David Lykken sits down with TRUE CEO Steve Butler to discuss why mortgage AI governance, data traceability and trust are becoming critical requirements for lenders.
AI has moved quickly from experimentation into real mortgage workflows. It’s being used to extract data, interpret documents, support underwriting and accelerate decisions across the loan lifecycle.
If someone challenges an AI-assisted decision months or years from now, can you reconstruct exactly what happened?
In this conversation, David Lykken sits down with Steve Butler to examine what Fannie Mae’s AI governance guidance means for lenders and why traditional approaches to technology governance may not be enough for GenAI.
The discussion includes the fundamental difference between probabilistic AI outputs and deterministic, repeatable results; why data lineage and validation are becoming essential to enterprise AI; and why lenders should be evaluating the governance architecture behind their AI vendors, not simply the accuracy of their models.
Key topics includes:
- Why accuracy alone isn’t enough. A highly accurate model doesn’t necessarily give a lender the evidence needed to defend an output later.
- The growing importance of traceability. Can you trace a value back to its source document, understand how it was validated and reconstruct what ultimately entered the system of record?
- Why GenAI changes the governance equation. The same prompt can potentially produce different results over time, making preservation of evidence and validation increasingly important.
- What lenders should be asking AI vendors now. Governance needs to become part of technology evaluation—not something addressed after deployment.
- Why governance and efficiency don’t have to compete. Steve explains how creating trusted data earlier in the process can both strengthen governance and reduce the repeated human review that adds cost throughout the loan lifecycle.
Take the Next Step: Assess Your AI Governance Readiness
The interview raises the questions. Our Executive Guide to Preserving Data Traceability in AI-Powered Mortgage Lending helps your organization start answering them.
Built around Fannie Mae’s AI governance guidance, the guide provides a practical framework for evaluating the traceability and governance behind the AI systems your organization is already using or considering.
Inside, you’ll find five questions to ask AI vendors, an Mortgage AI Governance Readiness Checklist, guidance for evaluating data traceability, and practical considerations for preparing your organization for LL-2026-04.
The goal isn’t simply to check a compliance box. It’s to help your risk, compliance, technology and operations leaders answer a much more important question:
