A five-person team built an artificial intelligence platform used by 30,000 real estate agents across 240-plus brokerages in 16 months, all without outside capital.
Now BrokerBot, a Tucson, Arizona-based company operated by Ribera AI, Inc., has announced on Friday that it closed a seed funding round led by Grand Ventures, with participation from Second Century Ventures, the strategic investment arm of the National Association of Realtors. The company also joins the 2026 NAR REACH cohort, SCV’s flagship scale-up program for real estate tech.
Financial terms of the round were not disclosed.
Jerimiah Taylor
BrokerBot trains AI assistants exclusively on a brokerage’s own documents, policies, and forms, not generic internet data, and deploys them across chat, SMS, voice, web and email. Brokerages report the system resolves roughly 83 percent of real estate agent questions automatically, saves more than 33 staff hours per brokerage per month, and has driven a 70 percent reduction in minor contract corrections.
”We’ve spent the last 20 years operating real estate brokerages and the last year building the AI teammate we always wished we had, one that knows our brokerage inside and out, never sleeps, and earns trust by being compliance-first, not hype-first,” Jerimiah Taylor, BrokerBot co-founder and CEO, said in a statement. “With Grand Ventures leading our seed, Second Century Ventures investing alongside, and NAR REACH supporting us, we’re moving from answering questions to actually getting work done across the brokerage stack.”
‘They come back 6 months later’
The company’s pitch to brokerages is compliance-first AI, a direct counter to the “just build it with Claude Code” impulse, which Taylor says he constantly runs into among brokers trying to cut costs.
“What I run into is, ‘Oh, I could just build this with Claude Code,'” Taylor told Inman. “And then six months later, they come back and say, ‘Well, that didn’t work out. Can I buy your thing now?'”
Taylor said the average broker underestimates the liability exposure created by running consumer-grade AI tools, models that aren’t enterprise-contracted, don’t have data retention agreements, and can surface client information in unpredictable ways.
The high compliance requirements for AI in real estate, Taylor said, force brokers into one of three decisions: prohibit agents from using AI, which is very unlikely; require agents to use enterprise-grade AI and place the onus on them; or provide a brokerage-wide solution themselves.
“In all those scenarios, we’re in a really good spot because we’ve made a significant investment in providing a truly secure platform,” he said. “It sets the bar high for anyone trying to chase us.”
Arizona, where BrokerBot is headquartered, gives the company a structural runway to test that message. The state’s Attorney General’s Office operates a “regulatory sandbox” that allows companies working in regulated, licensed industries to operate without full licensing requirements while they develop.
Taylor sees NAR — “possibly the most powerful lobbying arm in the country,” in his framing — as a vehicle to shape AI regulation at the national level before states diverge too widely.
“We’ve taken the time to understand the rules at the national, state, and local levels, and I think NAR can help us have a seat at the table when regulations are created,” Taylor said.
Built lean, by design
The seed round will go primarily toward building out BrokerBot’s sales organization, Taylor said. The engineering headcount stays lean by design. The company’s development team produces roughly half a million lines of code per quarter using AI-assisted tools, and a two-person customer success team currently supports all 30,000 users.
“Building a company in this environment, even just from a year ago, is really different. It changes how you forecast headcount and growth,” Taylor said.
Platform metrics the company disclosed: 291,000-plus unique visitors, more than 500,000 messages exchanged, 22x message growth over 10 months, and 69 percent of activity occurring after business hours.
The numbers point to a pattern, Taylor said, that repeats across every deployment. Managing brokers get relief from the repetitive-question queue, and real estate agents get answers when no one else is available.
Current integrations include Keller Williams Command. BrokerBot was named to KW’s initial third-party developer lineup when the company launched its Command platform in February, alongside Google Calendar, Gmail, Microsoft Outlook, Canva and Follow Up Boss.
Deployments span independents like Baird & Warner, which runs a BrokerBot assistant called “Remi,” as well as franchisees at Keller Williams, Coldwell Banker, Better Homes and Gardens Real Estate, United Real Estate and Engel & Völkers.
What’s next
Seed capital and a REACH seat point toward a clear next phase: agentic workflows.
Rather than answering agent questions, Taylor wants BrokerBot to complete tasks by integrating with transaction management, e-signature platforms, and MLS data to take action on behalf of real estate agents and staff, rather than routing information back to them.
On the product side, the company has also built what Taylor described as a proprietary video-processing layer. The system opens video files in cloud containers, uses AI agents to watch and listen simultaneously, and indexes the output so agents can query spoken content by the second.
The use case: surfacing a specific moment from a training video or recording in response to a natural-language question.
“It’s probably the nerdiest thing we’ve built yet,” Taylor said. “But it was hard, and it’s really slick when it works.”
Tim Streit, co-founder and general partner at Grand Ventures, said BrokerBot was the first platform he’d seen train on a brokerage’s own workflows and clear compliance requirements at scale, and that the 240-plus live deployments provided the verification.
“We’re proud to lead this round and to back the team building what we believe will become the operating layer for the modern brokerage,” Streit said.
