Capital One, ranked 82nd on the Fortune 500, has long been recognized for its data-first approach and early adoption of cloud and AI. Amy Lenander, Chief Data Officer, has spent over two decades helping shape that evolution — transforming the company from a data-savvy bank into a leader in AI-driven innovation.

Today, Lenander leads Capital One’s enterprise data strategy, advancing its vision of using data to deliver smarter, more impactful financial products.

In this conversation with CDO Magazine, she reflects on the company’s data transformation journey, the foundations for scaling AI responsibly, and how a “data-as-a-product” mindset is fueling enterprise-wide innovation.

Edited Excerpts

Q: You’ve been with Capital One for over 22 years. Can you walk us through the company’s digital transformation journey? How has your data strategy evolved, especially as AI becomes increasingly central to the business?

Data has always been at the heart of Capital One. The founding idea of the company was to use data and analytics to deliver better financial products and services tailored to customers’ needs. Since then, the technology to realize that vision has only gotten better.

We became the first U.S. bank to go all in on the cloud, enabling us to leverage data and technology more nimbly and effectively than ever before. This opened up a lot of opportunities for us in data and helped us to develop a modern data ecosystem that would be ready for data, machine learning, and AI at scale.

Today, with even greater advances in AI and even more data, there are more ways we can capitalize on our data strategy and ecosystem. It means we’re looking closely at things like unstructured data and ensuring data is available in real-time in order to use AI models in real-time.

Tapping into our own data continues to be an advantage for us, and now we have much more of it. Our approach hasn’t changed, but execution has evolved to keep pace with the growing volume and importance of data.

Q: Having previously served as a CMO and a CEO of Capital One UK, how has your background on the business side shaped how you lead in data today? 

Across every role I’ve had with the company, I’ve used data to drive business decisions. That experience helped to build my intuition on what type of data is most important for business teams across the company, which helps me understand which capabilities are most important to focus on.

My experience working in the business also helps me have empathy for the pressing business needs of our business partners when I need to ask them to prioritize something that is important for us to forward our enterprise data strategy.

Q: What inspired the launch of Capital One’s Chat Concierge AI for car buyers, and how does it reflect your broader approach to responsible innovation?

We built Chat Concierge as a proprietary multi-agentic conversational AI assistant for car buyers and dealers in early 2025. It can perform tasks like comparing vehicles to help car buyers decide on the best choice for them and scheduling appointments with salespeople. 

The framework we built to support Chat Concierge can be extended to other customer- and internal-facing use cases. Whenever we look at cutting-edge technology, we must balance these opportunities to enhance both customer and business experiences with a well-managed, risk-centered approach.

Q: What are some of the biggest challenges in deploying AI models at scale within a financial institution like Capital One? 

Even before AI or generative AI, we’ve been leveraging machine learning at scale across our business. We know that when it comes to addressing challenges of deploying AI models at scale in a well-managed way, enterprises need to have a set of core capabilities in place. 

This includes the ability to prepare and curate high volumes of high-quality data, have capabilities for observability, monitoring, benchmarking, and analytics, and be ready to refine and improve the models over time post-deployment, to name a few. 

We’ve found the best way to manage these challenges is by having central platforms that can scale. Of course, all of this relies on certain precursors for success, like the readiness of your tech stack, the ability to trust and use your own data with a strong data ecosystem, and top talent. We’ve been investing in those foundations for years, which have prepared us well for this AI moment. 

Q: How has the “data-as-a-product” approach helped unlock new business value or innovation in your organization?

We have invested in building out a “data-as-a-product” approach that curates our regularly used data and data that has the potential to drive innovation in the future.

Making that data easily available with a high level of quality is already speeding up our ability to identify innovations and implement them, even though we’re still relatively early in our journey. We continue to bring more and more data into this framework and champion the teams that are building data products.  

Q: What technologies or frameworks are most critical for enabling scalable data product management?

One of the things that is most critical for the success of data products is elevating the role of the person who’s defining the intent and building the data product: the data product manager. 

This is a critical, senior-level role that is focused on not just making data standardized and accessible, but also thinking about what the future of what potential uses of that data might be, and driving innovation from the data product itself. This also involves thinking about the things we aren’t doing today, but which data could enable, and helping advocate for that across the organization.

Q: If you could automate one annoying part of your day-to-day with AI, what would it be?

For me, it comes down to something more mundane: everyday home life. It would definitely be nice to have AI autonomously remind me — or even help me — with my list of house chores and ongoing maintenance in a proactive way. I’d love it if AI could source trusted local and cost-effective service providers to help with this work and schedule them in a way that works with my schedule.