Embracing the Future of Mortgage with AI: Key Takeaways from July’s California MBA MQAC Webinar

Jeff Flory

By: Jeffrey Flory, CEO of QC Ally & CMBA’s Chair, Mortgage Quality and Compliance Committee

During our recent California MBA Mortgage Quality and Compliance (MQAC) Committee webinar, I had the honor of moderating a powerhouse panel of leaders pushing the boundaries of mortgage technology. We dug into what’s real, what’s hype, and what’s next when it comes to artificial intelligence in mortgage lending and compliance. And if there’s one thing we all agreed on, it’s this: AI isn’t coming. It’s already here. And how we choose to engage with it today will define the industry’s future.

Here are the biggest takeaways from our conversation.

  1. The Current State of AI: We’re Further Along Than You Think
    • Shelley Leonard, President of Xactus, said it best: we’ve been using AI longer than we realize. From machine learning in credit scoring to AI-assisted customer service chatbots, automation is already embedded in our processes. But CEO and co-founder of Two Dots, Henson Orser, emphasized that we’ve moved beyond the days of 97% accuracy being “good enough.” Today’s best-in-class AI is achieving 100% data extraction accuracy without humans in the loop. What was once theory has become reality, and the tech exists right now to automate entire underwriting processes end-to-end, powered by smart document understanding and orchestration tools.
  2. AI Brings Real Opportunity If You’re Willing to Start Small
    • The biggest myth? That you need a pristine tech stack and perfectly formatted data to get started. Andrew DeGood, CEO and Co-Founder of Ask Bob AI, made a strong case that this is holding lenders back. You don’t need to overhaul your entire operation to see value. Start with one process. See the ROI. Build from there.
    • And Chief Risk Officer and Head of Strategic Partnerships at Gateless, Katie King, echoed that sentiment: AI doesn’t have to eliminate jobs. Instead, it can elevate your teams. Think of it as augmenting your people, freeing them up for judgment-based decisions and complex scenarios.
  3. Compliance and Governance Are Non-Negotiable
    • With all this potential comes responsibility.
    • Senior Partner of Garris Horn LLP, John Levonick, laid out a crucial lens: Every AI function must be assessed through the relationship between tech and human responsibility. Especially in a heavily regulated space like ours. When AI takes on tasks that used to require human licensure or specific regulatory scrutiny, lenders must understand and document those shifts.
    • It’s not enough to say your tech is compliant. You need clear frameworks and governance to demonstrate how it’s compliant. That includes explainability, auditability, and clarity on who does what, when, and why. As John said, governance is the connective tissue between innovation and compliance.
  4. Adoption is Stalled by Fear, Not Facts
    • There’s a natural human instinct to fear what we don’t understand especially when it threatens our norms. But we can’t afford to sit this one out.
    • As Shelley said, the biggest risk isn’t AI. It’s resisting change. Our industry has a track record of being slow to evolve. But this time, the disruption won’t wait. If we don’t lead the change, it’ll be led by outsiders who don’t understand our business or our borrowers.
    • Andrew added another key point: we accept 85% accuracy from humans, but we expect 100% from AI. That’s a double standard. AI is an evolution, not a replacement. It won’t be perfect, but it’s already outperforming many manual processes, and it’s only getting better.
  5. The LOS is the Elephant in the Room
    • We can’t talk about AI without talking about our tech infrastructure. The LOS (loan origination system) has become a system of record, deeply embedded in our workflows. But as Henson noted, these legacy platforms are hard to rip out, and AI’s full potential can’t be realized until we rethink how these systems integrate and interoperate.
    • John was blunt: current LOS vendors are capturing value at every point of integration. For AI to truly drive cost savings and speed, we need lenders to reclaim their tech destiny, knitting together the best tools for the job instead of being locked into monolithic systems.
  6. Practical Next Steps for Lenders
    • The message was clear: Just start. Don’t wait until your data is perfect. Don’t wait until you’ve mapped out every future scenario. Pick a use case like document classification, income verification, or fraud detection, and pilot it.
    • And bring your risk, compliance, and legal teams to the table from the start. As John put it, they shouldn’t be handed a fully baked AI tool after the fact. They need to shape the strategy with you. Katie encouraged us to “trust, but verify.” Start with a pilot. Watch the ROI. Once you see the system working, you’ll know when it’s time to go all-in.

The Bottom Line

AI is no longer optional in the mortgage space. It’s the next wave of transformation, and we have all the ingredients to make it work: real tech, industry-wide interest, and a growing body of best practices.

But the biggest risk? Is doing nothing.

As Shelley warned, if we keep our heads in the sand, someone else will build the future for us, and it may not be a future we like. So now is the time to act with purpose, with partners you trust, and with your eyes wide open. Let’s get to work!even prouder of the way our community continues to show up, ask the hard questions, and support each other.

Contact Us to Learn More