EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
How Agentic AI can transform enterprise operations through co-innovation
In this episode of the EY Microsoft Tech Directions podcast, explore how Agentic AI is transforming enterprise operations through the power of co-innovation.
Step into a lively discussion where experts from EY and Microsoft share how Agentic AI is changing the game for enterprise operations—especially within the SAP ecosystem. Rather than sticking to traditional, reactive AI, the conversation explores how intelligent agents are now able to sense, reason, and act, bringing a proactive edge to business processes.
You’ll hear how this evolution isn’t just about adopting new technology. It’s about building trust, developing expertise, and having a strategic vision to truly unlock value. The speakers dive into the real-world impact: organizations can move faster, make smarter decisions, and get more out of their SAP investments, all while adapting to today’s unpredictable business landscape.
The episode doesn’t shy away from challenges. Leaders talk candidly about overcoming resistance to change, tackling talent shortages, and ensuring strong data governance. There’s also a thoughtful look at workforce transformation—while some worry about job displacement, the consensus is that Agentic AI frees people from repetitive tasks, letting them focus on work that matters most.
Throughout, the importance of upskilling and continuous learning is emphasized. Success with Agentic AI, especially in SAP environments, depends on a culture of curiosity and collaboration between humans and AI. If you’re interested in how co-innovation is driving measurable outcomes and reshaping enterprise systems, this conversation is for you.
Speakers:
Mark Luquire: Global Microsoft Alliance Co-innovation Leader; Managing Director, Microsoft Alliance, Ernst & Young LLP
Pat Sullivan: Global Digital Engineering Lead; Ernst & Young, LLP
Mark Skoog: Managing Director, Strategic Partnerships, Microsoft
Key Takeaways:
Agentic AI Represents a Shift from Reactive to Proactive Intelligence
Strategic Advantages and Workforce Impact
Adoption Challenges: Data Quality, Talent and Change Management
Measurable Business Outcomes Through Partnership
Curiosity, Continuous Learning and Human-AI Collaboration Are Essential
Mark Luquire
Hey, everyone. This is Tech Directions, where EY and Microsoft professionals explore transformative enterprise solutions. I'm Mark Luquire, the global co-innovation leader at EY for the Microsoft Alliance.
Today, we're exploring a shift that's quietly but rapidly transforming enterprise operations, Agentic AI, and what it means for organizations running SAP. We've all heard the promises of AI, but Agentic AI is about intelligent agents, software that can sense, reason, and act, embedded in your most critical systems.
Why now? Businesses are under pressure to do more with less, move faster, and unlock more value from existing investments, all in an unpredictable environment. For interconnected platforms like SAP, Agentic AI is fast becoming table stakes.
But technology alone isn't enough. You also need trust, deep expertise, and the ability to turn innovation into results at scale. That's why today's episode brings together leaders from EY and Microsoft, two organizations at the forefront of SAP transformation, to unpack what's changing, what's possible and how companies can take their next steps with confidence.
And as always, before we get started, nothing in this podcast should be relied upon as accounting, legal, investment or other professional advice. Please consult your own advisors.
So today, I want to introduce a couple of guests that we have. Mark Skoog, you lead SAP Innovation at Microsoft. Where are you joining us from today?
Mark Skoog
I'm joining us from Chicago today.
Luquire
Great to have you. Pat Sullivan, you focus on enterprise transformation at EY globally. Where are you joining us from?
Pat Sullivan
Hey, Mark. I'm actually joining from Dallas today.
Luquire
Awesome. It's great to have you guys, and glad we could get together and do this.
So, Mark, what makes Agentic AI different from other forms of AI?
Skoog
I think it's understanding that it's a profound shift from reactive to proactive. If you think about artificial intelligence, in the past, it was more in a reactive versus collaborator state. And now you think about Agentic, it can take initiative. It can be proactive and engage. I think that is one of the major differences between where we were and where we're at.
Luquire
Yeah, it's really evolved very rapidly, right? What do you think some of the biggest misconceptions and sticking points you hear from business leaders are?
Skoog
I think there's a couple. I think one that we hear about quite a bit is, how many years is it going to take to build a platform? What's the final state of what I need to build out?
I think when you think about Agentic, it's really about figuring out how you take advantage of it today and understand it's going to change over time, but the tools help you get there. And you can use the Agentic capabilities to help morph over time.
I think one of the things we think about is, how do you provide a consistent user experience to these changes? Betting on one single model or one single technology, I think, is high risk. That's why in Microsoft, we're thinking about how do we connect to the tens of thousands of ISV applications or over 11,000 language models today.
We see this fundamental shift from labor arbitrage to the next 30 or 40% of value in the enterprise through agentic, and so that's an increase from 10 to 15% on the average customer.
Luquire
Yeah, it's pretty significant. I know Microsoft has really leaned in in terms of giving people choice as well, the different models available across the Microsoft Cloud and other capabilities. Pat, from the outset, what should leaders really understand about agentic AI?
Sullivan
I'd say number one is you need to understand it, right?
Almost from the very simplistic root of just what agentic can be and what it does.
But then when you get into it, it's really getting into, I would say, four or five things like strategic advantages. What are those advantages that agentic can bring to businesses, to leaders, to companies, whether it be operational cost, decision making.
The impact on a workforce, there's a lot out there as it relates to just agentic, and generally, a lot of that news is negative. In my business, we work in a knowledge space, and that's what AI is great at, is knowledge. But actually, the impact on a workforce, I think, is going to be incredibly positive.
Another key thing just that I think leaders have to be aware of is the regulatory landscape. We spend a lot of time with our clients around compliance. This is evolving. It's going to continue to evolve as countries and unions are going to enhance what they do, especially around data privacy. AI is going to be front and center. So understanding that regulatory landscape is very important.
This is something where EY and Microsoft, between understanding the business side and the technology side, we actually pair up very well together with clients.
And, Mark, I'll say the last piece that leaders really do need to be focused on [is that] overall long-term vision. Everything with AI right now, I like to tell people, is that the worst AI is ever going to be is literally right now.
But having a long-term vision instead of a short-term myopic one is going to be very important because what Agentic can do is totally different than what we've seen in the past with process automation and things of that nature.
Luquire
I think the pace of change is unprecedented, and we're going to continue to see that.
What pitfalls or missed opportunities do you see, especially for organizations with SAP at the core?
I always try to help people understand that it's time to re-imagine everything. It's not just time to automate the process that you have, but re-imagine how you get to the outcome.
Sullivan
Yeah. Look, first off is just adoption and understanding.
The technology is moving at such a rapid pace that understanding how Agentic within an SAP environment can really enhance how all the data that sits inside those systems can be leveraged for a corporation is huge.
There's going to be resistance from employees. Just it's a fear factor. So overcoming that and having a game plan on that is really important.
But then, honestly, Mark, right on the other end of this is that there's a talent shortage and a skill gap in the market place right now. Just because there's Agentic and things of that nature that are out there in great technologies, they still need to be integrated. There has to be data that is sourced for this. Those are things where we need human talent to help derive that in order to set this up to get the best success from it.
Remember, you can have as much AI as you want, but if your data is not good or it's not set up properly, you're not going to have very good AI. It's very, very important to get that understood and have a plan and then govern that very meticulously as you go along.
Luquire
I think that's spot on. You mentioned upskilling and training. I know we've partnered with Microsoft and launched the AI skills passport training as well, which has been really valuable in helping in some of that.
Let's dig a little bit deeper in the real-world impact and power of partnership.
Mark, you've had a front-row seat to SAP transformation on Azure and using Microsoft AI and other capabilities. Can you walk us through an example where Agentic AI is delivering measurable change?
Skoog
There's a couple that come to mind, but if I look at Microsoft, one of the things that we looked at was our forecasting, where we had hundreds of people actually building our forecast instead of doing innovative work on figuring out how to increase margin and other work. And in the end, we were pretty accurate. We were 3%, but it was a lot of effort to get there. Our CFO often said, I have to make adjustments to this.
Using Agentic, not only is that process much quicker now, but we can look at things like a pipeline. In our case, it's Dynamics CRM, but it could have been Salesforce or another tool. And looking at that to actuals, which are in this case for Microsoft and SAP.
And tying that together and have the Agentic be predictive about that. We bought our forecast accuracy, and now it's within 1%. So not only better results, but we're pulling people out of the tedious work to high-value work.
Not necessarily about offsetting labor, but how do you get people to do more with less and do more high-value work?
The other thing that's interesting is rather than training everyone in the back-end tools as this process, people are asking questions right in the tools they use every day, like Teams.
I rarely go into SAP or even our dynamic CRM tool without a question, I just ask it, and Agentic goes out as an assistant for me, right? I can even prompt it to bring information for me daily on my meetings and what I'm going to do.
And so a lot of value with that. But I think it's getting to those measurable results that becomes really key. So we think of everything as a process, and is that process adding value on what we're putting Agentic in.
Then to Pat's point, we don't try to solve world hunger, but we look at the data and governance and compliance aspects of that as well, and we're taking it a piece at a time as we work through there.
I think the other misconception is you have to really build all of this first to get the value. We think of it as, where do we get immediate value in this highly transformative world and build that piece at a time, thinking of the long term strategy in the process.
It's been great. EY has been part of a lot of our journey, even in our internal business processes and partnering, bringing that subject matter expertise.
Luquire
Mark, what role does Microsoft's cloud native architecture and the AI-powered capabilities play in making these results possible?
Skoog
I guess a couple of things. One is you don't have to worry about the infrastructure behind it and the technologies behind it. They'll come as part of the tool.
We took that same example Pat talked about with EY's assistance and did a similar shift on Microsoft in our close from 14 days down. Leveraging the technology meant we didn't have to think about acquiring all of that technology, but leveraging the investments we're making. We packaged that up as part of our core product. If you think a lot of this is the agents as a service that you can leverage, all the backends eventually are going to have AI in them and the backend ISV applications, is that orchestration that becomes really critical and important in that process.
Microsoft already has a huge user base, as you can imagine, in the productivity space of billion users. We have over 100 million users on copilot, right? That consistent experience, no matter where you sit and go, I think is really critical.
Then it just scales as part of the tools. I mean, it's part of every tool that I use today. Whether I'm in Word looking at a contract or helping edit a contract or whether I'm in Teams just collaborating with folks internal and external to Microsoft, collaborating with EY. It's in the native experience and extension.
Then just the enterprise scale. If you look at the data platform, SAP is critical data for Microsoft, 300, 400 terabytes of data. But in our enterprise data warehouse sitting on Azure, we have 17 exabytes.
We're not only thinking about the tools that we provide today, but as we move forward, how we get into more of the compliance, not just through the one location but the global compliance, and then the critical integration, of course, for services with EY to help understand how to implement that.
Luquire
Helping people get value out of, whether it be Power Platform and AI capabilities there to simplify the creation of it or the full commercial AI capabilities that live within Azure. It really makes a lot of sense, and Microsoft has a lot of capabilities in that space. Pat, you talked about a great example, but what aha moments have clients had when they see Agentic in action? I think sometimes we have to help them imagine the future, right? And especially across SAP and other systems.
Sullivan
Oh, gosh. Well, I mean, we had one, right? For two weeks to four days. That's an “aha” moment.
And I think if you put yourself in the shoes, say, a finance professional in an enterprise, that's what you were trained to do was finance. And what happens all too often now with technology and just the tech that we have is you're loading data, you're manipulating data, you're not really doing finance.
There are so many different use cases where the major thing what it comes down to is time and time saved. It's the one thing that we can't buy and we can't create, is time.
What AI is doing to essentially free up the ability for people to do the jobs that we originally hired them to do, I think that's the biggest “aha” moment, and that would be right now.
As we continue to mature, and Agentic is maturing so fast, as we continue to mature, we're going to start to see where we can put specific numbers on this. We saw a 90% improvement here, an 80% improvement there, cost reductions. We're starting to see some of those already, but this is only going to progress, and it's exciting.
Luquire
Yeah, and I think we're going to start to see use cases that were never possible before even, right?
And as part of that, as we look at enterprises, a lot of them have a lot of different enterprise systems, not all SAP, in a media running service now. Salesforce, Dynamics, or other things. How does Agentic AI connect all of this across the platforms?
Sullivan
What I love in terms of what we're doing with Microsoft on the alliance side here is, looking at Copilot is almost like an enterprise foundation.
So we can have all the agents from these different platforms connect in. But the so what of that is combining all of this together, orchestrating it and understanding how these data areas connect across the different systems.
So our ability to either use some copilot agents off the shelf or honestly, Mark, create them, which is what we've been doing, is I really like that angle because now you're truly taking an enterprise approach to it.
Mark, I know you've got some views on this, and you've dealt with this already with some clients, specifically with SAP, but then the rest of the enterprise.
Skoog
Yeah, I think one of the things that we look at is where the customer is at in their journey.
You don't have to upgrade today I think that's been key for us is to understand you can work on the systems the way they sit today. It was also important for us to understand even 8% of our processes run on SAP at Microsoft for the business processes. We had hundreds of other systems that it was integrating to.
[A lot of that is which system you're going to go into, which is going to be the master.] With Agentic, you don't have to worry about that. I think the thing that's often missed is you're talking about creating these high value agents from EY, they all plug in to Copilot. At the end of the day, it becomes the orchestrator and the consistency as this technology morphs.
Then the expertise from EY allows you to take in and not have to worry about, am I compliant in the region I'm running in for finance or procurement?
The other “aha” moment for me really was that just how different all of the integration can be. If you think about how things work today, I sat down with our purchasing group and found out, originally, we were taking contracts as they sit, and it have them on one display screen, then they were in a Ariba on another one, entering data from the contract in manually.
We took 35 human rout of that process, more into high-value work and negotiating better contracts and getting better insights through Agentic versus entering the data.
I think the last thing I'll just mention is it's critical for us to have this partnership. We bring a ton of great technology and evolution in that technology, but we don't bring the subject matter expertise in finance, procurement, and the other areas that you bring in thought leadership by industry at EY. That partnership, I think, allows us to do this transformation faster with customers and bring that expertise together.
Luquire
Let's spend a little bit of time talking about challenges, adoption and really what's next in this space. Mark, what would you say are really the biggest challenges when embedding Agentic AI into SAP ecosystems?
Skoog
There are two things that we look at.
One is leveraging the tools that are there already. We have a billion users on our identity platform. As SAP moves away from direct identity and integrating in, Entra is the preferred platform for that. So everything at Microsoft is fed starting from SuccessFactors, which is our HR tool, into our identity. And then that identity persists across all of our line of business systems for consistency. That's out-of-box capability.
Then the second is I think people are looking at where they have to morph the technology to integrate in or where they're going to build or buy integration. SAP is a great example. We have a whole series of engineers that have been working together for over a decade, integrating these products together.
Thinking about how you're going to build that, if you just take what's already been built and used at scale, the ability for us to even leverage our own product that way and not have to build it like we would have in the past is really fundamental in allowing us to move fast.
We're taking projects that were 12, 24 months, 36 months, and we're shrinking those down to weeks. And we see this with customers as well.
I think the first step for me is really to look at what's there and what you can leverage versus building. That's one of the key observations made.
Luquire
I like that. Pat, for companies just starting out, what does a smart, low-risk adoption roadmap really look like for them?
Sullivan
I think EY is a great example. We've partnered with Microsoft quite a bit on this, and it's what has allowed us to take our client zero experience and bring that out to the market to our clients.
What we started with Mark was back in 2023, our own version of a large language model internally within EY. We call that EYQ. What that has allowed us to do is, quite frankly, put AI in the hands of every single person within EY. That's built within an Azure container. It's a private instance of OpenAI and the technology side of that.
But what that's driven, just from an adoption perspective, is we've had all 400,000 of our employees over the last year and a half have used this system. We've had over 20 million visits. We've got close to 200 million prompts. We've generated over 600 billion tokens.
There's been some really impressive stuff. But just that entry has then allowed our people to understand the power of how AI can be used and how we can start to use that within the work that we do.
Taking that to the next level in what we've done in terms of partnering with Microsoft with the different copilot agents that we can now deploy and bring to the market, which has led us to what we call EY.ai, which is our platform, which now serves our clients. That's built on a Microsoft foundation. That's now powering the different whether it be consulting or tax or assurance and the different areas that EY serves our clients with across the globe, we're embedding AI into the work that we do for our clients so we can deliver faster with more quality and just give an overall better experience for our clients.
Arming our people with the knowledge of how to use AI and how to make it work is huge. Now, what Microsoft and EY can do for our clients, specifically with SAP, is I think we have a head start and an advantage because we took that original entry to drive adoption and get our people educated.
I'm really proud of that. I'm proud of the work that we've done.
Luquire
So, Mark, you've described in the past three layers of Agentic AI adoption, and this really aligns with Microsoft's frontier firm, from helping humans to the human in the loop inspection and really a full orchestration as well. Can you walk us through those?
Skoog
Absolutely. I think we started similar to EY, where the first was assistance for the humans, going on in your day-to-day job.
We used to have a lot of work where a lot of our managers and executives would ask questions of finance every day. And so we created Ask Finance to help out.
In this area, we were able to remove all the finance folks from the process. Over 99% of the requests handled from AI on the assistance of day to day.
I think that's phase one, is really how do you make better use of the knowledge you have, the policies, the procedures, even taking that then to step two, which is having the agents work on your behalf. Still humans operate and oversee it, but in those cases, we're doing automatic approvals of things. The agents are working with some human oversight around the logic behind that. Any high-value use cases, humans are still approving before cash transactions are happening. Then I think level three of that really is where it's interactive, humans and agents working together day to day. Now you have agents inspecting agents.
The beautiful thing about Copilot is it becomes the orchestrator of that. Now I can look at all these agents and let the agents interact with each other in a way that's more natural.
I don't have to worry as this technology evolves over time, I can just take more advantage of it. So pretty exciting time.
Luquire
Pat, how important is it that people and process transformation is there in terms of ensuring Agentic AI delivers full value?
Sullivan
Oh, Mark, it's massive, right?
Look, in our business at EY, essentially, our product is our people. We're a people business in terms of how we serve our clients every single day.
Now with AI, it's a true collaboration.
Today, EY has 400,000 employees. As we continue to grow, I think we'll have 400,000 employees, and we'll have 400,000 agents that we deliver to the market with our work, and it's going to be a true collaboration on that.
In essence, it's like we're staffing AI agents as part of our program. But another part of this, just in terms of the true collaboration, again, I'll come back to, if we can't do it to ourselves at EY, it's tough for us to go tell clients what they should go do. And one of them is just we started this project where we started using GitHub Copilot for the systems that we developed that we use to deliver work, to do modernization of code or put in new features, enhancements, continue to add to our products and things like that.
So we still have to approve that as humans, and we still have to do the quality assurance over the work that AI is doing. I think that's going to be a very, very critical role as we continue to move forward.
But if you're not updating how people's jobs and roles are going to be performed in collaboration with AI, you're going to have a long journey. Your ROI is not going to be what you hope it's going to be.
It's really important for that human in a loop, but more importantly, to empower our people as we go through it.
Luquire
Yeah, I think that's a really important point, especially you know, we work with Microsoft. We have the pleasure of being able to partner with Microsoft, but also we're a customer of Microsoft, and Microsoft is a customer of ours as well. It opens up a lot of possibilities.
For both of you, as you look ahead, what trends or technologies and enterprise integration and Agentic AI that we haven't already talked about should leaders really prepare for now?
Sullivan
I'll go ahead and kick this one-off, Mark.
I think the most important point for leaders is a mindset of curiosity. That curiosity has got to lead to action. Because there's going to be a lot of disruption. There already has been, and it's going to continue to work. But if you have a curious mindset and you're trying to find ways in which AI can improve different elements within your organization, I think you've already won half the battle. I really do.
And with our clients, I have this curiosity between what Mark and I get to do with our clients, it's really magical, the type of things we can do.
So I would say stay curious, be curious, and always look for ways on how potentially AI can work to improve how the business runs.
Skoog
Great insight, Pat.
I think the curiosity in building an organization of curious people. I think too often we sit back into our patterns, and it's got to be top-down led, like you've done at EY or we've done at Microsoft.
That top-down curiosity of, just because we did it that way in the past, are there brighter ways to do it?
Funny, it goes back to not something new, but something old. People, process and technology. In this case, I think the technology is going to stay ahead of where the people and process are. But to think differently on that, enable the people in this curious mindset.
I think the smart workers are going to determine that the faster they adopt this, the more secure they're going to be going forward, right? Because I think that's to be key to that. If you create that culture and you enforce that culture with the right behaviors, I think it's critical.
Then once you're there, I think the whole organization changes. People have to be open to that change. At Microsoft were truly using our copilot technologies. I would go into Teams and that's my AI, and that's my interface to all the applications we use in the enterprise, primarily. I no longer go into the back-end tools.
That's a fundamental shift. But the usability of that is I don't have to worry, is this in SAP, or is this in Dynamics CRM or ServiceNow? It doesn't matter, right? And so that curiosity gets us to think differently.
Luquire
I think that's right. I think the pace of change is going to drive this continuous learning in a way that we haven't really seen before as well.
So one last final question for both of you. What's one piece of advice for leaders on getting the most from Agentic AI and SAP integration? Something that we haven't talked about just yet.
Sullivan
I love this question, Mark.
I used to start presentations years ago saying, who's used AI today? And at the time, you get a quarter of the room or what have you. The good news is that people were using it five, six years ago. And now when I ask that, you get most of the hands up. And for whoever doesn't raise their hand, I like to tell them, hey, if you're not using AI, it's been using you all day long. But my point on that is we can't be afraid of it. It's here. It's only going to get better, and it's really exciting.
And there's this mantra that I hear often, especially in my line of work, where people say, you got to fail fast. Sure, don't be afraid of failure, but let's succeed fast. We can vastly shrink the amount of effort that goes into different programs to get to the end result.
Then furthermore, that also drives this inclusiveness across our teams to want to achieve something, and everyone likes to be part of something. I think we can do it at an incredible pace, but we can really be part of a movement that is going to just write the future for decades. It's a really cool time.
Skoog
Pretty amazing. At Microsoft, we let the business innovate now at their pace, which is a fundamental shift. We put all the guardrails and boundaries on from IT and the enterprise guidance and architecture, but we're no longer a blocker for the business.
I think the other one is we're thinking of things in individual problem statements now as we move through this process and we build these problem statements out to the results.
Where that's an advantage is if you look at EY's capability of tying that business to technology and your ability to help guide our customers, we're finding this every day that it allows us to have a tighter link between the business and technology to solve these problems quicker.
That time to market has just been great. We had a large manufacturing customer that was just looking at exceptions on billing, where they had people in that process. In 12 weeks in a production scenario, they were able to avoid mistakes in the billing and rectify those billing entries quicker. They're going to have tens of millions of dollars saved year one in that one scenario that's a 12-week implementation, and that's against their current SAP system.
They have a vision to get to Rise. They'll get even more great capability from SAP on Rise, but it's something they could do today with EY and Microsoft.
Luquire
One of the things I always talk to leaders about is think differently.
It's a real opportunity to think and act differently than you've ever been able to before. You have to help people to be able to do that.
I was working with a team that was trying to figure out how to leverage AI to deliver capabilities, but they weren't using AI to help them figure out how to leverage AI.
I think just think differently and the doors really open up.
Well, Mark and Pat, thank you so much for your time today and for sharing your insights.
Sullivan
Yeah, likewise. Thanks, Mark.
Skoog
Thanks for having us.
Luquire
Today's conversation shows that Agentic AI is more than a buzzword.
It's a capability that when paired with the right strategy, technology and partners, can help businesses move faster, think smarter, and really stay ahead.
If you'd like to learn more about how EY and Microsoft are working together to unlock value from SAP and other enterprise systems, visit the link here that we'll share in the podcast.
Thanks for joining us today, and we'll see you next time on the EY and Microsoft Tech Directions podcast.
A reminder, the views of third parties set out in this podcast are not necessarily the views of the global EY organization nor its member firms. Moreover, they should be seen in the context of the time in which they were made.
For more insights from EY and Microsoft leaders, visit us on ey.com/techdirections.
