AINEWS
Real Brokerage says its in house AI logged 1,992 conversations with cold leads in 30 days. The customer who fed the leads says the same system replaced his inside sales staff. Both claims are now on the same record.
The Real Brokerage ran an experiment: take roughly 3,000 leads that had gone cold, drop them into an AI assistant called Leo 2.0, and see how many conversations the system could run without a human in the loop. In 30 days, the company says, the AI logged 1,992 conversations. At the moment of the interview, 444 of them were live.
That figure, reported to HousingWire. The Real Brokerage's marketing line for the bet is "replacing the work, not the agent." The same customer's on-the-record account puts a number on the contradiction: the AI, he says, replaced his entire inside-sales-agent (ISA) staff.
ISA is the industry's term for the agents, often unlicensed, who sit in front of a brokerage's CRM and work a lead list: calling, texting, qualifying, and handing the warm ones to a licensed agent to close. It is the first rung on the residential brokerage ladder, and it is the rung Leo 2.0 is being deployed to replace.
Keener told HousingWire he uploaded the 3,000 unresponsive leads into Leo 2.0, an AI assistant embedded in the company's reZEN platform, the proprietary software stack Real Brokerage uses to run agent back-office work. The product's "AI Relationship Manager" reads client interactions and sentiment, then flags the ones ready to talk to a human. A second product, HeyLeo, gives consumers a branded property-search tool the brokerage can put on its own site. Together they form the public face of the rollout, which is in beta with "thousands of agents," per the company.
Real Brokerage's Damani. "I don't think this industry has technology right," he told the outlet. The bet is that the same AI stack can run lead follow-up, marketing, and back-office work, all under one roof. Damani uses a long-standing industry rule of thumb, that 1% to 2% of any agent's database is buying or selling at any given time, to argue that the rest of the database is the high-leverage target.
The numbers are Real Brokerage's and Keener's; there is no independent measurement. Keener is one customer, not a survey. HousingWire's own reporting on brokerage AI adoption is that holdouts are thinning out, and Morgan Stanley's 2025 outlook on AI in real estate treats automation of brokerage back-office and lead work as a near-term cost story rather than a remote one. None of that resolves what Keener's account now puts on the record: that "replacing the work" can mean replacing other agents' jobs.
The "replacing the work, not the agent" slogan names the work in the abstract. The work it is replacing is the work other agents used to do: the ISAs, the junior agents pulling lead-qual duty, the inside-sales teams that sit between a CRM and a closing table. Keener's account is the first on-the-record case of a brokerage customer putting a headcount number on what the slogan means in practice. Whether the licensed-agent model survives once that floor is taken out from under it is the next test, and it is the test Real Brokerage's own slogan will be measured against.
Real Brokerage is one brokerage; Keener is one customer. The next test for the rest of the industry is whether the slogan stays abstract or gets the same kind of headcount number attached to it.