X

Black Entrepreneur Engineers First-Ever Conversational AI System That Transforms Mortgage Lending

Nationwide — Troy Smith, Founder and CEO of The G1 Group, has engineered Carol iE3, the first and only artificial intelligence system in the NonQM mortgage industry that handles complete loan origination through natural conversation. The groundbreaking platform, accessible at grow.glwins.com, represents a technological milestone that positions independent mortgage professionals to compete on equal footing with the nation’s largest lending institutions.

Smith, who engineered the system from concept to deployment, built Carol to democratize access to enterprise-level mortgage technology. In an industry where advanced automation has historically required million-dollar investments, Smith’s innovation delivers institutional-grade capabilities through a conversational interface accessible to any broker with internet access. The platform directly addresses systemic barriers that have limited Black and minority-owned brokerages from competing against well-capitalized firms, creating pathways to wealth generation through homeownership and entrepreneurship.

As both engineer and CEO, Smith developed Carol specifically to address workflow bottlenecks that disproportionately impact smaller brokerages and independent loan officers. The system processes pricing scenarios, analyzes bank statements, manages loan pipelines, and handles submissions without requiring users to master complex software platforms or navigate multiple screens.

“I engineered Carol because technology should expand opportunity, not create barriers,” Smith said. “As the first conversational AI in NonQM lending, Carol gives independent brokers, including many minority-owned businesses, access to tools that match what billion-dollar institutions have. This competitive advantage translates directly into closed deals and wealth creation for professionals who have been locked out of expensive enterprise technology.”

The platform analyzes loan scenarios in seconds, processing credit profiles, property details, and program requirements to deliver rate quotes across NonQM products. Loan officers describe scenarios in plain language and receive formatted rate tables with payment calculations and program recommendations. The technology handles complex NonQM scenarios, including bank statement income calculations, asset depletion qualifications, and investor property evaluations.

Carol processes bank statements in under 30 seconds, analyzing 12 to 24 months of financial documentation to extract deposit patterns, identify income sources, and flag potential documentation issues before underwriting. This front-end validation enables loan officers to address client questions about qualification during initial consultations rather than discovering problems after application.

The system maintains context across interactions, allowing loan officers to reference previous conversations when making decisions about specific loans. Users can check pipeline status, retrieve condition lists, update loan phases, and verify submission readiness through conversational queries. The platform provides access to 697 guideline sections from four policy manuals, enabling real-time verification of program requirements during client conversations.

User management operates through the same conversational interface. Branch managers can onboard new loan officers, update company information, and modify access permissions without accessing separate administrative systems. The platform generates credentials and sends notifications automatically, streamlining team expansion for growing brokerages.

Integration occurs through embedded widgets that brokers can add to existing websites or use within The G1 Group’s platform. The system authenticates through API keys while maintaining security protocols, allowing mortgage professionals to deploy the technology across different digital properties. Session data persists across interactions, eliminating the need to repeat context when discussing specific loans.

The launch establishes Smith and The G1 Group as pioneers in mortgage technology innovation. As the first and only wholesale lender to integrate conversational artificial intelligence directly into NonQM origination workflows, Smith’s engineering breakthrough represents a fundamentally different application of technology. While mortgage companies have deployed chatbots for customer service, Carol handles operational functions that directly impact loan processing timelines and officer productivity.

Smith’s innovation creates pathways for mortgage professionals seeking to scale operations without proportional increases in overhead. The platform handles higher loan volumes while maintaining personalized client service, with technology managing routine operational tasks so loan officers can focus on relationship building and business development. This approach particularly benefits minority-owned brokerages that have historically faced challenges accessing capital for expensive technology infrastructure.

For more details, visit grow.glwins.com

The G1 Group operates as a division of Oaktree Funding Corp., based in Benedict, Maryland. Under Smith’s leadership, the company specializes in NonQM, Small Balance Commercial, and Residential Transition Loans for mortgage brokers and independent loan officers nationwide, with a strategic focus on technology investments that level the playing field for independent professionals.

About The G1 Group
The G1 Group, a division of Oaktree Funding Corp., 3133 West Frye Road, Suite 205, Chandler, AZ 85226 (NMLS # 71640), provides wholesale mortgage lending services from offices located at 7252 Benedict Ave., Benedict, MD 20612. The company focuses on NonQM and specialty loan products designed for the independent mortgage broker channel, with continued investment in technology that creates competitive advantages for broker partners.

X

Headline

You can control the ways in which we improve and personalize your experience. Please choose whether you wish to allow the following:

Privacy Settings