Dark Matter Technologies announced on October 30 that it has acquired the AI agent platform and related intellectual property of House Numbers, a venture-backed company that builds AI agents for home lending workflows. Alongside the technology, Dark Matter has brought on the team that built it, led by House Numbers co-founder and chief executive Jeff Levinsohn, who joins as chief AI officer. The terms of the transaction were not disclosed.

The deal is a buy rather than a build. Dark Matter chief executive Vikas Rao told HousingWire that the two companies first discussed a partnership before he proposed an acquisition, so the technology could be integrated deeply into Dark Matter’s products instead of offered as a separate service. In the announcement, Rao said bringing in proven technology and the team behind it gets agentic AI into clients’ hands faster than building from scratch, with the payoff in borrower experience, loan officer productivity and a lower cost of origination.

What Dark Matter bought#

Levinsohn founded House Numbers in 2021 with Benjamin Hoffman and Arthur Granado, and led the platform from concept into live production for home lenders. Its autonomous agents constantly monitor loan files for changes and respond by working loan conditions, taking actions, driving recommendations and communicating with borrowers and lenders by email, text and voice, with loan operations personnel supervising.

That production history is the point of the purchase. The platform has been live for two years and has processed loans for more than 100 lenders, according to Dark Matter, and Rao told HousingWire it has 19 agentic AI use cases in production, with more in development. Levinsohn put the hard part plainly: orchestrating complex agentic work safely inside a regulated process and executing valuable tasks consistently at scale, “with people in control of the decisions that matter.”

Loan folders moving through review stages toward a human reviewer
AI-generated editorial illustration.

How the agents are kept inside the lines#

House Numbers built the platform for a regulated industry, and the controls read like the pitch. A workflow orchestration engine is paired with a Model Context Protocol (MCP) gateway, so each agent gets only the context and capabilities its task requires, and external AI assistants can work a loan file with the same permissions and controls as a human user. Confidence gating, a code-enforced compliance blocklist and human oversight keep agents within defined limits, and a complete audit trail records every action taken by a person or an agent. Per-workflow model tiering, multi-provider failover and per-loan cost attribution are meant to keep the system running efficiently at scale.

In day-to-day use, the agents help borrowers complete applications, understand outstanding requirements and get feedback when they submit incorrect or incomplete documents. They also review incoming documents to determine whether loan conditions have been satisfied, check closing packages for missing signatures or pages, send reminders and flag potential issues for loan officers.

House keys, signed papers and a pen on a closing table
AI-generated editorial illustration.

The cost argument#

Rao framed the acquisition in dollars. He estimated that the total cost of producing a loan remains around $12,000, with roughly $6,000 attributed to operational expenses, and said: “Even shaving $1,000 off of that $6,000 is game changing.” He added that the technology is not about removing people from the process: “We’re not talking about eliminating humans,” Rao said. “What we’re talking about is these AI agents are performing these tasks instantaneously. But then all of this is being reviewed by a human.” Lenders will decide when agents can act independently, make recommendations or flag items for human review, and will be able to train agents to follow their own policies and procedures.

Dark Matter plans to run the acquired technology within its Amazon Web Services infrastructure, with controls designed to address security and regulatory requirements. The acquisition supports three AI platform commitments the company lists in the announcement: AI embedded in every workflow, an AI agent assistant for every user, and the shift from doing to reviewing across the loan lifecycle.

What to watch#

Dark Matter expects customers to begin seeing new capabilities within weeks of the announcement, with additional releases planned over the following months — a fast clock for technology that has to work inside a regulated process at the scale of Dark Matter’s largest lenders. Watch whether the $1,000 saving Rao sketched shows up as a measurable drop in origination cost, and whether the 19 production use cases grow once the team is working across Empower and the wider product line. Rao also said Dark Matter hopes to pursue more deals in 2027, with parent company Constellation Software encouraging portfolio companies to consider complementary acquisitions. This one sets the template: proven agents, bought with their builders, and wired directly into the workflow.

Sources: Dark Matter Technologies’ announcement, “Dark Matter acquires House Numbers’ agentic AI platform for home lending, names Jeff Levinsohn chief AI officer”, and HousingWire’s report, “Dark Matter acquires House Numbers’ agentic AI platform”. Images are AI-generated editorial illustrations.