M&T Bank, a Fortune 500 institution with around 1,000 branches, over 22,000 employees, and a 165-year history, operates as one of the largest regional banks in the U.S. The bank serves millions of customers across the Northeast and Mid-Atlantic and is known for its disciplined approach to risk, its community banking ethos, and long-standing focus on customer trust. Against this backdrop, the bank’s data and AI transformation has become an essential pillar of how it competes and delivers value at scale.

M&T’s Chief Data Officer (CDO) Andrew Foster sits at the intersection of data strategy and AI adoption, overseeing both the enterprise data agenda and the bank’s approach to emerging AI tools. In this first part of a three-part interview with Parker Thompson, Regional VP and General Manager, East at Denodo, Foster outlines the milestones behind M&T’s data strategy, the disciplined path to deploying AI tools to 17,000 employees, and the foundational governance required to protect customer trust while accelerating value creation.

Edited Excerpts

Q: You have been with M&T for about two years. What would you highlight as your key milestones in developing and executing the data strategy?

Data is a tough topic, and data professionals sometimes want everyone to understand our data geek language. That does not always work when you are partnering with heads of business who have deep domain expertise but are not experts in data.

So the first step for us was to ask how we know where we are. We used the EDM Association’s Data Management Capability Assessment Model (DCAM) to do a current state assessment and then determine where we wanted to go. That became our first two-year strategy from October 2023 through October 2025. We have executed successfully on it, secured good investment, and now the focus is on looking ahead to the next two years.

Q: As you mentioned, now that the data strategy is directionally where you want it, you are pivoting toward AI and particularly Copilot. How do you see that journey unfolding?

It is interesting. I just came off a call with one of the top four consulting firms, and they were talking about their Copilot deployment. I firmly believe that Copilot, or tools like it, are table stakes for any big organization now.

That has been a great success for us. We have rolled out Copilot to about 17,000 employees, and now the focus is on unlocking the value we know we can create. The first hard step is done. With banks and other well-regulated institutions, you must have your compliance and governance ducks in a row. You need to invest the time in that and prioritize safety first. That is what we have done.

Now it is about creating value. We are well underway. We have rolled out Microsoft Copilot Chat and Microsoft Copilot 365 to 17,000 employees.

Q: Working in a heavily regulated industry, what sensitivities did you need to address before deploying at that scale?

People in my role can feel pressure based on what they see in the media. Every next article is about someone doing the next best thing in AI. It is important to ignore that. My mantra is always, “What do we ground ourselves in?” And we ground ourselves in being a community bank and having the trust of our customers. Job number one is to keep doing those things.

When we looked at Microsoft Copilot, we ran an extended proof of concept for about six months. I expected the technology to work, and it does. The real work was bringing all our partners to the table. Compliance, privacy, legal, HR, first-line risk, second-line risk, and audit. It’s one large group that needed to be educated and then aligned with our corporate partners at Microsoft. We needed to reach a place where we could say, “This is how we are going to implement it and this is how we are going to stay true to our core principles.”

Q: The rate at which you are executing, both on the data strategy and the incremental AI approach, is unusually fast in banking. Please tell us more about it.

I will answer that in two ways. When I look at our investment in data and data-enabling cloud-based technology, we have moved quickly. Since April 2024, we have onboarded four to five new target-state enabling technologies, focused on data governance and the data science workbench space.

With Copilot and AI, we are probably in the middle of the pack. There are large banks that are far ahead. They were early at the table and have invested heavily. There are also organizations that are lagging. For us, it is about the right pace for M&T. Step one was building the control conditions we needed to feel comfortable adopting. Now the priority is accelerating value creation. That is what matters.

Q: How did data governance and value capture factor into your AI roadmap?

AI is broad and all-encompassing. It means different things to different people. You need a way to frame it for your organization. We look at it as six or so categories of investment, such as personal productivity, cyber, credit, and other use case categories.

In terms of capabilities, you have Copilot in its different forms, embedded AI within third-party applications, and differentiated capabilities that you build yourself. When you look at that matrix, the intersection point is talent.

We have pilots and real deliveries underway within those investment categories. Copilots, for example, are a good personal productivity example. But we must decide how best to go after each category. The intersection point must be grounded in talent.

If you are talking about a new form of AI enablement, you need a leader who will drive that forward in the organization, someone who can re-imagine how they approach their business unit with these technologies so they can be successful over multiple years. You need people who are willing to lean on their domain expertise and also reassess what it means in the context of new technology. Then you must bring those elements together in a way that allows you to succeed.

CDO Magazine appreciates Andrew Foster for sharing his insights with our global community.