Artificial intelligence: Creating value through experimentation
Artificial intelligence (AI) is not new to CIBC, but thanks to Generative AI, the ways we use it are evolving faster than ever.
Ozge Yeloglu
Jan. 30, 2025
3-minute read
As VP, Enterprise Advanced Analytics and Artificial Intelligence at CIBC, Ozge Yeloglu has deep experience in the field of GenAI and regularly shares her insights at conferences. She believes in the power of AI to elevate CIBC’s capabilities in a client-centered, thoughtful way.
Artificial intelligence (AI) is not new to CIBC, but thanks to Generative AI (GenAI), the ways we use it are evolving faster than ever. As our clients and team members manage through the pace of change in the AI landscape, it’s up to organizations like CIBC to adapt and learn about the ways we can use AI to better support and protect our clients and our bank. Recently, I spoke at the Evident AI Symposium in New York on a panel of peers about how experimentation is driving innovation across the AI landscape.
Here are some key takeaways, and some further thoughts of my own.
Traditional AI has always relied on experimentation. After all, it’s rooted in data science, with the key word being “science.” With GenAI, it seems like there’s a new large language model (LLM) every few weeks. It’s critical we stay on top of these advancements to make sure we’re working with the best solution for the problem at hand. We leverage our Azure Innovation Lab to test these models and stay in constant touch with providers like Open AI, Cohere and many others to help achieve this.
Experimentation is a key step in bringing an AI use case — how data scientists refer to AI projects — to fruition. There is a famous saying in the technology world about how to approach experimentation: “move fast and break things.” At CIBC, however, we’re inclined to take a more thoughtful approach. We’re a people-first bank, and believe in a strategy-led approach: being purposeful about which AI use cases go into production, and being crystal clear on what success looks like. We use our homegrown tools like Business Opportunity Assessment and AI Value Calculator to stay aligned with our strategic focus.
Another important aspect of our experimentation is how we measure value. The challenge is that not all use cases are the same. To help partners overcome this, CIBC created an AI assessment tool that evaluates AI use cases based on the monetary, social or cultural value they bring to our clients, as well as our ability to deliver that value. Most importantly, we incorporate our Trustworthy AI Principles into our AI Risk Assessment Process so we understand how best to manage risks and use AI responsibly and ethically.
Buy-in and literacy start at the top
None of this would be possible without executive support and understanding of how AI will push our strategies forward. Over the past year we’ve embarked on important AI training and education for our most senior executives, in partnership with academic experts from MIT Media Lab and Rotman School of Management. This training helps ensure AI is understood and considered in our most senior leader conversations and approaches. Training sessions included topics such as data quality, governance and risk, and understanding how to use AI for high-impact use cases and to drive business value.
I strongly believe experimentation is a crucial part of our thoughtful AI journey. No matter the outcome, we always learn from it and continue to move forward to achieve our purpose of helping clients achieve their ambitions.
Ozge Yeloglu
Vice President, Enterprise Advanced Analytics and Artificial Intelligence
