August 17, 2026
Part 4 - So What’s Missing From Your AI Strategy?
By Suha Beidas Zehl, CMB®, AMSP™

In this ongoing series, I’ve been sharing what I believe is missing from most AI strategies in our industry. I’ve published three articles so far: in the first article, I argued that what is missing from most AI strategies is an AI-ready culture, built on three pillars. The second made the case for staying human-centered, and landed on this: keeping humans at the center only works if those humans keep getting better. The third took that further, arguing that talent development is an ongoing discipline, not a training session you check off.
Which leaves the question I ended on last time. Developing your people continuously is the right instinct, but developing them in what?
And that takes us to the third pillar: the deliberate development of how our people think and lead.
We talk about critical thinking as though it were a personality trait, something people either have or do not. Grant and McCann, writing in Harvard Business Review, make the opposite case. They describe four “types” of thinking leaders need to practice and teach: expert thinking, critical thinking, strategic thinking, and systems thinking. Their most important point is easy to miss.
Read that again, because it changes who is responsible. If these skills have to be taught, who is responsible when a workforce cannot think critically about AI output? Leadership.
Each of the four types of thinking maps cleanly onto AI-era lending.
Expert thinking is the seasoned underwriter’s pattern recognition. It is the instinct that says something about this file does not add up, before anyone can articulate why. It is exactly what lets a person catch what the model missed, and it is built through reps, exposure, and years of seeing loans go right and wrong.
Critical thinking is the discipline of interrogating AI output rather than accepting it. The model returns a calculated income figure. Does your team know what it included, what it excluded, and what question to ask when the number looks clean but the file does not? Confidence in an output is not the same as accuracy, and the gap between them is where risk lives.
Strategic thinking is the ability to ask what a tool makes newly possible, not just how it makes the current process faster. Most lenders use AI to run yesterday’s process at a lower cost. Fewer ask what parts of that process should exist at all now. Rather than doing the same with less, why not do more with the same? That question is worth more than the efficiency.
Systems thinking may be the most underrated of all in our business, because a change in one part of the pipeline ripples everywhere. Automate income calculation upstream and you have just changed the work of quality control, investor delivery, and repurchase risk downstream. Leaders who cannot see the whole system will optimize one function and quietly break three others, then wonder why the savings never showed up in the cost per loan.
Notice what all four have in common? Not one of them is a tool skill. You cannot buy them, license them, or install them. They are built in people, over time, by leaders who make it their business to build them.
Which brings up the harder question... Are our leaders equipped to do that?
The role of the manager itself is shifting under this pressure. A Harvard Business School study found that generative AI is flattening hierarchy and moving organizations toward a more skills-based model, blurring the old line between manager and associate as routine coordination gets automated. Think about what that means for a mortgage operation. A significant share of what a manager does today is coordination: chasing status, putting reports together, moving work between queues, answering where things stand. And that is precisely the kind of work AI excels at.
That is a real shift, and I would argue our industry has not prepared for it. We promote our best producers and our best operators into management, then measure them on throughput and turn times. Few of them were ever taught to develop judgment in someone else. Now that is becoming the core of the role.
Consider how much of that skill lives in feedback. Recent Harvard Business Review research on how great leaders handle feedback found that people do not learn simply because they receive it. They learn when the emotions the feedback triggers keep their attention on improving the work rather than protecting themselves. Every leader reading this has given feedback that landed as a threat instead of a lesson. The message was accurate. The learning never happened.
And as machines absorb more of the analysis, a leader’s ability to develop judgment in another human being, through feedback, coaching, and challenge, becomes the differentiating capability of the organization. It is also the heart of what Allie K. Miller describes as the defining skill of an AI-first operator: teaching others how to think about AI, not just how to use it.
So there you have it. The three pillars missing from your AI strategy: staying human-centered; treating talent development as an ongoing discipline; and building the leadership and thinking skills that turn information into judgment and elevate data into wisdom.
Read separately, each one is good management advice. Read together, they are something more specific: the culture that decides whether the money you are spending on AI turns into anything at all.
More on that in the next and final article in this series, where I’ll bring all this together.
This article series opened with a pattern I kept seeing in different forms. The budget was approved.
My Dad loves to watch news in the morning. Today was no different.
Putting This Into Practice
This is the work I do with lenders, banks, and credit unions. If your organization is working through these questions, these pages explain how I help.
