When Barbi Reuter, CEO and principal of Cushman & Wakefield | PICOR Commercial Real Estate Services in Tucson, Arizona, went to write an opening statement for an annual awards ceremony, she decided to open up the artificial intelligence chatbot ChatGPT to perfect her tone.

“I prompted, ‘Write a compelling opening for the CEO at the 37th annual awards luncheon,’” Reuter told CoStar News. “I opened our banquet with it and then fessed up on the source. The team was pretty incredulous!”

Since OpenAI’s chatbot launched in November, real estate professionals across the country have experimented with it to help with tasks ranging from copy editing to tweaking the tone of an email to a disgruntled client. But while executives and brokers are still exploring AI’s many uses — ChatGPT still has bugs and accuracy issues it’s working out — there’s agreement it holds the potential to make some industry tasks easier.

Right now, AI’s biggest benefit is efficiency, according to commercial real estate professionals who joined in on a Twitter conversation about ChatGPT. Users said they asked it to do once-overs for grammar and tone in marketing materials, property descriptions, letters of intent and right of first refusal contracts, to name a few examples.

The technology can also help with lease reviews, comparing information against an agreed-upon letter of intent and avoiding bias in favor of the landlord or tenant, users said. Beyond that, it has basic coding capabilities and can perform data analysis to identify market patterns and risks.

Some industry professionals are projecting more significant capabilities in the future.

“In its current state, I think of it as a tool to assist in improving daily efficiency, but in the long term, I think AI will be transformative to the industry by eliminating repetitive and low-barrier-to-entry jobs,” Bronwyn Scrivens, an associate broker at Omada Commercial in Edmonton, Alberta, wrote in an email to CoStar News.

AI has the potential to reshape the brokerage and real estate listing service sectors the most, according to Kunal Lunawat, co-founder and managing partner of Agya Ventures, a San Francisco-based real estate technology fund.

ChatGPT isn’t yet able to access real-time real estate data, but other AI technology has the potential to significantly improve real estate listing searches. For example, it could narrow down a search based on a client’s preferences to about 90% accuracy, and then brokers could help pick through the remaining 10%, Lunawat said.

Individual brokers, especially the most technologically savvy, have been quick to jump on the ChatGPT trend, but it may take longer for companies to embrace it. Although the chatbot has more than 100 million users, it’s been around for less than three months and still has some major limitations.

OpenAI acknowledged the bot can sometimes write incorrect answers or get confused in what’s called an artificial intelligence hallucination. Its understanding of more recent events that occurred since 2021 can also be muddled, users have found.

“There is potentially concern for plagiarism and fostering the ‘lazy’ tendencies of not coming up with your own ideas,” Scrivens said. “Commercial real estate organizations should get ahead of this technology and invest in learning the best ways for their employees to guide the conversation while we all learn best practices together.”

CoStar News reached out to OpenAI for comment but did not immediately hear back.

The importance of authenticity and personal relationships, which is particularly significant in commercial real estate, could slow the adoption of AI, Lunawat and Reuter both said. Professionals are required to reflect a personal or corporate brand in their communications, which are often more consultative and less transactional than residential.

“If we’re borrowing someone else’s words and they don’t sound like our own, we dilute our credibility,” Reuter said.

While companies grapple with the ethics of AI, they could be missing major opportunities, however.

“There are like billions of dollars of industry value up for grabs,” Lunawat said. “Any sophisticated brokerage house should be thinking about how to inject some of these AI, large language models as effective tools for their brokers.”

Real estate startups that are more up on technology trends and less risk averse may jump ahead of big-name companies when it comes to adopting AI, Lunawat said. Facing an economic downturn and a general slowdown in commercial transactions, many of the larger players may have too many other concerns to invest heavily in the ambitious bet of AI.

It doesn’t help that the real estate industry has historically been slower to embrace new technology relative to other industries, said Casey Flannery, a senior associate at Foundry Commercial in Nashville, Tennessee, who hosted an online forum about ChatGPT in February.

“We have to think creatively how we would use it in our business, and it is a bit hard to see at the moment with it being so new,” she said.

Residential listing marketplaces should be thinking “deep and hard” about how to incorporate AI if they want to stay relevant, as AI’s capabilities are only going to snowball, Lunawat said.

Large language models like OpenAI’s chatbox use reinforcement learning, meaning it can grow more accurate as it works with a user to search for information. Other up-and-coming AI prototypes are able to not only render multiple designs within minutes but also predict emissions levels and material procurement information.

“We haven’t even seen [the next generation] GPT-4 come out yet,” Lunawat noted. “There’s folks in my portfolio who have seen it, who had early access, and they’ve been blown away by what it can do.”