As artificial intelligence rapidly reshapes mortgage operations, lenders and servicers are grappling with a central question: where does automation truly add value, and where does human expertise remain essential?
That question was at the heart of “The Human Multiplier: Where AI Stops and Expert Ops Begin,” a recent Five Star Webinar sponsored by Global Strategic and featuring Brian Flaherty, CEO of Global Strategic, and Cade Thompson, Co-President and Chief Growth Officer at Rocktop Technologies. Moderated by Five Star Institute and MortgagePoint Editor-in-Chief David Wharton, the discussion cut through AI hype to focus on practical, real-world applications across origination, servicing, and capital markets.
Here are the top five takeaways from the conversation.
- BPO Isn’t a Trend: It’s a Structural Necessity for a Cyclical Industry While offshoring and outsourcing are often framed as cost-saving tactics, panelists emphasized that Business Process Outsourcing’s (BPO) real value lies in its ability to help lenders survive extreme and unpredictable market cycles.
As Thompson explained, outsourcing was never about chasing trends. Rather, it was about scalability and resilience.
“We didn’t get into offshoring because it was trendy. We needed scalability, cost efficiency, and resilience.”
Flaherty reinforced that outsourcing has continuously evolved alongside technology and regulation, rather than being displaced by it. “BPO is not new. It’s gone through many iterations and a lot of evolution, and we’re facing another one right now.”
In short, BPO has become embedded in mortgage operations because it absorbs volatility in a way that fixed, onshore staffing models often cannot keep up with.
- AI’s Biggest Wins Today: Speed, Consistency, and Pattern Recognition
Despite the excitement around generative AI, Thompson explained that today’s most reliable AI use cases are not creative or autonomous. Rather, they are operational.
“AI today is outperforming humans on things like doc classification, data extraction, consistency in mortgage workflows,” Thompson said.
Rather than replacing decision-makers, AI excels at front-loading analysis and reducing friction, performing tasks such as quickly analyzing files, reconciling data, comparing against guidelines, and highlighting anomalies.
Thompson added, “It can do that lower-level work to present information to the universe of humans that can do more critical thinking work.” The panelists stated that this shift allows organizations to move faster while improving data quality, both of which will aid lenders facing tight margins.”
- The Myth of “Fix It and Forget It”
Both speakers warned against the idea that AI can simply be deployed and then left unattended. Flaherty put it bluntly, saying, “The misconception is you’re going to fix it and forget it, that … it’s going to solve for everything, and I can just go sit on the beach and drink Mai Tais while all the work gets done. That’s not the reality.”
AI systems require ongoing governance, fine-tuning, and human oversight, especially when operating within heavily regulated environments such as the mortgage industry.
- Human-in-the-Loop Is a Requirement, Not a Weakness
Rather than being a sign of an immature technology being deployed too soon, both panelists argued that human involvement is fundamental for risk management, compliance, and continuous improvement.
Moreover, Thompson pushed back against the idea that automation should be fully autonomous.
“With risk management, regulatory management, and those types of things, humans are how the models get better,” Thompson said. “It’s [about] finding the edge cases [and] feeding the learning back into it. You’ve got to have that human critical thinking in there to make the tech better.”
Flaherty noted that regulators already require human review in many adjacent industries, and mortgage will be no exception.
“You can automate this, but at this stage, I still want a human.”
- Finding the ‘One Plus One Equals Three’ Model
Rather than viewing AI and BPO as competing forces, both speakers emphasized the combined potential of these two areas. Flaherty described how AI can enhance, rather than replace, outsourced operations.
“We’re using AI to streamline procedure mapping,” he said, to provide “line-of-sight oversight into our own [internal processes].”
Thompson framed the future as presenting an elevation of human roles, rather than their elimination.
“AI today compresses the lower-skill tiers of work. It should shrink the amount of critical tasks that are available to be done by the human element,”
he explained.
The result, in theory, is fewer people doing basic, repetitive tasks, leaving more team members free to focus on exceptions, quality assurance, and higher-order decision-making.
“If we can elevate people to a level where they can get value out of their jobs … by reducing the mundane aspect of it … it’s a win.”
While neither AI nor BPO alone will define the future of mortgage operations, success will likely come to those who masterfully combine automation, expertise, and human judgment while remaining flexible enough to adapt to ongoing changes in both technology and regulation.
As Flaherty concluded, the goal isn’t to “be slaves to the tech,” but rather “to be masters of the tech to make our lives better.”
