By Brian Hemmi

Senior Director, Emerging Technology

The commercial real estate (CRE) industry has undergone significant and transformative changes over the recent decades. These changes are primarily driven by advancements in technology, including enhanced connectivity of devices which facilitate a greater geographical distribution of projects. Additionally, improvements in hardware and software have enabled broader participation and increased transparency in operations. Most notably, recent milestones in the artificial intelligence (AI) landscape are reshaping the industry. In this article, we explore some of the most recent historical milestones in AI, some current applications in the CRE sector, and potential future developments in the near and long term.

A brief history of AI advancements

To better understand where the future may take us, we will first look at some of the most recent (last ~15 years) advancements in the AI landscape and how these innovations build on each other to better understand how this technology may evolve in the future. During this time, we have progressed from simple models to more advanced, data-driven models that are used in all of our lives and the interactions we have every day. Each of these phases has played a crucial role in laying the groundwork for the next whether they contributed significantly on their own or not.

Pre-Deep Learning (< 2010)

Before deep learning became mainstream, due to the implementation of graphics processing units (GPUs), the AI landscape was characterized by lengthy training times. During this period, computing power limitations was a considerable bottleneck in model training. As a result, the models developed were slower and primarily suited for simpler tasks.

Deep Learning and GPUs (2010+)

The 2010s marked a significant shift with the rise of deep learning and the exponential increase in computing power, leading to massive increases in supercomputer compute capacity with the largest supercomputers from the beginning of this period to now increasing capacity per computer by 500x. This surge allowed the training of more complex models and sparked an AI boom. Applications such as language translation, image recognition, and recommendation systems became mainstream, transforming various industries, including CRE.

Reinforcement Learning

Reinforcement Learning (RL) has further advanced AI by facilitating self-guided training. Unlike supervised learning, which relies on humans to pre-tag data, RL allows models to learn by interacting with their environment and optimizing their actions to achieve predefined goals and objectives. While not a major milestone on its own, this capability paved the way for the advanced models we use today.

Large Language Models (LLMs) (2020s)

The release of ChatGPT by OpenAI in 2022 marked a noteworthy shift in the sentiment of AI and made the technology much more real to the average person. Before this, most interactions with AI systems were behind the scenes where the underlying model was task-based and made for one specific purpose. The emergence of LLMs marked the first time direct interaction was enabled with a system that anyone could use for a multitude of purposes.

These models are having a profoundly transformative shift on the AI ecosystem and have the capability to do for information what search did for the internet. Simply put, LLMs provide everyone easy access to the vast data they were trained on and are even beginning to be used to perform search functions by fetching the relevant data and filtering out the unnecessary information and condensing it into a ready to consume format.

This advancement has also allowed organizations to implement AI into their business operations with relative ease and significantly enhance employee efficiency by providing access to information they need to perform their jobs more effectively.

These models have also sparked a new rush in development and investment in the AI space, which was starting to wane due to diminishing returns. This increase in funding has driven the exponential growth of these models, so they can now answer more complex questions and decreasing the time it takes to make improvements.

What the future may hold for CRE

As AI continues to evolve, its potential applications within the CRE industry are expanding rapidly. The future promises many innovations that will further streamline operations, enhance decision-making, and improve efficiency. By harnessing the power of AI, CRE professionals can anticipate advancements in real-time valuations, underwriting processes, asset surveillance, and market monitoring. Additionally, long-term developments in environmental modeling and AI-driven architectural design will reshape how properties are planned, built, and managed. Below, we explore both the immediate and long-term impacts of AI on the CRE sector.

Short Term

AI Agents and Automated Workflows: AI agents (or teams of models) running independently and coordinating their specific goals and tasks will be able to monitor the market landscape around the clock and process information automatically helping human employees to be the decision makers and the final say. They can take inputs from a company’s systems, use that to create information gathering tasks for other agent-based models to go out to the open internet to gather, aggregate and refine. This refined information is then turned into ready to ingest reports for individuals and systems to create continuous alerting and monitoring, creating for a smarter workforce that is performing higher value, more thought-based tasks rather than data entry and research.

Real-Time Valuations: AI is set to enable real-time property and position valuations in CRE finance. By swiftly processing documents, cash flow statements, and gathering relevant market data, AI can provide accurate and timely property valuations. This capability will provide quicker decision-making and improve investment strategies.

Faster underwriting (UW) and Transaction Time: AI can more adeptly speed up the underwriting process by analyzing cash flows, financial structuring, market/appraisal data, cost of capital, and other firm-specific calculations. Automated underwriting will allow for a higher volume of potential deals to be assessed, providing more comprehensive information for informed decision-making and reducing transaction times and the subjectivity of these decisions.

Long Term

Better Development Planning: Advances in environmental modeling, such as what is happening with Nvidia’s Earth-2, will provide more accurate planning of development projects. By utilizing the vast compute power and modeling capabilities of these simulations, property developers will be able to plan projects around what the environment is like today and likely to be in the near future. This would allow them to create buildings more suited to their environment and capable of withstanding the storms and disasters of today and the future. Climate modeling will also support the construction of more efficient buildings increasing the sustainability of building development and reducing operating costs.

AI Architects: Future AI systems will design buildings more efficiently, incorporating advanced construction techniques and materials to reduce costs and enhance energy efficiency. AI architects will streamline the approval process with local municipalities and create designs that are both sustainable and efficient, ultimately transforming the landscape of commercial real estate development.

Conclusion

The ongoing evolution of AI will continue to have profound impacts on the commercial real estate industry. As models become more advanced, every aspect of the industry has areas that will be impacted for the better. From decreasing the cost and time required to develop new buildings and making them more resilient to their climate, to increasing the availability of information which speeds up and enhances decision-making on potential deals. Consequently, we can expect to see the acceleration of numerous positive impacts on the way the industry operates today.

---

Brian Hemmi, Senior Director, Emerging Technology at Trimont, has seven years of experience in real estate technology and currently leads Trimont’s AI initiatives.

---

CRE Finance Focus is Trimont’s bi-weekly LinkedIn newsletter delivering distinct perspectives on the most pressing topics of today from industry experts. Keep your finger on the pulse of the #CRE market and subscribe today.

Have questions or topics you would like to see covered? Submit them here.