The old playbook is dead: Now redefine non-bank lending with AI

The non-bank lending industry has hit a plateau of sameness. After years of weathering economic volatility, regulatory pressure, housing market swings, and ongoing consolidation, most lenders find themselves mid-way in their AI journey. Among the 257 US non-bank lenders surveyed, only 21% consider themselves innovation leaders, while over half admit to operating at competitive parity—doing just enough to keep up, but falling short of true differentiation.

This sentiment was echoed clearly at the May 2025 digital roundtable, hosted by HFS and Cognizant, which brought together senior leaders from Angel Oak Mortgage Solutions, Carrington Mortgage Holdings, Cenlar FSB, Lakeview Loan Servicing, Onity Group, Primary Residential Mortgage, Cotality, Planet Home Lending, and Guaranteed Rate. The goal wasn’t to hear lofty prognostics but to surface realistic actions. Rather than debating the merits of AI innovation, the focus was on a more pressing question: how should the non-bank mortgage industry prepare for the inevitable transformation it will bring?

The message was clear: the success of AI hinges not just on investment, but on how intelligently and strategically it’s applied. The discussion will strike a chord with the mortgage industry, as it captures the real-world hesitations, challenges, and actionable insights shaping lenders’ AI trajectory.

Compliance drags and vendor platform tethering hinder AI scale

Despite the promise and heightened focus on AI, many non-bank lenders remain ill-equipped to deliver and scale the technology. One of the biggest challenges is the burden of a compliance-heavy environment, where many core processes are still anchored in frameworks built over a decade ago. As a result, deploying AI at scale requires a full-stack rethink over legal, compliance, operations, and technology.

Offering a candid reality check, a delegate remarked, “There’s no shortage of AI ambition in the industry—but scaling is another story.”

The structural challenges run even deeper. The industry is undergoing rapid consolidation—whether for capabilities or to build scale, these moves only add to the complexity of the technology stack and are reshaping the competitive landscape. The industry remains highly fragmented. Within an organization, there are often multiple legacy platforms, rigid data architectures, and vendor ecosystems. Furthermore, many lenders remain tethered to LOS giants whose innovation cycles are slow and inflexible. Even if a lender is ready to move fast, their progress is only as good as their slowest platform partner. Some of the most promising solutions may come from agile, AI-native vendors, but only if they meet the enterprise-grade standards for security, compliance, and scalability.

Hesitancy still lingers across the industry, which many leaders are not proactively addressing. While the challenges are real and complex, leaders must be prepared to act. Accelerating AI-driven transformation requires a deliberate and holistic approach: defining a clear architectural vision of using AI versus humans; building strong data foundations; and redesigning workflows with legal, compliance, technology, and operations all at the table. Equally important is ensuring financial viability to support and sustain this transition.

Winning the AI case means proving ROI against cost to originate and service

Non-bank lending leaders made it clear that the economics of mortgage origination and servicing are increasingly unforgiving. Most operate in a hyper-competitive market where the largest cost—people—can’t scale efficiently, pricing power is limited, and borrower loyalty is fleeting. Originators face constant margin pressure: commissions eat into profits, excessive fees drive churn, and rising cost-to-serve erodes the value of every deal. Worse still, when a loan doesn’t close, the sunk cost is real and painful.

Today’s underwriting landscape only adds to the complexity. It’s no longer about greenlighting straightforward, cookie-cutter loans. Borrowers are more complex, and products more specialized. That’s why manual underwriting still matters—it allows lenders to serve edge cases, uphold loan quality, and deliver personalized value where automation can’t. However, routine decisions—such as standard credit checks or simple income verifications—can be automated with AI, allowing underwriters to focus on those complex, high-value cases.

Here’s the reality: some non-bank lenders believe AI isn’t bridging that gap, at least not yet. While the market is flooded with AI tools and vendor promises, most solutions fail to move the needle on the core economics of the business. If non-bank lenders keep waiting for AI to drive down the cost to originate and service, they’ll be unprepared to meet the demands of tomorrow’s market.

All is not bad news—real success story emerges

Amid the hesitation and challenges facing non-bank mortgage lenders, a standout success story is unfolding: AI agents are moving from hype to meaningful impact. A key insight from the HFS-Cognizant roundtable was the growing recognition that AI agents are no longer just peripheral tools but central actors in a reimagined operational model. Leading lenders are piloting custom-built AI agents assigned to underwriters, processors, and even HR teams. These agents are more than generic chatbots—they’re intelligent, role-specific copilots that understand institutional data, workflows, and business goals from day one.

As a delegate from a firm already scaling with AI agents put it, “Our goal is for everyone to have a co-pilot—an AI agent that understands their role, learns with them, and helps them work smarter from day one.”

The implications are profound. As described by the delegates, picture a loan not as a transaction to be manually moved through siloed systems but as a product on a digital assembly line, with AI agents orchestrating tasks, enforcing quality, adapting workflows, and flagging exceptions. Each loan moves through a sequence where document ingestion, data validation, compliance checks, and customer interactions are intelligently automated—with humans only stepping in where judgment truly adds value. This is not automation for automation’s sake—it’s AI-led orchestration that delivers scale, accuracy, and speed.

Communication and collaboration are essential to scaling AI

Leaders at the virtual event emphasized that cross-functional collaboration is key to scaling AI and reimagining operations. Progress happens when CIOs, business leaders, legal, compliance, and operations align early—whether through shared roadmaps or regular advisory syncs to prevent fragmented efforts. Successful organizations foster radical transparency and shared ownership, intentionally bringing skeptics into the room alongside AI champions to critically shape direction.

As one participant summarized, “You can’t deploy AI in silos and expect system-wide impact. The future belongs to the organizations that communicate like they build openly, collaboratively, and with purpose.”

The Bottom Line: Winning with AI-led innovation for non-bank lenders means overcoming institutional complexities and scaling with purpose to unlock true competitive advantage.

The industry feels divided. Yes, there are real success stories of AI agents at scale, but the hesitation still lingers. Many remain entangled in institutional complexities, waiting for a watertight business case or a definitive ROI before making a move. But non-bank lenders can’t afford to wait for perfect conditions to embrace AI if they want to streamline operations and secure long-term growth.

The better path is to start now. Collaborate and communicate how you envision and execute your AI strategy. Those that build the organizational muscle to scale AI with purpose will be ready to meet tomorrow’s mortgage demands head-on.