Colin Terry

Chief Product Officer

LoanPro

Roadblock Title:

Trust Was Always a Bad Idea: Lessons from Building AI in Lending

Time:

Tuesday - 1:00 PM (Windows)

Abstract:

AI in lending isn’t going away, but at LoanPro we learned that it can’t be trusted blindly. That tension is the central challenge for any fintech operator trying to move fast without breaking compliance.

Colin will walk through the story of the different options we considered and the different AI tools clients were asking for, and why so many of them either created massive legal liabilities or manual reviews that negated the efficiency gains we were looking for in the first place. We eventually realized that in a heavily regulated industry like lending and credit, you need a built in review layer that relies on neither AI nor manual review.

For LoanPro, the solution we landed on was Model Context Protocol (MCP) architecture. Colin will walk through what that solution does for us, and when it might—or might not—be right for you.

Learning objectives:

Understand the compliance and audit requirements that make “just trust the model” a non-starter in lending and other highly regulated industries.
Walk through what we tried first, from prompt engineering to manual review cycles, and where each one broke down once it got past the idea stage.
Based on our own team’s engineering journey, evaluate why we arrived an an MCP architecture as both a tool to streamline AI implementation and automatically review outputs and agentic behavior.

Bio:

Colin Terry is a fintech operator and product leader with a decade of experience modernizing how lenders build, service, and scale credit. As Chief Product Officer and COO at LoanPro, Colin has developed product strategy from the ground up, most recently leading the launch of LoanPro MCP, an industry-first AI gateway enabling financial organizations to deploy compliant, agentic loan servicing at scale. He brings deep operational and product expertise to one of fintech’s most critical infrastructure challenges: making AI-powered lending both innovative and defensible.