With assets exceeding $250 billion and a history spanning over 160 years, Fifth Third Bank stands among the top U.S. financial institutions — known for its innovation in digital banking, customer experience, and financial data insights. As the organization accelerates its data and AI transformation, it’s also redefining what it means to responsibly deploy AI in one of the most regulated industries.

The first part of this three-part series unpacked the bank’s evolving ModelOps journey and how it is laying a foundation for trusted, explainable AI. For the second installment, Shanna Anderson, Senior Director of Customer Data and Insights at Fifth Third Bank, sits with Or Zabludowski, CEO of Flexor, to share how the bank is blending disciplined governance with innovation — ensuring that every AI initiative has a clear, measurable impact while keeping people at the center of decision-making.

Building operating rhythms for responsible AI

Anderson opens the discussion by grounding Fifth Third’s approach to AI in strong data governance. She emphasizes that before any organization can truly scale AI, it must first get its data house in order.

 “Data is the crux of that and the quality of what you’re able to do with AI,” she says. For Fifth Third, the groundwork lies in creating the right operating rhythms across both business and technology teams.

To ensure alignment and focus, the bank has established dedicated AI focus groups tasked with exploring advanced use cases, generative AI applications, and intelligent automation. These cross-functional units operate in tandem with the bank’s broader technology teams, all under the same governance framework.

“We spend a lot of time on the governance foundation first,” Anderson explains, highlighting the importance of applying a disciplined lens to any AI deployment. “Spending time on the maybe non-exciting things first has helped us. We’re able to accelerate and understand what those use cases are that make sense for us and also yield productivity.”

4 pillars of AI strategy

Fifth Third’s AI playbook is guided by four key pillars:

1. Empowering employees
2. Enhancing engineering and business processes
3. Improving customer engagement
4. Delivering outcomes through aligned value streams

These pillars ensure that every initiative is anchored to tangible business outcomes and value streams, going through rigorous steer codes and governance reviews before deployment. “It’s focusing on the things that we think are going to make a difference,” she says, adding that each use case must “show up for the right development and productivity use cases.”

Real-time decisioning: Speed meets compliance

Anderson shares that her team is “working on models that will enable real-time decisioning,” though they have not yet reached production. These models, she explains, will support use cases like credit and customer decisioning, requiring both technical precision and regulatory care.

Recognizing the high stakes of decisions made at the edge, especially in regulated industries like banking, Anderson is clear-eyed about the importance of compliance and control.

For now, her teams are investing heavily in testing, compliance validation, and other areas, “making sure the models are fully tested and meet the right type of governance and compliance.”

Still, she acknowledges the imperative to move fast where it matters: “Speed to decision is definitely [important] for us in terms of a customer wanting a decision as fast as possible.”

Yet, Anderson underscores the importance of balancing speed with integrity: “We’re not looking to deploy anything that we don’t feel really good about.”

Human feedback in an AI-driven world

Beyond the algorithms and models, she is just as focused on the human side of AI — both customer and internal stakeholder experiences. One example is the bank’s Genie Chatbot, which assists customers on mobile and online banking platforms. “It’s not necessarily making decisions for the customer,” Anderson clarifies, “…but it’s giving the customer information based on what the customer desires.” The chatbot not only helps customers get answers but also collects their feedback to continuously refine the experience.

The feedback loop extends to real-time decisioning initiatives as well. “We still keep a human in the loop so that a human can still look at the types of decisions being made.”

Rethinking dashboards in the age of GenAI

As the conversation turns inward, Anderson discusses how AI is transforming how internal teams access and interpret data. Historically, internal “customers” at Fifth Third relied on a robust suite of dashboards to assess performance, customer engagement, and operational metrics. But that is evolving rapidly.

“With what we’re starting to see with some of the generative AI capabilities and large language models,” she says, “it starts to pull in a question of, like, what’s the need of having that robust dashboard suite in the future?”

Instead, she foresees a shift toward natural language querying, where bankers and business leaders can simply ask questions and receive instant insights. “It’s thinking more in terms of how do we just provide the right access and use natural language where we can just query what’s needed just in time for those types of analytical or business decisions,” Anderson concludes.

CDO Magazine appreciates Shannon Anderson for sharing her insights with our global community.