Inman

Real Brokerage CTO: AI ‘extremely useless without proprietary data’

Craig C. Rowe; Canva

This report is available exclusively to subscribers of Inman Intel, a data and research arm of Inman offering deep insights and market intelligence on the business of residential real estate and proptech. Subscribe today.

Like so many of its peers in proptech, The Real Brokerage has plowed into artificial intelligence, announcing in May its own GPT integration, or Chat Generative Pre-Trained Transformer, the same type of tech powering the well-known ChatGPT. The brokerage is calling its bot Leo.

But even as some in the real estate industry quietly fret about the new technology, and its potential threat to humanity, the brokerage’s chief technology officer, Pritesh Damani, told Inman the tool offers few benefits without information exclusive to the company responsible for collecting it.

It’s “extremely useless without proprietary data,” Damani told Inman.

The comments come as the brokerage races to integrate artificial intelligence into its existing tech stack. For now, its using GPT to assist agents in day-to-day responsibilities. Executed from a chat interface, agents can engage Leo with a wide array of questions, many of which would normally materialize as an email to a team lead, broker or admin, likely resulting in a day’s long back-and-forth before an answer is surfaced.

Leo is also up to speed on company events, training seminar schedules and other noteworthy company happenings.

With Leo, agents can inquire about their cap status, when their license expires, the size of their network or a pending revenue forecast. Leo was taught to respond concisely, even a bit curtly. But, the idea is to be quick and accurate. “Super key to this is the clarity of the message,” Damani told Inman.

Because Leo has absorbed each agent’s accumulated data as well as The Real Brokerage’s information, its presence eliminates what would normally be a countless stream of banal conversations flooding the inboxes of company personnel.

“Leo also will be customized for each agent, allowing it to understand and cater to their unique needs and working style,” the company said in May. “This personalized approach ensures that each agent receives the precise support they need, when they need it, making their job easier and more efficient,“

Damani said that without teaching artificial intelligence tools like GPT what your company does and how it functions, its large language modeling won’t benefit an organization. It’s just a giant machine waiting to be told to turn on. “Otherwise, why would it be exciting?,” he said.

Damani also insists that in order for GPT to fully learn a system — and, ultimately, help agents and decision makers — everyone needs to be using the software from the beginning.

Artificial intelligence tools struggle to advance an enterprise that’s divided among multiple CRMs, transaction tools and lead systems. It needs to be fed standardized data. This gives Real an advantage when it comes to GPT, for example, as it’s developed an in-house transaction coordination solution and it partners with a single third-party provider, Chime. It’s tech-stack is short.

Damani wants the tool to eventually send agents marketing templates as soon as they enter a listing, some of which have been designed by company president Sharran Srivatsaa. And these marketing plans will be consistent with the agent’s location and the listing’s pricing tier. Leo will also learn to send task lists and other business suggestions, becoming in many ways a coach.

“This isn’t customer support; it’s business insight,” Damani said. “We expect Leo to eventually answer everything.”

Email Craig Rowe