Brian Abent spent 12 years at Ceros running product and engineering. $0 to $50M ARR - PE exit. Final year he transformed the operations team around AI at the core. After leaving, he sold his house and discovered that an industry processing trillions in annual transactions couldn't give a buyer a working first impression. The agent's best tool was a yard sign with a text code that responds with a scammy message. He ripped the sign out of the yard and started building.
The MLS reduces every home to a spreadsheet row - aluxury renovation with solar panels looks identical to a 1970s flip. Buyers can't search for what actually matters to them. Sellers spend years perfecting a home, then hand the first impression to a yard sign with a contact form nobody checks. Agents field unqualified leads all day. The infrastructure is 30 years old and it shows.
Brian built Alma as a seller who wanted the buyer to experience his home the way a friend would - not through a broken yard sign. That instinct became a founding principle: put the buyer first. Each listing gets a dedicated website with a 24/7 AI that knows the home inside and out - the light, the character, the details that make someone fall in love with it.
Real estate software is built entirely for high-volume agents. It ignores the humans on either side of every transaction - the buyer and the seller. Alma doesn't. Every buyer is also a seller - either selling another home right now, or selling this one in the future. It's the same person. Alma is built for the humans inside the houses, across the entire arc of homeownership.
Not a portal. Zillow and Redfin are listing aggregators - search engines for real estate. Alma is the home itself, talking. It's the engagement layer that lives on every listing, every sign, every agent's site. Zillow shows you what's available. Alma helps you figure out what you actually want.
A GTM lead joins at close - someone who's already been working on Alma for six months, unpaid. Outbound to top brokers, connecting with Redfin's partner team, coordinating with agents, evangelizing the product. Not a hire we need to find. A hire we need to pay.
The unit economics are simple: 1,000 agents at $999/yr = $1M ARR. No marketplace dynamics to bootstrap. No network effects required to deliver value on day one.
When a buyer asks about natural light in three different listings, asks about noise levels, asks about outdoor space - Alma is learning how they actually live. Not filters. Real intent. Across enough listings in a market, Alma builds a per-buyer preference model and starts matching proactively. That's when buyers come to Alma before Zillow. The agent tool gets us there. The taste graph is what keeps us there.
At close, Alma transfers to the buyer and keeps growing. Every renovation documented for resale value. Every system tracked for maintenance. Insurance-ready records from day one. When it's time to sell again, the listing writes itself - one click, years of context. Today's buyer is tomorrow's seller. Alma is already there.
Per-buyer preference data from millions of real conversations is irreplicable - Zillow has search behavior, Alma has conversations. And buyer trust: Alma is the only platform in the stack that is structurally on the buyer's side. That trust, at scale, is the real asset.
SAFEs. 18–24 months to prove the model in Orlando, then replicate nationally.