, , , , ,

Using ChatGPT as a Real Estate Analyst – Sales Comp Analysis (Updated Aug 2023)

In early 2023, we here at A.CRE began exploring use cases for AI in commercial real estate. Since that time, we have released nine such posts covering topics such as using AI in the real estate job search, using AI to expand on Excel’s capabilities, using AI for real estate research, etc. However, this post on using ChatGPT as a real estate analyst is truly eye opening for me. It confirms for me that AI will fundamentally change how real estate professionals work.

As you may know, last week OpenAI released its new plugin feature (in beta). ChatGPT Plugins open the GPT-4 natural language model to external data sources previously unavailable to it. One such data source is Zillow. With the release of the ChatGPT Zillow plugin, it’s now possible to access Zillow data as part of a GPT conversation.

So, why does that matter? In this post + video, I’ll show you why it matters and hopefully give you a glimpse into what real estate analysis will look like in the future.

  • (Added Aug 2023) At the end of this post, I’ve added a second tutorial where I show you a Sales Comp Database ChatGPT Plugin I created to pull and analyze sales comps from my own (hypothetical) database.

A back-and-forth with ChatGPT as my real estate analyst in performing research for a hypothetical property tax appeal

What is GPT, ChatGPT, OpenAI?

Given that we’ve now shared numerous posts in this series, I assume you’re well aware of what ChatGPT and natural language models are. I also assume you’re familiar with its creator, OpenAI and the significant advances in artificial intelligence (including advances in artificial intelligence in real estate) that are currently underway.

Therefore, I won’t rehash in this post what GPT, ChatGPT, or OpenAI are. However, if these concepts are new (or you’re looking to learn more) you might check out our Using ChatGPT in Real Estate guide or our list of AI resources for real estate professionals.

What are ChatGPT Plugins (or Bard Plugins)?

Plugins are a feature that extends the capabilities of natural language models (NMLs) such as ChatGPT and Google Bard. With plugins, an NML gains the ability to access external data sources or services, enabling it to provide real-time information, answer questions based on the latest data, or perform complex computations.

For example, a plugin could connect ChatGPT to a weather API, allowing it to provide current weather conditions in response to user inquiries. Plugins also provide a means for integrating specific knowledge domains, such as legal expertise or property/market data enabling ChatGPT to offer more accurate and contextually relevant responses in those areas.

In my view, NML plugins will serve as valuable tools for commercial real estate professionals. It will enable us to access commercial real estate data from sources like CoStar (if/when they develop such a plugin), and thus use NLMs like ChatGPT or Google’s Bard as a personal real estate analyst to analyze markets, submarkets, rents, supply/demand, property values, etc.

Using ChatGPT as a Real Estate Analyst in a Hypothetical Property Tax Appeal

So, how might natural language models such as ChatGPT (or Google Bard) be used as a real estate analyst? In the following video, I connect ChatGPT to Zillow data via a Zillow plugin. Once ChatGPT has access to that data, I instruct the AI on what information I need it to source and how I want it to organize the data as part analyzing a hypothetical property tax appeal.

The traditional role of a real estate analyst is to source data (e.g. information from offering memorandums, comparable rents and sales, market data, etc), organize that data per the instructions of their superior, and then assist the firm making some real estate decision. I would estimate that at least 75% of an analyst’s time is spent sourcing information, and entering/organizing that information into some format and medium (e.g. an Excel or Argus model, Word Doc, etc) so that a more seasoned professional can digest it.

In the following video, you’ll see ChatGPT perform that exact role: source information and then organize it per my instructions. While plugins have not yet been created for commercial real estate data sets yet, I’m confident they will. And when that day comes (and it’s soon), the role of an analyst in commercial real estate will fundamentally change.

  • Click here to read the entire ChatGPT conversation shown in the video

Creating a ChatGPT Plugin to Pull and Analyze Sales Comps (Added Aug 2023)

Since I created this post in early 2023, I’ve begun to learn to build ChatGPT Plugins. This weekend, I created my own plugin that interfaces with a (hypothetical) proprietary sales comp database, enabling ChatGPT to source and analyze comps from that database. It showcases the potential of this technology with a deployable, real-world application.

In the video below, you’ll see how I employ the plugin to search for hypothetical comps within the sales comp database. ChatGPT then helps me organize those comps. The next step for me is to add my actual sales comps to the database, which will significantly accelerate my sale comp and valuation analysis process.

For those with OpenAI Developer access interested in creating a similar Plugin for their sales comp database, feel free to reach out. We’re more than willing to share the Replit project and Google Sheet template with Apps Script code to create your own Sales Comp Plugin.

About the Author: Born and raised in the Northwest United States, Spencer Burton has over 20 years of residential and commercial real estate experience. Over his career, he has underwritten $30+ billion of commercial real estate at some of the largest institutional real estate firms in the world. He is currently President and member of the founding team at Stablewood. Spencer holds a BS in International Affairs from Florida State University and a Masters in Real Estate Finance from Cornell University.