Two years ago, McKinsey estimated that generative AI could yield $180 billion-plus in value for the real estate industry; a figure I’d imagine has only gone up since. The potential of AI in our industry is huge, yet I’d argue that many real estate professionals are leaving a lot of its value on the table.
When ChatGPT was released in late 2022, excitement quickly reached a fever pitch. Usage surged in early 2023. However, by the following year, usage had stabilized and growth had flatlined. In discussing this and other AI trends with brokers and advisors at Engel & Völkers’ EVX conference, many corroborated this experience in their own businesses.
They were eager to harness the power of AI and quickly jumped into using tools like ChatGPT. But once the novelty had worn off, they realized they were essentially using the tech as a substitute for Googling or, perhaps, asking it to execute simple tasks like writing a listing description, with varying degrees of success.
The good news is that getting more out of AI doesn’t necessarily require a huge monetary investment or a complete business plan or tech stack overhaul. A few simple changes to the way real estate professionals are using existing AI tools can yield big benefits.
Here’s how:
Step 1: Document your processes
Driving incremental value with AI really comes down to how you’re training your tools and the quality of prompts you’re creating. We all know “garbage in, garbage out”; by the same token, quality prompts and comprehensive knowledge bases yield quality results.
Many brokers and agents can feel overwhelmed by this task; after all, they are real estate specialists, not AI specialists.
San Francisco advisor Blakely Hull tackled this challenge in perhaps a surprising way; she began by documenting all her processes the old-fashioned analog way, with pen and paper. Of course, agents and brokers can document their own processes in any way that works for them, but the important takeaway is to start doing it.
Begin with one task at a time, like everything you do when you take a listing, and keep going from there. Pretty soon, you’ll have your own “special sauce” documented, and you can then train your AI tools with this intel. Now, you’re beginning to move away from generic AI responses to creating a tool that can automate your specific business processes.
Step 2: Use your own data sandbox
Training AI on your own data goes hand-in-hand with documenting your processes. Incorporating the data that makes your business unique, and pairing that with external data around market trends, etc., is what will allow AI to really perform.
Geoff Bray, a license partner in Minneapolis, found that he was really able to see the benefits of incorporating AI tools into his business when he opened up his data to Claude.
Claude, and other AI tools, give users the ability to save project folders with prompts for things they do on a regular basis. This becomes especially powerful when users allow the tool to access their own data instead of just relying on what AI can pull from the web. The more specific and unique information a user gives, the more the AI tool will sound like the user and incorporate brand-specific language.
When we give AI access to our own data, we begin to see true competitive differentiation because now you’re using the technology in a proprietary way.
Step 3: Move beyond using AI as a search engine
With processes documented and data sandboxes open, real estate professionals can really begin to watch the magic happen. Here is where users can make the move from asking AI questions (What were the top five sales in my market last month?) to giving it true commands (You are a copywriter specializing in real estate marketing. Taking into account the top five sales in my market last month, write a listing description that will most resonate with millennials with no children, earning a household income above $300,000 who already own more than one home).
Hull uses Sidekick AI to create detailed buyer personas that enable super-precise marketing. As an example, she shared the process for a recent $2 million listing. The home was a straight shot to downtown San Francisco. It was spacious, with three bedrooms and two baths, and had a lot of stairs. Using all of the listing information and Hull’s own data, AI helped to create buyer personas for those most likely to purchase the home.
In this case, some of the personas that emerged were tech workers being called back into the office who don’t want to live downtown but want an easy commute, or families looking for more space with kids who are old enough to manage all the stairs.
An important takeaway from Hull is not to hold back on the information you’re giving AI. Pepper it with information even if it doesn’t seem relevant: All of the pros and cons of the property, its location, its specs, etc. Now you can tailor your marketing efforts with a new degree of specificity.
This is the difference between asking AI to simply create a listing description or social media copy for a listing and asking it to create these things targeted to a specific persona.
An added benefit is that this forces users to be extremely thoughtful in their approach to marketing and may result in rethinking some practices that have become the status quo.
It’s been said before, but it’s worth repeating: Agents who use AI will replace those who don’t. However, it’s not a magic bullet. AI alone is not going to uplevel or enhance your business. Agents, brokers and real estate professionals across the board have to be creative and intentional in setting up a solid foundation from which the AI can work in order to see real value over the long term.
Colin Hull is the VP of digital marketing, Engel & Völkers Americas. Connect with him on LinkedIn.
