Valuation sits at a critical point in every loan. AI can make valuation workflows faster with fewer delays. It can also make them fragile. Getting it wrong means problems in compliance, regulatory exposure and biased outputs. The challenge is not capability alone, it’s trust and accountability. That line needs to be walked on carefully. Here is how we are approaching it.
The industry still has untapped opportunities, largely due to regulatory constraints and the current limitations of how AI operates. For example, in 2026, origination volume has picked up again. But the cost to originate remains high, and teams are under pressure to do more with the same resources. In theory, AI could address this. In practice, we are not yet seeing workflow intelligence mature to a point where systems can independently manage and execute such work. Right now we are seeing assistive AI like Chat bots and document sorting. These are reliable and reduce manual effort, but they stop short of driving decisions. The next step for the industry is AI-powered, proactive decision support.
Valuation, in particular, is a highly regulated and accountable discipline. Judgment, rationale, and responsibility for every outcome sit with a qualified professional. The role of AI here is not to replace that expertise, but to remove the interpretive overhead that comes before it. The scanning, the piecing together of information, and the effort that consumes time without adding value.
Adoption has been slower largely due to trust. An AI model can process more data than any person and surface patterns no one would catch manually. What it cannot yet do is fully understand context or explain its reasoning in a way that satisfies regulators or borrowers. Lenders operate under strict regulatory scrutiny. Borrowers deserve to understand why decisions are made about their loans. An AI system that cannot explain itself and operates as a black box becomes a liability. This also ties into the challenge of data overload. If AI produces incorrect outputs, teams need to be able to identify and correct them. This is why human intervention remains essential, particularly for interpretation. In this context, assistive AI plays a meaningful role by reducing redundant effort without overstepping accountability.
When we built AI Summary, by SAM we started from that principle. The feature does not make decisions. It removes the manual effort of scanning and consolidating order history, giving users a clear view of where things stand and what needs to happen next. It is designed to be useful within trusted boundaries.
SAM, our AI assistant, is the umbrella under which all of our AI initiatives will be developed. Every feature we build going forward will be part of what SAM becomes. AI Summary is our starting point. What we are building toward is more agentic AI with capabilities like vendor recommendations, proactive alerts and daily briefing. The goal has always been to make workflows that are simple, smooth, and fast. As the industry evolves, our focus is on building systems where efficiency is matched by accountability.
