Underwriters Who Use AI Will Replace Those Who Don’t

Introduction

At the NAGGL 2025 Annual Conference in Colorado Springs, one of the sessions on the main stage opened with a quick audience poll: How many attendees use artificial intelligence in their personal lives? Nearly everyone raised their hands. And how many use it at work? Less than half. When asked “who uses AI for credit underwriting?”, less than a handful of people raised were brave enough to say “I do!”. While the use of AI in lending is still in its early stages for some bankers, others see it completely differently. Some are already benefiting from major time savings and strengthening their position inside their organization, most are still ironically afraid AI will replace them.

Omri Yacubovich, CEO of Lama AI, was invited to speak at the conference’s main stage. Rather confirming existing concerns, there is another perspective to this story: “AI will not replace underwriters - but underwriters who use AI will replace those who don’t.” The discussion at Colorado Springs moved beyond technology, into themes of trust, accountability, and innovation. Here are some key insights that emerged from the session:

1) AI isn’t here to replace us, but to empower us.

More than anything, AI isn’t here to replace us, but to empower us. AI should help us make better, faster, and more precise decisions, and automate technical, time-consuming, and lower-value tasks. By taking over repetitive and time consuming tasks, AI enables teams to dedicate their expertise where it counts most - evaluating risk, exercising judgment, and driving smarter lending outcomes. AI doesn’t remove the need for expertise, it helps teams make decisions that are faster, more balanced, and more consistent.

“AI will not replace underwriters - but underwriters who use AI will replace those who don’t.”

2) Don’t use AI for the sake of using AI.

Success starts with clearly defining the business problem, the success metrics, and the measurable value it’s meant to deliver, not by saying “we want to automate”.  Our recommendation is to start where ROI is highest and operational risk is lowest. Focus on areas such as document processing, initial credit screening, narrative generation, and compliance validation. These use cases deliver immediate, tangible impact without disrupting core lending operations.

3) Keep humans in the loop, always.

AI supports decisions, but it doesn’t make them. When humans stay in the loop, technology becomes a force multiplier for transparency, consistency, and better decision making. While technology and AI can provide powerful insights, automate data collection, and highlight potential risks or opportunities, the final decision should always be made by a human. It’s the lender’s expertise, context awareness, and ethical judgment that ensure each credit decision aligns with regulatory standards, institutional policy, and the borrower’s real-world circumstances.

Summary

The race has already begun. Those still debating whether to go AI-first, cautiously “add AI,” or simply wait and see will soon find themselves outpaced by those already moving. The choice is simple: do you want to be on the side driving change, or on the side trying to catch up?

So, which side are you on?

‍