Sela Secures $21M as AI Mortgage Agents Drive $1B+ Monthly Loan Originations

Sela, an AI mortgage agent platform, has secured $21 million in funding while demonstrating that its technology can significantly outperform traditional human-led sales teams at scale. For fintech professionals, this marks a pivotal shift in mortgage origination, proving that voice AI can drive higher profitability and scale more effectively than legacy call center infrastructures in high-stakes lending environments.

What was announced

Sela has officially confirmed the closing of $21 million in total funding, a figure that encompasses both its Seed and Series A rounds. This capital injection was led by Costanoa, providing the financial backing necessary to scale a platform that is already facilitating more than $50 million in mortgage originations on a daily basis. The company’s rapid ascent in the fintech space is further evidenced by its financial performance, reaching an annualized run rate exceeding $10 million within just 18 months of beginning its operations.

The core of the announcement centers on performance metrics derived from large-scale implementations that challenge the traditional reliance on human sales forces for mortgage lead conversion. In a recent performance evaluation involving a substantial sample size of over 10,000 individual borrowers, Sela’s AI technology delivered 40% higher profit compared to the lender's existing call center infrastructure. This suggests that the AI agents are not merely a cost-saving tool but are actively more effective at driving revenue through the mortgage funnel.

Furthermore, Sela provided data from a separate head-to-head trial involving more than 7,000 leads. In this instance, the platform demonstrated 41% better results than a competing voice AI platform, highlighting a significant performance gap between specialized mortgage-focused AI and more general voice technology solutions. The platform’s utility is currently being leveraged by five of the ten largest independent mortgage banks in the United States, indicating a rapid adoption rate among the industry's most significant players.

The companies involved

Sela is a specialized fintech firm that has developed an AI-driven platform designed specifically to serve as an automated mortgage agent. Unlike general-purpose customer service bots, Sela’s technology is engineered to handle the complex, multi-stage interactions required for mortgage origination, a process often characterized by high friction and significant regulatory requirements. The company has focused its market strategy on independent mortgage banks (IMBs), which are non-depository institutions that play a dominant role in the American residential lending market. By securing five of the ten largest IMBs as clients, Sela has established itself as a major infrastructure provider for high-volume lenders.

Costanoa, the venture capital firm that led the $21 million funding rounds, typically focuses on early-stage investments in enterprise software and technology startups. Their decision to lead both the Seed and Series A rounds reflects a strong conviction in the platform's ability to disrupt traditional human-centric sales models within the financial services sector. Sela remains an independent company and has reached its current market position without the backing of a larger corporate parent or a history of previous names. Its rapid growth to a $10 million annualized run rate within 18 months marks it as one of the faster-scaling specialized AI firms in the current fintech landscape.

What this means

The mortgage industry has long struggled with the high overhead costs associated with human-led lead conversion, making Sela’s reported 40% profit increase a direct challenge to the viability of traditional call center models. This announcement suggests that voice AI has matured beyond simple triage and is now capable of managing the nuanced, high-value sales cycles required for home loans. For the broader fintech sector, the pressure is now mounting on legacy lenders to integrate similar automation or face a significant disadvantage in operational margins. The central question for the industry is whether human agents will eventually be reserved only for the most complex edge cases, as the standard for efficient origination is being fundamentally rewritten by specialized AI.