How EY’s finance transformation team is approaching AI strategy

As finance leaders face pressure to modernize and deliver ROI on their spending on technology and consulting, EY’s finance transformation team is focused on putting emerging technologies to work not just for clients, but inside the firm itself. The team, led by Deirdre Ryan, global finance transformation leader, is playing a dual role: Helping CFOs navigate AI adoption while also piloting those same tools internally across EY’s finance and consulting functions.

In a recent interview inspired by her session at the Gartner Finance Executive Conference last month, Ryan explains how EY is using agentic AI to reshape FP&A workflows and why being “client zero” is critical to building credibility in the market. She also discusses how CFOs can avoid repeating past mistakes from automation efforts, what it takes to lead a finance organization through transformation and how to do so with clarity, purpose and psychological safety.

Deirdre Ryan

Global Finance Transformation Leader, EY

Notable previous employers:

• Deloitte
• Dontech
• Dun & Bradstreet 

This interview has been edited for brevity and clarity. 

ADAM ZAKI: What types of new technology are you seeing success with internally? 

DEIRDRE RYAN: We feel very strongly that we have to be client zero. If we’re going to advise clients on new technologies, we need to understand them ourselves and use them in real scenarios.

We created a platform called EYQ. It’s essentially a private environment where our people can interact with large language models securely. We made it accessible on laptops and mobile devices, and it’s helped our consultants build hands-on understanding of the tools we’re asking clients to adopt. As of recently, we’ve had over 150,000 consultants using it globally. It’s one of the largest private LLMs in the world that EY developed in-house. 

In finance specifically, we’ve been building and piloting an agentic AI solution for FP&A. It looks like a normal dashboard, but what makes it powerful is that as actuals come in, it generates AI-driven insights automatically. That’s helpful, but the real impact comes from scenario planning. It’s built on driver-based forecasting, so we’ve identified the variables most correlated with forecast accuracy. You can adjust those and instantly see how the outlook changes.

It goes even further. We’ve modeled it so that there are three AI agents working like analysts, with a manager agent that synthesizes and returns the best answer. You can ask something like, “What would a one percent drop in GDP do to our forecast?” and it does the work. It’s not removing human oversight — someone still has to take action — but it’s changing the way FP&A work gets done.

One client saw it and said, “I have an army of silent FP&A analysts now.” That stuck with me because that is where the function is headed.

You mentioned the impact that tools like this can have on the finance team. How should CFOs be thinking about psychological safety, something your colleague Myles Corson stresses, during a transformation like this?

That brings up another important point. Psychological safety is something we talk about a lot. When tools like this are introduced, it’s natural for teams to wonder what it means for them. They may worry their work is being replaced or question what their future looks like in the function.

This is where leadership matters. People entering the workforce today don’t want to spend three days in Excel. They want to work with tools that help them think and act strategically. If you’re in FP&A and you’re given the choice between spending days building a model in spreadsheets or using agentic AI to get that answer instantly and then focusing your time on what to do about it, people are going to choose the latter. That’s how you retain talent. If finance doesn’t evolve, it risks losing its best people.

So yes, we’re advising clients on these tools, but we’re also living it internally. We’re applying it ourselves, and we’re navigating the same leadership, talent and change management conversations that our clients are. That’s what being client zero means.

You’ve said publicly that CFOs need to be “role models” when it comes to using AI. What does that look like in practice?

It’s difficult for CFOs today because they still play a very traditional role. They must protect and preserve the assets of the organization and mitigate risk, but now they’re being asked in a meaningful way to drive innovation within finance and across the enterprise. They need to understand disruptive technology well enough to make smart capital allocation decisions and guide the business forward.

So, CFOs have to start getting their hands dirty. A lot of people I meet have seen demos or presentations, but haven’t used the tools themselves. You have to understand the capabilities. Start small — maybe it’s a proof of concept to help your team come up the learning curve — but that gives you insight into what these technologies can do.

And from there, you can ask the bigger question: How do we apply this in a meaningful way to our finance organization? That’s why our team tells CFOs to not just look at the tech, but think about the end game. What do you want your finance function to look like once you’ve integrated these tools?

Is that something CFOs should handle with their teams in a structured way? Should they do a “AI tip of the week” meeting? Or should it be more ad-hoc, like, “Hey, I found this new way to do something, let’s explore it on a group call”?

You have to do some things in parallel, which is tough because CFOs are already being asked to do so much. But this is one of those areas where you can’t afford to take a one-track approach.

You don’t want to repeat what happened with robot process automation. Very few companies realized the value they expected. It became very democratized — people used it to automate a few hours of work here or there — but it didn’t lead to large-scale transformation.

That’s the risk with AI and generative AI. The technology is unbelievably powerful, but without a strategy, you end up with fragmented efforts. You have to ask: Where is the puck going to be, and how do we get there? That means setting a clear end state, helping your finance team come up the learning curve, and avoiding what I call death by a thousand cuts — a little pilot here, a tool there, but no cohesive vision.

So, yes, you want experimentation, and maybe that’s informal — sharing a cool use case in a meeting. But it also needs to be backed by a very intentional strategy tied to how your finance function delivers value.

For CFOs who are trying to find this sweet spot between productivity and ROI in their technology implementations, how would you advise them on their initial approach?

For this, there are two big buckets I talk about with CFOs. One is productivity, and yes, you can absolutely drive productivity using these tools. We have great examples of that. And honestly, any of my competitors could give you the same 200 use cases for technology within finance. So I’m not saying that’s a bad thing, but many of those use cases have been around for a long time. So if you’re going to pursue productivity, you need to ask where you’re going to get the biggest ROI. What’s going to move the needle?

The second bucket, and this is where I think the real value is, is decision insight. That’s about using these tools to provide better analysis that helps your peers in the C-suite make smarter, faster decisions. And while that’s much harder to quantify, I think it’s equally important.

Sometimes I ask CFOs to imagine a scenario. Let’s pretend your data is perfect, it comes in on time and everything is consistently defined. Of course, that never happens, but let’s just pretend. What is the kind of analysis you’d want to do on demand that could give you a competitive advantage?

And it’s interesting, because many CFOs haven’t even had the [capacity] to think that way. They’re so tied to traditional metrics like revenue and profitability that they haven’t had the chance to ask, “If I had access to better data and AI tools that let me explore it faster, what decisions could I make differently?” That kind of thinking is where AI can really change the game for finance.

There’s a debate in finance between those who believe the “single source of truth” requirement is a myth and those who think it’s pivotal to infusing new technology into finance. Where do you stand on this?

I think it depends on what you mean by “a single source of truth”. We all know CFOs need to ensure the financial statements are accurate. And with technology of all types, there has to be a level of trust that the data is producing results that fairly represent the performance of the organization.

Do I know any company whose data is 100% perfect all the time? No. Especially not large, acquisitive organizations. But what I always tell clients is, you have to prioritize. Not every data point needs to be perfect, but the ones that drive the most value do need to be consistently defined and captured across the enterprise. You could spend the next 10 or 20 years cleaning data, and it still wouldn’t be perfect.

The better approach is to identify the data that will drive meaningful analysis and ensure that it’s reliable. That way, when you present insights to the executive team, you have confidence in the underlying information and the decisions it supports. It’s about being intentional. Know what value you’re trying to unlock, and focus on the data that supports that value.