We’ll Have AI as a Top-line Budget Item for 2025 — Fifth Third Bank Director of IT Data and AI Governance

(US & Canada) Jonathan Paul, VP, Director IT Data & AI Governance at Fifth Third Bank, speaks with Peter Geovanes, Chief Innovation and AI Officer, in a video interview about the approach to adopting latest AI advancements, the economics of adopting new AI technologies, and building the AI innovation team.

Speaking about the approach to coping with the rapid pace of AI advancements, Paul points out the importance of building systems with agility to incorporate new technologies as they emerge. His team is often quick to identify and share the latest innovations, whether it's from OpenAI or other major players, allowing them to stay ahead of the curve.

Paul likens the situation to solving a Rubik's Cube, where understanding the bigger picture is crucial to progress. He highlights that the AI landscape is highly volatile, with some technologies being replaced in a matter of months. This evolving environment necessitates a flexible, modular architecture that enables quick integration of the latest AI services and components, regardless of whether they are open-source or proprietary.

Additionally, Paul stressed the need for adaptable solutions that can support a range of use cases, from text-to-speech to more complex AI models. As technology evolves, it’s critical to maintain access to the best models available, knowing that today’s cutting-edge solution might be surpassed within days.

Speaking of balancing the costs of using AI tools, Paul emphasizes that economic factors play a significant role in selecting AI models, though they are often overshadowed by the initial excitement surrounding new technologies. He notes that, much like cloud services, where pricing models must be adjusted based on usage, AI model costs require careful consideration. Teams responsible for taking AI projects from concept to production must thoroughly assess both performance and cost efficiency. Paul's organization conducts a rigorous comparison of models, evaluating them on multiple criteria beyond just speed or size. Ultimately, they select models that align with the use case and fit within budget projections.

When asked about building a talent pool for AI, Paul describesthe team he assembled as a dynamic blend of talent, ranging from seasoned data scientists to early-career professionals. He mentions how younger team members are gaining valuable experience, particularly by experimenting with large language models through internships and cooperatives. This approach allows the organization to leverage fresh talent while also drawing on the expertise of more experienced modelers, especially in areas like traditional machine learning (ML) and AI, to fine-tune technical parameters.

In a similar vein, Paul also highlights the importance of translating technical knowledge into accessible language for broader organizational understanding and says that the team collaborates closely with ethics committees to ensure responsible AI usage. The legal and compliance functions play a critical role in this structure, contributing by piloting tools like document summarization systems. The innovation team is constantly pushing boundaries, exploring new technologies and use cases, and considering both in-house development and external solutions.

Further, Paul likens the current landscape to the dot-com boom, noting that while there’s tremendous opportunity, there are also risks. The AI and fintech industries are rapidly evolving, requiring a strategic approach to integrate these emerging technologies into their ecosystem.

Sharing his thoughts on budgeting for AI given the rapid rate of adoption, Paul says that the organization is making significant strides in integrating AI, with 2024 focused on establishing key governance foundations and rolling out technologies to employees. He notes that a commercial chatbot was successfully deployed across the entire workforce, alongside generative AI training made accessible to all bank employees. This initial phase aimed to spark interest and gauge the organization's readiness for broader adoption.

Looking ahead, Paul reveals that AI will become a top priority in the 2025 budget, with a focus on strategic business deployment and ROI. The goal is to move beyond basic productivity enhancements, like using chatbots to summarize documents or draft emails and shift toward intelligent automation. By linking these tools together, the organization plans to streamline processes and eliminate inefficiencies, unlocking the full potential of lean process automation—something that wasn’t feasible just a year ago.

CDO Magazine appreciates Jonathan Paul for sharing his insights with our global community.