Banking and financial services are in a tricky position when it comes to AI. Being highly regulated, intensely risk-averse, as well as massive targets for cyberattacks, means that there is a lot standing in the way of meaningful progress. However, with mounting pressure from customers asking for better experiences, even the most risk averse are exploring the possibilities of AI. 

At our recent Banking Innovation Summit, it became clear that AI isn’t just a new technology, it’s a whole new ball game. For senior banking leaders steering this transformation, success depends on mastering two imperatives: trust and scale. 

Customer Trust in the AI Era 

Trust is foundational in banking—and that doesn’t change in the AI era. What does change is how it’s earned. Customers expect transparency, control over their data, and assurance that AI systems act in their best interests. 

Speaking at our keynote panel discussion, Dr.Denise Turley, the VP of Corporate Systems at U.S. Chamber of Commerce, emphasized that it’s best to move “deliberately and slowly, particularly in banking and these regulated industries, because trust is paramount.” 

This was echoed by Srini Nidamanuri, VP of Digital, Data & Analytics at Equifax: 

When an industry is as deeply ingrained in trust as banking is, if you lose a customer’s trust, you lose the customer. Gone are the days of implicit customer trust. Robust governance, transparent AI interactions, and clear data consent protocols are no longer optional, they’re non-negotiable. 

Your Employees and AI 

It’s not just your customers you need to get on board. Internal trust is just as critical. Employees must understand how to use AI tools confidently, recognize their limitations, and validate outputs before customer-facing use. 

Ongoing training, process design, and human oversight are the safeguards that prevent AI errors from becoming business risks. 

Appeasing the Regulators 

What about the regulators? If there is anyone that needs assurances your AI tools are trustworthy, it’s them. As Scot Lynch, Executive Director at Morgan Stanley, so eloquently put: 

This means that your AI models must be transparent, auditable, and explainable. Turley added to Lynch’s point, likening the immature regulatory landscape to “driving the car while we’re building it.” 

When it comes to driving a half-built car, all our panelists were aligned: having evidence and being able to demonstrate exactly how your models operate is essential for any AI implementation. 

You Want AI? How’s Your Data? 

While trust is foundational, implementing and managing AI effectively in banking is fundamentally a problem of scale. 

Data and operational scale are immediate challenges. AI is a data game. Data security in large financial services organizations, however, is described by Lynch as “a game of locking the data down.” While AI is really good at scale, the underlying task remains immense. 

Automating processes, understanding complex transaction patterns at scale, and building a “knowledge fabric” from vast amounts of transactional and third-party data can all leverage AI. Getting your data to a point where it can make that happen, securely—that’s the challenge. 

Infrastructure & Energy 

In all the excitement of AI, something that doesn’t get mentioned enough is the reality of the infrastructure and energy challenge. 

Large AI models and the hardware supporting them are increasingly power-hungry, running faster and hotter all the time. Competition for the resource to run your AI models is also heating up. 

Leaders are concerned that if this trend continues, we’re going to reach a point where electrical power governance is introduced. Forward-looking banks are rethinking infrastructure strategy, exploring smaller, distributed data centers in lower-demand regions. Proactive planning today could be the differentiator tomorrow so maybe stay away from Silicon Valley. 

Governance at Scale 

Of course, for any AI implementation to get off the ground, you must be able to adapt to new governance at scale. Fortunately (and unfortunately) the governance model for AI is currently still in the teething stage. 

Without much guidance to work with, scaling your governance is a matter of continuous effort. There is a constant need for feedback loops, monitoring model performance, adapting to data drift, and the integration of new models and data patterns. 

Navigating the Future 

For banking leaders, success with AI will rest on two pillars: earning trust and enabling scale. From customer-facing tools to back-office automation, these imperatives will determine whether AI becomes a competitive advantage or a compliance nightmare. 

Trust demands transparency, ethical oversight, and human judgment. Scale demands strategic data, resilient infrastructure, and agile governance. Get both right, and the promise of AI is well within reach. 

To continue exploring how your peers are navigating these challenges, and shaping what’s next, join us at one of our upcoming Banking Innovation Summits. 

To see all our upcoming summits, visit our events page.