^{Image Source: ChatGPT generated image showing Generative AI in Commercial Real Estate} 

The commercial real estate (CRE) industry is undergoing significant changes, influenced by economic trends, shifting market demands, and technological advancements. 

An article released by Business Opportunities stated that “the real estate industry is ripe for disruption. Despite hesitancy in the industry, innovative technology is already making headway and changing how real estate transactions work.” 

This is particularly relevant today, as the US real estate market continues to expand, with a projected increase to $119.80 trillion by 2024, according to a Statista report. 

In this changing landscape, Generative Artificial Intelligence (AI) is emerging as a key technological development. Unlike traditional AI, which focuses on data analysis, generative AI can create new content and solutions, offering powerful tools for CRE lending and asset management. This technology is beginning to automate complex tasks such as property valuation and loan underwriting, making these processes more efficient and accurate. 

In asset management, generative AI is starting to transform investment strategies, predict market trends, and augment building management with data-driven insights. Its impact on the industry is becoming increasingly apparent: generative AI is not simply enhancing existing processes; it’s a catalyst sparking a new era in how we approach commercial real estate. 

Adding to this, a recent survey conducted by EY Financial Services in August 2023 sheds further light on the impact of Generative AI. 

Executives and managing directors from wealth and asset management firms, each with over $2 billion in revenue, were asked to identify the top three areas where Generative AI could most significantly affect their organizations. The responses indicated a wide range of impactful use cases throughout the value chain. 

Data ingestion to drive alpha-generating strategies and investment operations emerged as the primary areas of impact, followed closely by middle office operations, client onboarding, marketing, and client acquisition.  

However, it is important to acknowledge that the full potential of AI remains largely uncharted, even to those working within AI research labs. Companies such as Google, OpenAI, and Microsoft are exploring the frontiers of AI, but even they cannot fully predict the implications of AI on specific jobs or industries. The capabilities and impacts of AI are, to a large extent, still being discovered. 

Research into Large Language Models (LLMs) and their optimal applications is ongoing, with numerous significant studies already published. However, it’s clear that we are still in the early stages of this technological transformation. 

The best way to understand and leverage AI in any industry is through direct, task-specific applications. Each industry can gradually uncover how AI can be most effectively used, learning and adapting as the technology evolves. 

This article will explore how generative AI is changing the face of CRE lending and asset management, providing new opportunities for innovation and efficiency. 

1. Generating Real-time, In-depth Analysis 

2. Enhancing Underwriting 

3. Comprehensive Deal Due-Diligence for Lenders and Asset Managers 

4. Streamlining Credit Memo Preparation 

5. Facilitating Clear and Efficient Communication 

6. Automating Legal Review Processes 

To fully capture the transformative potential of generative AI in CRE lending and asset management, we have structured this blog into two parts.

Understanding Generative AI 

What is Generative AI? 

Generative AI refers to a type of artificial intelligence that can generate new content, ideas, or data that didn’t exist before. Unlike conventional AI which analyzes and interprets existing information, generative AI creates original output, such as text, images, or code, based on the data it has learned from. 

In the world of AI models and applications, there are currently a few standout players to consider. ChatGPT is regarded as state-of-the-art at present. 

In addition to ChatGPT, other leading Large Language Models (LLMs) include Claude 2, developed by Anthropic, and Gemini by Google.  

Key Technology Behind Generative AI: 

• Machine Learning (ML): This involves algorithms that allow computers to learn from and make decisions based on data. ML is the foundation of most AI systems, including generative models. 
   
• Neural Networks: These are a subset of machine learning inspired by the human brain. They are composed of layers of interconnected nodes or ‘neurons’ that process information and learn complex patterns in data. 

• Deep Learning: A more advanced form of machine learning, deep learning uses large neural networks with many layers (hence ‘deep’) to analyze and interpret complex patterns in large datasets. 

• Natural Language Processing (NLP): For generative AI dealing with text, NLP is crucial. It enables the understanding and generation of human language by machines. 

Differences Between Generative AI and Traditional AI: 

Purpose: Traditional AI is designed mainly for analysis, interpretation, and decision-making based on existing data. Generative AI, on the other hand, focuses on creating new data and content. 

Data Handling: While traditional AI models might classify, sort, or respond to data, generative AI models use their training to produce entirely new data that are similar to but separate from the training data. 

Complexity and Computational Power: Generally, generative AI models are more complex and require greater computational power than traditional AI models. This is because they need to understand and replicate patterns in data to generate new, coherent outputs. 

Generative AI represents a significant leap in the capabilities of artificial intelligence, moving from understanding the world as it is to imagining and creating things that never existed before. 

Beyond Basic Applications 

The field of AI holds seemingly limitless potential for business applications, yet its utilization in the Commercial Real Estate (CRE) sector appears somewhat relatively new. The question arises: how can AI be effectively deployed in CRE, and what new opportunities might it unlock? 

In a report released by ULI, an executive specializing in AI application remarked that “AI is going to change so many things so quickly at such scale that I don’t know that we have a good mental model to process what it means, not only for real estate but for society at large.” 

However, the current applications in CRE have mostly been mundane. The said executive also added, “When I think about all the ways in which AI may nip around the edges of real estate by changing how you interact with a customer service agent or property management, it still feels very low level.”  

In the same report, a real estate investment firm developer echoed this sentiment, with numerous tech tenants, including AI companies, concurs. “There are ways we can’t even fathom that will be helpful in all businesses. But the one I’ve heard of more recently is administrative tasks. It basically serves as your superpower assistant: lunch, property tours, running in the background handling all those sorts of things.”  

The Impact of Generative AI Today 

Generative AI is rapidly transforming the way businesses operate, bringing about significant improvements in task completion and efficiency. By automating complex, repetitive tasks, this advanced form of AI allows for quicker and more accurate completion of work. 

According to McKinsey, generative AI has the potential to automate up to 70% of tasks across various industries. This automation goes beyond simple routine tasks, extending to more complex processes that traditionally require human intelligence. 

The productivity gains from the implementation of generative AI are substantial. In data-heavy sectors such as finance and technology, generative AI is not just a tool for efficiency; it’s a catalyst for innovation. 

The integration of generative AI can lead to a productivity increase equivalent to up to 5% of industry revenue. This is because generative AI doesn’t just speed up processes; it also reduces errors and generates new insights, leading to higher-quality outcomes and more informed decision-making. 

Furthermore, the rise of generative AI is reshaping the nature of work through human-AI collaboration. This technology is not about replacing human workers but enhancing their capabilities. 

It supports a collaborative environment where AI handles the heavy lifting of data processing, and providing insights, enabling humans to focus on strategic, creative, and decision-making roles. 

As detailed in an Ernst & Young article, Generative AI is an emerging force in AI applications. It promises exceptional capabilities in handling tasks such as information search, retrieval, and synthesis, especially with unstructured content, along with remarkable content generation across various formats like text, images, and code. 

The proficiency of Generative AI in processing and contextualizing large volumes of information, and its ability to auto-generate responses that are easily comprehensible to humans, is inspiring leaders in the C-suite. They are envisioning how Generative AI could disrupt traditional business value chains, strategically position enterprises for future challenges, and create substantial value for all stakeholders. 

The influence of Generative AI extends beyond the corporate sphere, as it is poised to have a significant and enduring impact on society at large. 

Dr. Andrew Ng, an AI pioneer, famously compared AI to electricity, emphasizing its transformative potential—- “Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years.” 

In CRE, for instance, this technology is not only streamlining operations but also paving the way for more strategic decision-making in investment, lending, and asset management. Let’s explore the specific advantages that generative AI brings to commercial real estate industry finance. 

AI and Productivity: Transforming Workflows with Smart Technology 

As artificial intelligence (AI) continues to advance, compelling evidence is emerging about its substantial impact on individual productivity. Early controlled studies have shown remarkable time savings ranging from 20% to 70% across various tasks, coupled with an enhancement in output quality when AI tools are utilized. 

An experiment conducted by a team of social scientists in collaboration with Boston Consulting Group has further highlighted the transformative potential of Al in professional settings. This extensive study, considered the largest of its kind, explores the future of professional work in the AI era and involves 18 different tasks representative of those performed in elite consulting firms.  

The results were striking: consultants utilizing ChatGPT-4 consistently outperformed their counterparts who did not use the AI tool, excelling in every performance metric. 

In this study, together with social scientists from Harvard, UPenn, and the University of Michigan, revealed that consultants utilizing ChatGPT-4 achieved 12% more tasks, completed tasks 25% faster, and, most critically, the quality of their output was 40% higher than those not using ChatGPT-4.  

It’s important to note that these findings predate the latest enhancements to GPT-4. The recent advancements, such as an advanced data analytics mode, the integration of plugins, and updated web search capabilities, were not part of the study. This suggests that the productivity gains and quality improvements observed in the study might be even more pronounced with the latest upgrades to GPT-4. 

These insights offer a glimpse into the profound effects that AI, particularly advanced models such as GPT-4, can have on productivity and efficiency in professional environments. As AI continues to evolve, its role in reshaping workflows and enhancing productivity in various industries becomes increasingly clear, marking a significant shift in how work is approached and executed in the modern age. 

Generating Real-time, In-depth Analysis 

Generative AI is starting to transform the way investment analysis is conducted in the commercial real estate (CRE) sector. Traditional methods, where analysts would spend extensive hours analyzing data, are being replaced by AI algorithms capable of processing vast amounts of information in minutes. This technological advancement leads to faster and more accurate analyses, greatly enhancing decision-making efficiency. 

A key player in this transformation is GPT technology, a type of generative AI. 

As an AI-driven system, it has the capability to gather financial and operating data from loan -related materials and external sources such as borrower and property financial statements and budgets, property rent rolls, leasing agreements, market reports, etc. This significantly shortens the time needed for data intake, thereby speeding up the entire origination process 

The benefits of generative AI extend to the processing of extensive documentation such as information memorandums, market overviews, or due diligence materials. AI shows a promising efficiency in this domain, with the ability to scan, process, and condense large volumes of text into accurate summaries. This capability is particularly valuable in the CRE sector, where the ability to quickly synthesize and analyze complex documents is essential. 

This significant reduction in time—from weeks to just a few days—is a testament to the role of GenAI in streamlining the origination process. This initial stage, enhanced by GenAI, lays down a foundation of accuracy and efficiency vital for the subsequent stages of the CRE loan cycle. 

Limitations in Advanced Data Analysis

While GenAI has transformed data analysis, it’s important to recognize its limitations for effective application. 

• One key constraint is that GenAI primarily processes structured data. For example, it cannot directly analyze a scanned financial statement. However, this limitation can be overcome with third-party applications that preprocess and standardize data into a format conducive to GenAI analysis. 

For instance, Smart Capital Center serves this purpose, converting various vast amounts of unstructured data, including financial documents, statements, rent rolls into a standardized dataset for analysis. 

It takes various forms of data – scanned financial statements, rent rolls, excel spreadsheets, text entries, or complex financial reports, and converts them into a uniform dataset ready for analysis. This standardization process not only ensures compatibility with the analytical capabilities of Generative AI but also enhances the accuracy and reliability of the insights generated. 

• Another limitation lies in the accuracy of the output of Generative AI systems: 

Users must review and verify the results, especially in scenarios where precision is critical. The efficacy of GenAI is optimal when exact precision is not the paramount concern or when users have a clear expectation and seek visual confirmation or clarification from generated charts. Moreover, users can modify prompts to refine results. 

For example, consider generating a unit mix using a specific methodology that excludes vacant units. If a table includes a subtotal at the bottom, GenAI might initially misinterpret it as a separate unit type. 

However, in Advanced Data Analytics mode, GenAI can identify such outliers, suggesting they are likely subtotals to be excluded, but still necessitating user confirmation. This synergy between the system and human analysis proves most effective. 

Explicit instructions, such as directing GenAI to ignore potential subtotals, can yield accurate unit mix results.  

Hence, at Smart Capital, leveraging the power of Generative AI, we’re introducing an option for users to request customized charts and analytics generated in real-time, tailored to their specific use-cases.  

This dynamic feature is engineered to accommodate a vast array of specific requirements and scenarios that may be required in various situations. If an analyst needs a visual representation of a certain specific financial trend or a certain comparative analysis that is not part of the standard set of visualizations, the GenAI-powered system can generate these in real-time based on user prompt. 

For instance, a user might need to analyze a unit mix while excluding vacant units or require a financial projection that accounts for variable market conditions, the Gen-AI powered platform is equipped to handle such nuances and it is designed to interpret the subtleties of user request in a much simpler, faster way, requiring much less manipulations from the user. 

We’ve built our platform to ensure that users remain in control. After the initial generation, users can review and verify the results. 

Enhancing Underwriting 

Generative AI is transforming the field of underwriting for lenders and asset managers by bringing automation and better insight into many steps of the underwriting process. 

In the underwriting process, generative AI systems synthesize vast amounts of financial data, interpret complex patterns, and predict future financial health, which are pivotal for making lending decisions. 

AI can help integrate and interpret data, such as data from financial statements, balance sheets, rent rolls, development budgets, and financial, and operating projections, to assess a borrower’s financial stability and the profitability of a property or construction project. Advanced algorithms evaluate historical financial performance and project future cash flows, enabling underwriters to identify risks and opportunities that might not be immediately apparent.  

Automation systems powered by generative AI can model debt service capabilities by incorporating real-time data feeds, and keeping underwriting assessments up-to-date with current property performance and market conditions. It can quickly adjust analyses based on fluctuating interest rates, changing property performance, market dynamics, and other factors.  

This dynamic approach to financial analysis ensures that lenders and asset managers have access to the most relevant and accurate financial information when making underwriting and portfolio management decisions. 

By leveraging the power of generative AI in financial analysis underwriting, the process becomes not only more efficient but also more predictive and adaptive to market changes. This results in more informed, data-driven decisions that can better mitigate risks and capitalize on viable lending opportunities. 

Document Review and Approval 

With AI, the review and approval of documents, such as insurance policies, appraisal reports, environmental assessments, etc., become less time-consuming. 

AI algorithms can quickly scan through documents to identify key terms, conditions, and clauses. This capability allows for the fast assessment of whether a document complies with set standards and regulations, drastically reducing the time needed for manual review. 

AI systems are also equipped to detect potential risks and issues within documents by analyzing patterns and inconsistencies that may not be immediately apparent to human reviewers. This includes identifying discrepancies in appraisal reports or spotting potential issues in environmental assessments ensuring that all potential risks are flagged for further review. 

By automating the workflow, AI can route documents to the appropriate parties for review and approval based on predefined criteria. This not only speeds up the process but also ensures that all necessary approvals are obtained in the correct sequence, minimizing bottlenecks. 

Beyond just reviewing documents, AI can provide predictive insights based on the content of the documents and historical data. For example, AI can predict the likelihood of an insurance claim based on the details of an insurance policy and historical claim data, helping in decision-making. 

To discover further applications of generative AI in CRE, head to Part 2 of this article. 

---