- Key insight: Banks should look to their history and be deliberate about how they integrate AI into their systems.
- What's at stake: Many institutions are trying to apply AI to workflows
that are not yet well understood, not consistently executed or not even clearly owned. Automating confusion produces faster confusion at scale - Forward look: The institutions that win at this moment will be the ones that had the patience to build the foundation first, the courage to say no when necessary and the confidence to move at a pace that matches the responsibility of the industry they serve.
Every earnings call
That pressure is exactly why banks need to slow down.
The institutions that will see durable efficiency gains from artificial intelligence will not be the ones that rush to deploy tools for the sake of signaling innovation. They will be the ones that treat AI the way banks have always treated balance sheets, credit risk and core conversions: methodically, deliberately and with a sharp eye out for unintended consequences.
There is a growing misconception in the industry that speed itself is the strategy. That if a bank is not actively piloting multiple AI use cases, it must be falling behind. That belief is not just flawed, it is dangerous. Banking is not a social media platform, and operational shortcuts rarely end well in regulated environments built on trust.
The most effective efficiency gains in banking have never come from shiny technology alone. They come from clarity. Clarity around processes, ownership, data quality, accountability and outcomes. Artificial intelligence does not replace the need for that clarity. It amplifies the cost of not having it.
In my experience, many institutions are trying to apply AI to workflows that are not yet well understood, not consistently executed or not even clearly owned. Automating confusion does not produce efficiency. It produces faster confusion at scale. When leadership teams ask why early AI initiatives fail to deliver meaningful return, the answer is rarely that the models were not powerful enough. More often, the underlying processes were never ready to be automated in the first place.
There is also a tendency to frame AI as a shortcut to transformation. A way to leapfrog years of operational discipline with a single implementation. That framing misunderstands both AI and banking. The real value of artificial intelligence inside a financial institution is not in replacing judgment, but in supporting it. It works best when it augments well designed human decision making, not when it is asked to compensate for gaps in governance or institutional alignment.
Banks and credit unions are pairing AI-driven efficiency with stable staffing and cross-training to scale mortgage production as originations rebound and technology expands capacity.
The banks that are making real progress tend to start small and unglamorous. They focus on narrow use cases with clearly measurable outcomes. They invest time in data classification and access controls before experimenting with automation. They treat model outputs as inputs into existing controls, not replacements for them. Most importantly, they accept that some efficiency gains only appear after months of work, not after press releases.
This approach can feel unsatisfying in a market obsessed with speed. It does not generate flashy headlines or immediate validation. But it builds institutional muscle. It creates repeatable patterns for evaluating risk, measuring value and scaling responsibly. Those patterns matter far more than any single AI deployment.
There is also a cultural dimension that rarely gets discussed. Rushing AI initiatives sends an implicit message to staff that speed matters more than understanding. That adoption matters more than trust. In an industry already grappling with change fatigue, that message can quietly undermine engagement. When AI is introduced as a deliberate tool to remove friction and improve judgment, rather than a mandate imposed from above, adoption follows naturally.
None of this is an argument against artificial intelligence. Quite the opposite. AI represents one of the most meaningful opportunities banks have had in decades to improve efficiency, consistency and customer experience. But opportunity does not require urgency for urgency's sake. It requires discipline.
Banks have survived and thrived for generations by being skeptical when others are euphoric. By asking uncomfortable questions when markets are loud. By choosing resilience over optics. Artificial intelligence should be no exception.
The institutions that win at this moment will not be the ones that chased the trend hardest. They will be the ones that had the patience to build the foundation first, the courage to say no when necessary and the confidence to move at a pace that matches the responsibility of the industry they serve.
Slow is not a lack of ambition. In banking, it is often a sign that something is being done right.
