and risk management. Despite the benefits, adopting generative AI requires careful consideration of its challenges and responsibilities. Real estate organizations must navigate the complexity of data accuracy, privacy concerns, and ethical considerations. Integration of generative AI requires real estate organizations to proactively invest in their information technology infrastructure and embrace the new technology landscape, while ensuring it is used effectively and ethically. 

¹ JLL. “Is Your Real Estate Technology a Value Driver?” JJL. Accessed February 3, 2024. 

² Fitzpatrick, Matt, Vaibhav Gujral, Ankit Kapoor, and Alex Wolkomir. “Generative AI Can Change Real Estate, but the Industry Must Change to Reap the Benefits.” McKinsey Quarterly. McKinsey & Company, November 14, 2023.