How Microsoft Is Rewriting Enterprise AI Adoption | Erin Chapple, Microsoft CVP

Samuel:
Erin, I’m really excited to have you on the show. Welcome.

Erin Chapple:
It’s exciting to be here. Thank you for inviting me.

Samuel:
My pleasure, and thank you for accepting the invitation. Erin, you started at Microsoft 30 years ago as an intern. So first, that’s really impressive — congratulations. You spent almost all of your career in the product engineering group. You played a pivotal role in Windows Server and then Azure, moving the whole world to the cloud. And one year ago, you made a major career shift and moved into our Commercial Solution Areas, which is basically in sales. So what inspired you to move from engineering into the sales organization? And so far, after one year, what has surprised you? What have you learned from this transition?

Erin Chapple:
Well, I wish, Samuel, I could say that I had the inspiration, that it was part of a master career plan that I had, but that’s not the case. I think I had spent my entire career here in the product engineering organization, and I think in my heart, I’m all about solving problems and I’m all about really ensuring that we’re building products that meet customer needs. That’s where I get my energy from. I love spending time with customers and I love understanding what the challenges are they’re facing from a business perspective and what can we do, right, at Microsoft in order to provide either infrastructure platforms or solutions to really meet their needs.

So I have to say, I was really happy. I was leading the Azure Infrastructure Product Design and PM organization, and I was enjoying it. I feel like I had the best job on the planet in many ways. And so I will say I was surprised in some sense when I got a knock at the virtual door from Judson, who leads MCAPS — our sales organization, our Microsoft Customer and Partner Solution team — to explore what I wanted to do next and whether or not I would be interested in coming into MCAPS. And first I was like, “You have the right person?” you know, in that sense.

But through the conversation, really, it makes a lot of sense, right? If you think about where we are in just the evolution of technology. You know, I think that historically there was much more separation between what was done in product engineering and what was done in sales and in the field and customer-facing roles, because the timeline for delivery was just much longer, right? I come from a world of operating systems that not too far ago were three-year release cycles. And so there was just much more separation.

And the product management role has always been that voice of the customer within the product engineering organization. And given the speed at which we’re just seeing technology change with AI and the transformation that that’s bringing, we’re creating the technology and the product at the same time that customers are adopting it and evolving and understanding how it meets their business needs. And so that loop of feedback between product engineering and the customer has gotten so much — or needs to get so much — smaller. It has gotten, and I think we need to continue to push on that.

So through the conversation with Judson, I thought, “Wow, this makes a lot of sense.” It gave me the opportunity to scale from having always worked in the infrastructure space to really be able to work across the breadth of the Microsoft Cloud and to represent that connection between the sales organization, the customer, and product engineering — to make sure that we were connecting that information and really making Microsoft seem smaller in terms of the information that we have to build great products that really drive customer success. So it was not something I was expecting, but it made a lot of sense for me in terms of continuing to have the impact that I care about for Microsoft and for customers.

Samuel:
I love that you mentioned making the portfolio look smaller, because over time — you’ve been 30 years at Microsoft — the portfolio has improved tremendously.

Erin Chapple:
It has. It has.

It’s one thing I really think — scaling into this role would have been challenging, or more challenging for me, if I hadn’t spent my time at the infrastructure layer. I am, at heart, a platform scaler. I really believe in the power of building those platforms that help enable scale in that sense. And so that gave me a really interesting vantage point, I would say.

But you asked a little bit like what have I learned and what’s been maybe some of the surprises in that sense. It is amazing what customers are doing — right across industry, across geography, across size and scale of organization — with technology today. And I think, you know, in having a focus at the infrastructure layer on what did we need to do to build the right platform, this new role that I’m in has given me a much broader perspective on just the nature of the relationship of technology in an organization and how it’s really driving business transformation — through AI transformation, but also the role that the partnership between customer and technology provider needs to play.

I really do feel like it is a partnership that we are jointly navigating. And that gives a different perspective, I think, on how you think about engagement, how you think about designing the product, how you think about delivery — not just with an end customer, like a large enterprise, but also with the partner ecosystem and ISV ecosystem, our software development companies out there, the service and solution integrators. It just gives you a different vantage point, I think, that helps me to have more breadth or sort of understand the dimensions of which we need to think about the end-to-end product delivery — from everything from the design through to how do we package, price, how do we go to market with it, what’s the feedback loop, how do we put that into place?

And so that’s been really important for me in that sense, and I think it makes me a better product leader. Because at the end of it, I think even though I’m in — I would say technically I’m in sales — at the end of the day, I think my whole job is about: are we building the right products for customers that delight them? And that breadth and that increase of perspective I think really helps to ensure that that happens.

Samuel:
So what I’m hearing is that the wall between sales and products is blurring and just separating.

Erin Chapple:
Totally.

Samuel:
And you mentioned simplifying the whole portfolio, and actually with the new fiscal year at Microsoft, the different solution areas have been simplified. Previously, there were Business Applications — CRM, Power Platform, ERP — and Modern Work, which encompassed M365, Teams, SharePoint, etc. That’s merged into one simplified AI Business Solution. Azure Core, which was previously three different solution areas with Data and AI, Infrastructure, and Data Innovation, has been simplified into Cloud and AI Platform. And Security has remained as one of the solution areas. I know it’s a key pillar for us.

What kind of impact will this simplified solution have on our customers and our partners? Because I’m asked a lot by — specifically — our partners, I’ll say.

Erin Chapple:
Yeah, look, I think you have to first step back and sort of understand what are we trying to do with the Commercial Solution Areas, or the CSAs. As I said earlier, customers don’t want to come and talk to Microsoft about a technology. They want to come and talk about a business problem that they’re facing and how can technology and the partnership help solve that business problem — help add value into the organization.

And so the Commercial Solution Areas have always been the way in which we represent the different products and capabilities to the customer in the conversation, because you want to have a conversation with the customer on their terms. What is it that they’re really hoping to achieve? And so the previous solution areas I think served us well. They served us well as an organizing function and a communication function to really illustrate the value that the Microsoft Cloud provides.

Now, in the world of AI, that conversation is changing, right? It used to be — prior to generative AI and kind of Copilot and all of these components — that you would have a more discrete conversation around Modern Work and what is it that I want to do to really modernize the employee experience within the organization. And that was, in some ways, a more standalone conversation. But AI is really blurring the lines, in some sense, of the capabilities and how that needs to show up in the context of the organization in order to really derive the business process and experience.

And so really, the simplification of the CSAs was meant to bring together the capabilities and the value that the Microsoft Cloud provides in a way that resonates with the conversations that we have with customers. And they really do now center on Security as the foundation — you don’t have an AI transformation without great security. You don’t have anything without great security. And then the Cloud and AI Platform — that fundamental set of building blocks that enables organizations to build applications and build AI applications that are really going to bring innovation into their space. And then the entire AI Business Solutions that really help with the employee experience, with the customer experience, and that fundamental way in which the business is run.

And so that much more logically maps to the nature of the conversation. So it allows us as Microsoft to show up better — in a way with the knowledge and capabilities — in order to be able to have those conversations directly with customers.

Samuel:
And it will also simplify, I suppose, the way we go to market, the way our partners have access — and customers have access — to programs to help empower them in deploying AI quickly. Because everybody wants to deploy quickly, but securely.

Erin Chapple:
You know, it’s not unlike, you know, when I came into the Azure infrastructure team many years ago. I came in and I was leading the Azure Compute organization. And you need to have the capabilities. You need to have compute. You need to have networking. You need to have storage. The way the customer purchases Azure, or what they want to do with Azure is—I have to talk about it—they don’t come and say, “Well, I want to purchase a bunch of block storage because I want to have it kind of lying around for a rainy day.” No, they come with a workload.

And so, I was very much a part of, within the Azure Core organization, really driving this workload focus and starting to invert that, because that’s the way in which we need to show up with customers, right? You want to show up in a way that says, “Here’s how compute and network and storage come together to deliver the best experience to host SAP,” or “Here’s how to deliver the best experience to drive whatever your application is that you’re building.”

And I think that’s sort of similar in this sense. We don’t want to show up and need to have 50 people in the room to have a conversation about building a new AI application. We want to be able to come in and really have that integrated view of how these components work together—both services that we provide, but also then, as you said, services that are delivered through the partner ecosystem, the marketplace, all of the third-party providers that build on top of Microsoft Cloud.

And we want to be able to have that conversation in a way that is actually coherent and doesn’t look like we’re showing up as, you know, 100 different teams. No, we’re showing up as one organization, one partner, one trusted advisor that can help the customer through that journey.

Samuel:
I’ve seen that on the field as well, where in the past we were, like you said, 50 people at the table trying to explain the whole technological landscape or offering from Microsoft. And nowadays there’s less people, more specialized, but we can really tackle the specific problem and business challenge from our customers. So I totally agree. Building on that—for organizations that want to stay ahead of the curve, the innovation curve that is technically a wall right now because it’s going so fast—how should they think about all those new solution areas?

Erin Chapple:
So in the last year in this role, I’ve talked to hundreds of customers. You know, one of the things I love to do is to spend the first 15–20 minutes of a conversation listening. What is it that you are trying to accomplish? What are you wrestling with from a business standpoint?

And really, the conversations can all map back into four key questions or four key goals or intentions, which I think are the foundation of what does it mean to become a frontier firm. A frontier firm is: how do I lead with AI in my thinking in that sense and infuse it, right? And move from it being sort of a tool to being an integral part of how you think about your overall business strategy.

But those four main conversations or main areas are:
First, how can AI really help enrich the employee experience? How can I think about my workforce and what their experience is—the productivity of what their role is? How can I help improve their satisfaction with their job by making the things that are more manual and repeatable easier to do? How can I give them access to a broader set of information? So I think that first is really anchoring on that employee experience within an organization.

The second conversation really comes into how are you reinventing customer engagement? And we saw with the adoption of the cloud, we saw the cloud change a lot of the capabilities that we had in terms of how do I deliver the experience to my customer, whether or not that’s through the application I’m building. In many ways, that was when we started to see that most companies became technology companies, because there was some way in which they were leveraging technology to really transform how they interact with the customer base that they have.

AI is taking that to the next level. Like, how can I leverage this in the customer experience that they have with my company?

The third conversation centers around business process. And this is a place that I think is just prime for AI to have an impact on what is done in an organization—from the automation of business processes through agents, through—

And then the evolution of that to, you know, in the future we see agents really becoming members of the team in order to be alongside the organization to really transform the way in which the business gets done within the organization.

And then the fourth conversation is just: how do I bend the curve on innovation? You talked about that wall in some sense. How can I leverage that in the context of whatever industry I’m in? So I’m in the pharmaceutical industry and I want to apply it to drug trials and how do I accelerate that? I’m in manufacturing and I want to think about it in the context of supply chain. But I think all of those pieces become really important.

So those four conversations—employee experience, customer engagement, business process, innovation—really are the foundation of which these solution areas work against or work to help realize what an organization can do to realize that value.

And so the AI Business Solutions conversation is really about how are we making it easier for customers to adopt AI across business process and workforce, right? Because we see those two being the core backbone in many ways of how the work gets done in an organization. So whether it’s Copilot or agents or a new AI experience in M365 or D365, bringing these together really gives that unified, outcome-driven experience so that it’s faster, cleaner, and really more aligned to their transformation goals.

So that sort of helps with that employee experience, business process, customer engagement.

Now on the Cloud and AI Platform side, we brought together all of the components of the Azure infrastructure. So this is almost, in some sense, an extension of that workload concept I talked about—around compute and network and storage coming together. Now infrastructure, data, AI, the developer experience, apps, PaaS services coming together—that really is the backbone of what helps customers move from on-premises to the cloud, but also then thinking about modernization and bringing AI innovation into the organization.

So that means fewer handoffs, there’s deeper technical engagement because we have the right people involved. And really, hopefully, that means faster time to value for the customer—especially with new offerings that we’re providing, but also ways in which the customer can pull these components together.

Hopefully that gives you the sense of, across those four conversations, how the solution areas really map in to drive that conversation and make it real for customers in a way that helps them to realize that value.

And I mean, can’t end this without saying: security, security, security. It’s just foundational. It is our number one priority. While it is sort of a standalone Commercial Solution Area for us, it’s also embedded by design across all of the solution areas so that we can ensure—whether it’s how are you securing and governing your data estate in order to be able to provide the right platform for all of that, the Copilot and Dynamics 365 experiences—so that you can really innovate with confidence in your business process, through to how it is core to the platform on the Cloud and AI Platform side to ensure that you can actually build that AI-native application with the confidence around governance, data, security—all of those pieces—that you’re protected from the start.

Samuel:
So: empowered workforce, customer engagement, business process, and bending the innovation curve. I’m really interested in this last one. What does that mean—to bend the innovation curve?

Erin Chapple:
Look, in every business, doesn’t matter what industry you’re in—manufacturing, retail—there’s a way in which you want to innovate in the context of your current business. For manufacturing, that could be: how do I actually improve supply chain? In drug trials, it could be: how do I actually identify, right, what should I be taking forward? In retail, it could be: how am I thinking about mapping, in some sense, what’s happening in market and trend and applying it to design and how I think about marketing of the retail store and whatnot.

But bending the curve on innovation is: how do I actually leverage AI in the context of the core business to bring that change faster into the environment? How do I leverage it to identify the preferred solution or solutions that might be more likely to succeed? How do I apply it?

All of the things that you think about doing today that are more manual in that sense, or that require a lot of research and discovery—how am I leveraging the AI tools in the environment that we have in order to really be able to move faster in that sense?

And so we see, right, in terms of organizations building applications that are AI-native applications in order to be able to leverage the information they have, the data that they have that’s specific to their business, in order to speed up the ability to actually take solutions and bring solutions into market.

Samuel:
My latest podcast episode was with the VP of Aldo. A big retailer here in Canada was explaining how it totally changed the way they operate. It’s really using AI to be able to innovate faster in your specific area.

Erin Chapple:
Yeah, both—like in that—probably both from an internal process standpoint, but also then in the context of whatever the domain is, right, that they are working.

Samuel:
Speaking of innovation, I think we all know this is moving faster than ever. And it’s really hard to predict what’s coming over the next 12 months. Five years, 10 years—it’s even worse. So how does Microsoft stay ahead of this curve? And how does Microsoft help their customers do the same?

Erin Chapple:
Look, I think we have a long history at Microsoft of being customer zero. Right? Of leveraging our own products and being the first. I remember—I often joke—you talked about, I’m sort of at the 30-year mark from my first internship. My first internship at Microsoft, I was working on an alpha version of Outlook—what became Outlook—running on top of a beta version of what became Exchange. And I never got mail, right? Because I was—I used to tell all of my intern friends, like, call me if you want to have lunch because I’m not going to get your mail, because we were running the product that we were about to ship, in that sense.

And so I think in order to really understand and predict what’s coming and be in a position to be able to speak from a place of understanding, we need to be, at Microsoft, customer zero across all of the products and services. So that means we’re not only developing new technologies, but we’re adopting them internally before anyone else.

So, for example, we rolled out Copilot to every employee in the company. And having that unique approach—to really have the whole breadth of Microsoft to test, to learn, to identify, fix potential issues before customers even experience them—puts us in a place where we can help to chart that path. And it doesn’t mean we’ll get it perfect, but it does mean that we’ll uncover a lot of the challenges.

And so I think that gives us a huge advantage as we innovate at the speed that you see the market developing and to be able to support customer deployments at scale. You think about what we see internally as driving utilization of Copilot. It’s about what is that immersive experience for the end user? How can it, firsthand, you know, improve their daily productivity?

Well, a lot of the tips and approaches that we’ve been able to help them sort of scale, right, through our go-to-market and our skilling and whatnot, and engagement with customers on things like promptathons and whatnot to really help change the way in which the workforce understands what’s accessible to them, came from our internal adoption and what we learned.

I know within my team, within the MCAPS organization, we run a Copilot Cup and a challenge over the course of the year to really drive utilization—and drive not just utilization, but you’re driving the utilization to drive the learning so that we understand what’s happening.

And so for any organization out there, my biggest encouragement is to just start. To just get it into the environment, because you don’t know how it’s going to help or what the impact is, or how your users or workforce are going to react to it, or what you might need to change, right, from a culture standpoint, from a governance standpoint, in order to drive that.

And so for me, I really think that what came out of the best practices to really scale Copilot adoption came from our internal awareness and understanding of using that. That helps us to bring that to customers in a broad sense much more quickly.

Samuel:
And we’re a big organization. I mean, we have a lot of employees. So I think we’re a good benchmark. And I experienced it as customer zero myself. And I’ve been through all the challenges that some of our customers might see. But since we’re close to the product team, it’s easier to try and adapt along the way. So it’s truly interesting.

Erin Chapple:
And look, we pivot based—early on, when we were just—the whole idea, in some sense, of generative AI and Copilot and this concept of chat was newer. We started to talk to both customers but also internally and see what are the AI patterns that are really driving the usage and driving the interest.

And early on we saw two really pop, right, in terms of the engagement. One was building your own AI assistant, and the other was chat with your data. And those were sort of the two AI patterns that were driving a lot of interest, as well as movement into production from the customer.

And so what did we learn from that? Well, we learned that instead of having everyone have to use these patterns and implement them in their environment, we were missing, right, in the strategy of M365 Copilot, this concept of Copilot Chat—that was this broadly accessible availability for users or customers to be able to quickly chat with their own data across a broad set.

And so we made that decision then, from a product strategy standpoint, to launch Copilot Chat. And so here is a way in which you’re seeing internal usage, understanding, and identification of patterns really helping us to understand what’s the next thing that we need to build.

And it’s that feedback loop, in some sense, from the sales organization and seeing what is actually resonating with customers when they’re adopting—back into product engineering—that really helps that evolution happen. To your point, right, when you don’t know where you’re going, you need to be able to be curious, learn, and react quickly and be able to bring those things into market.

And so I think really ensuring that we are staying engaged, talking to customers, meeting with them, learning firsthand what’s happening, and then taking that feedback back into the product engineering and design cycle is really important.

Samuel:
So this loop is applied internally as customer zero and with our customers as well. I mean, I’ve noticed as someone on the field that the time between a customer mentioning something about our product and the time we bring that to production is really faster than ever. What’s driving that? This new level of agility at Microsoft? And actually, can you expand on this feedback loop?

Erin Chapple:
Yeah, but it was one of the things—you know, I have to, like, I’ll admit—when I made the shift from product engineering to MCAPS and into the Commercial Solution Areas, I thought I knew how the field operated. I thought I understood the mechanisms. And I knew, like, this much of the entire process.

And so I think we were just missing, in some sense, that connection point, right? The connection point happened, I would say, more through escalation than it did through actually bringing the end-to-end set of teams together in this way.

And in many ways, the Commercial Solution Areas team was focused on supporting the field and connecting that back—and not, in some sense, putting equal weight on also a role supporting the engineering team and business planning.

And so part of it was just articulating the role of the organization and the role of what we call the loop—or this feedback loop—that visually I see as an infinity symbol that really is connecting all parts of the cycle that come together.

And so my sort of nerdy or geeky analogy is that the Commercial Solution Areas team is the corpus callosum of Microsoft. And if you know, the corpus callosum is the part that connects the left and the right hemispheres of the brain. And in order to perform most functions that the human body does, those two parts of the brain need to actually talk together.

If that corpus callosum is severed, you can still function, but it’s harder, because each sort of wants to do their independent motion. But if you can bring them together—and that’s always been one of our core values, this One Microsoft and the way in which we bring—but I think that loop is really what helps set that.

And now, that means we need to have product engineering more deeply engaged with field. It means we have to have field more deeply engaged with product engineering. Now, you can’t scale that to make sure that every person is connected to every person in the organization, but how do you put the systems in place?

And so we do have, as part of the Commercial Solution Area organization, we do have that regular connection point with the field where we pull that information together. And because we in the Commercial Solution Area look across all solution areas and look across all geographies, we’re able to get that signal and realize where are the patterns.

And in many ways, building product is a prioritization effort. You really want to focus on solving the things that have the most leverage in terms of helping the customer get through what they need to do.

And so I think that process of being able to connect that has helped us to both identify what the most important issues are, identify them quickly, and then by bringing the teams together earlier into that cycle—where product engineering can hear more directly from the field—we speed up the time it takes to just—you know, look, we’re in a large organization. That brings benefit with it, but it also brings challenges in terms of how do you make the organization act like a small startup and the agility that can come with that.

And so I think it’s this combination of identifying what needs to be done, having the right systems and processes in place, and then creating the right connections across the organization.

Samuel:
I really love this new way of operating, and I’m pretty sure our partners and our customers like it as well because we’re fixing things so much faster than we did in the past. Switching gears a bit—we’ve mentioned Frontier Firm a couple of times. What defines a Frontier Firm? I mean, it’s part of pretty much all of our presentations—the ones I’m doing—and I’m asked a lot: what does it mean exactly? What does the journey look like for an organization that wants to become a Frontier Firm?

Erin Chapple:
Look, simply put, I think that the Frontier Firm is about infusing AI in how you fundamentally think. And in many ways, it’s not different than the transition that we made from sort of on-premises into cloud, right? Where there was sort of a shift in how you thought about the value of cloud and how that actually should be infused in your business strategy and how you’re thinking about bringing applications and experiences to market.

Like if you can remember, I don’t know, 10, 15 years ago, just the way in which you thought about, well, the capabilities of the cloud enable you to have just a different way of thinking about how you actually approach application development or innovation. And so the same is true in the context of AI.

And so part of being a Frontier Firm is how do you infuse this in everything that you are doing? I think today we’re in a place where we’re sort of applying AI to the problems we know about today as opposed to thinking about how AI can help us to reimagine the way in which we could be doing things.

And so I think about the difference of sort of today’s first-stage application of AI. It’s more about how to use AI to speed up tasks. But if we think about the vision of where a Frontier Firm would be, a Frontier Firm would think about how do I build hybrid teams? How do I have humans with AI agents that are working side by side? And in some cases, maybe even AI agents that are leading specific workflows with human assist and the ability to provide that strategic oversight.

And so I think it is a journey. It’s a journey from AI as an assistant, helping an individual to be more productive, to AI as a digital colleague—how agents join a team and do tasks. You see that today in something like Microsoft Teams with the facilitator role and whatnot, where they’re actually joining and taking on the task. And you see that in GitHub with some of the developer experience.

To how do I get to a place where I have AI-led teams, right? Where those agents are operating autonomously—not completely autonomously, right—guided by human strategy. But how can you actually distribute that work in that sense?

And so, you know, we’re really using tools like Copilot to boost productivity, but we’re launching AI Centers of Excellence that actually help organizations to assess that maturity and scale. And so when we talk about Frontier Firm, we’re talking about this bold new way of work where AI isn’t just a tool, it’s that teammate, and it is infused in how you think about getting work done.

Samuel:
So summarizing, becoming an AI Frontier Firm means embracing AI in all your processes—from individual level to team level to even management level. And Copilot is probably the easiest way to start using AI. When we first released Copilot, I think it took the world by storm. It’s not that long ago when you think about it. I mean, GPT-3.5 was released in November 2022.

Erin Chapple:
Mm-hmm. Mm-hmm.

Samuel:
And then in 2023, we launched Copilot. So not that long ago. But one thing it does—it improves productivity, obviously—but it also exposed years of neglected governance in many organizations. How can companies address these challenges without slowing down their AI adoption? Because they don’t want to pause, they don’t want to stop, but still, governance is really important.

Erin Chapple:
Yeah, it’s interesting. There’s sort of two things I would say that I often try to present in somewhat of a humorous light, but it’s what we find in many organizations is when they enable tools like Copilot, they find data they never knew existed. Right? It just—in that sense.

And I often get asked, what is sort of the number one reason why an AI project fails? And I often say it fails because of data, right? And poor data strategy and data governance. And so what you’re saying is a real issue.

You know, I think that over time, having just worked in the infrastructure space and IT space, everyone knew they should be doing something with their data—it never hit the highest thing on the priority list. But just the confluence of AI and security and just the different dimensions of what’s existing, it’s really critical, I think, at this point.

Microsoft provides a comprehensive AI governance guidance to really help customers be able to responsibly adopt, manage, and scale those AI technologies. And that really is rooted in sort of a three-dimensional view—or, you know, 360 view—of ethical principles, regulatory compliance, and then what do we see as operational best practices.

So if I think about the stages of that, first it’s about establishing a robust AI organizational setup—an operating model. You know, when I talk to customers and partners—I was sort of on the road last year and doing a conversation about adopting AI—and the questions weren’t about tech. They were about the culture and the operating model and the governance within an organization of how you think about operating your AI strategy.

So that’s the first thing: how do you create that setup and operating model? And then I think there’s that governance model that needs to overlay on top to be able to take the siloed teams you have—because data is oftentimes a siloed problem—in that sense, and really ensure that you’re bringing together the siloed teams and those responsibilities so that you are implementing a set of structured governance stages or approach.

And then being able to really streamline AI use case intake so that you can get some—I hesitate to use the word control, but maybe visibility in some sense—and then prioritization around where you’re investing.

Because the other thing we see is oftentimes these pilot projects pop up all over the place, and it’s hard to understand, like, where are you really? And that’s where the governance comes into play—how can you actually understand what’s being utilized across the organization and then ensure that you’re making the right investments so that they align with business value and that you can drive impact and measure what’s happening?

And so I think some questions that customers should be asking themselves are: Do we have centralized Copilot and agent governance guardrails that are in place across all of the company to ensure we’ve got compliance? Do we have a data loss prevention strategy? Are we continuously testing solutions to make sure that we’re not oversharing, that we’re not actually slipping into non-compliance?

And so what we see is, as organizations are transitioning to a more strategic AI, really aligning the investment across the applications and platforms that they have—and the data—to put together that strategy as a whole really puts some of that governance and security in place so that they can actually scale from a growth step.

Samuel:
So it’s really three different layers: operating model—AI being part of the setup of the whole operating model—a whole governance layer, and finally, being able to prioritize all the AI use cases so you don’t end up with use cases or proof of concepts everywhere that might put you at risk.

Following on that—because governance is one part of the story—I think the other part of the story is security, right? And with generative AI come increased risks for different reasons, but especially around data security and compliance.

How can organizations embrace AI without putting themselves or their customers at risk—and again, without falling behind their competitors? Because same as with governance, people don’t want to stop or want to pause. They want to go with the flow and start implementing AI in the organization.

Erin Chapple:
Yeah, and this is—if you talk to CIOs, CTOs, CISOs, CEOs—this is a top concern. The majority of them, overwhelmingly, are worried about oversharing and leakage of data. They’re also worried about new threats coming from AI. They’re worried about regulatory compliance.

And so this is a challenging role to operate in. For example, if you work in financial services or healthcare or any regulated industry or government, there are concerns over the regulatory capabilities that any solution may have—or the regulatory adherence, right—that that solution, and implementing that, and how that might impede your ability to actually deliver solutions in market.

So we have, in Microsoft, a Cloud Adoption Framework for AI governance, and that’s really meant to provide structured guidance so that organizations can really take a look and assess their overall AI risks. They can put into place and document governance policies, and then provide the enforcement and monitoring compliance around that.

So that organizations can adopt AI securely by implementing that multi-layered governance model. If you just step back, visibility is really the key, I think, in this. You can’t protect what you can’t see, right?

And so tools like Microsoft Purview’s Data Security Posture Management for AI, Defender for Cloud Apps—really help organizations to monitor those AI interactions, to flag issues that they’re seeing, and then apply that data loss prevention policy centrally so that they can be more confident in their ability to adopt the technology.

Now, the other thing I’ll call out is that governance shouldn’t be seen as a bottleneck, right? It should be seen as more of a launchpad. And so before deploying AI, organizations should clean up that obsolete data. They should apply the sensitivity labels, right? All these things we should have been doing in the past—making sure we’re restricting access to high-risk repositories.

So that isn’t just good hygiene, it’s a strategic defense. And if you can do that in that place, it becomes this accelerator in many ways. And so I think that reframe of bottleneck to launchpad—and what can we do—is really important.

And then lastly, I think we need to think about how both we at Microsoft, but also organizations, are creating responsible AI frameworks. So if you think about just our internal development process—within Microsoft, teams go through these rigorous harm assessments and executive reviews before we launch AI features.

And that’s not just about ethics, it’s about maintaining trust. By actually having a risk team who is looking at this assessment, that can actually accelerate AI adoption because it means that you’ve got clear scope, you’ve got the right governance in place, you’ve got safe deployment—and these things become enablers, not blockers.

And so those who are most successful in this space will be the ones who architect for scale, because that reinforces trust, right? And they think about not just this world of AI, but the collaboration that needs to happen in order to become that Frontier Firm that we talked about.

 

Samuel:
I love this concept of governance as a launchpad instead of a bottleneck. I think it’s really important. People don’t want to tackle the governance, but they have to.

Erin Chapple:
And it’s interesting, there are parallels in some sense between the technology world, I think, and the human world in the sense of—oftentimes there’s this, usually quoted as an African proverb that says, “If you want to go fast, go alone, but if you want to go farther, go together.”

And in some sense, the governance is that piece of it, right? Like, how are you bringing together the different dimensions of your organization so that you’re actually working together and you have the right foundation in order to continue? And yes, if you just put things into production and you don’t think about them, and you don’t think about all of the strategy pieces, you’re gonna spend a lot of time afterwards cleaning that up. You might get something out faster, but is it gonna actually have the durability, the sustainability, the scalability—all of those pieces that are really important?

And that are really important, particularly in this day and age when the trust—right, trust with the employee workforce, trust with the customer—all of those things become really, really, really key.

Samuel:
Love it. Erin, we’re almost at the end of our time. But to wrap things up, I always like to ask two quick questions. One practical tip for our listeners that they can use right away, and one vision you have for the future and where it’s all heading. So first question: what’s your number one productivity tip when it comes to using AI in your own work?

Erin Chapple:
Look, you just have to start.

I think that people sometimes don’t know what to put into a prompt. They don’t know where to start. But I would say that you know how you work. And if you can just start using AI as a way and ask it the questions you’re asking yourself—right?—that starts you on this journey where then you start to figure out, well, wait a second. Like, I asked it this. I could ask it that.

Like, I started off just, you know, coming back from vacation or coming back from a long weekend—or even a weekend—and saying, “Look, please, Copilot, look across Teams and Outlook and give me the messages that have been directed to me that are most important, that I need to deal with,” right?

Or I use it when I’m going into a one-on-one to prepare with someone and say, like, “Hey, what were the last five mails that Samuel sent me? And have I answered all of them?” Like, some of those things that are just things that you would normally check as part of the way you do your work—right?—give those to Copilot because that’s going to help you.

And then you’re going to realize, well, wait a second. I can add this to that. I can refine. I could—and that, I think, gets you on that path where I think it has shifted from being something I have to think about doing to something that I more naturally have Copilot there, and I’m sort of asking it questions while I’m in a meeting because someone says something and I want to research the piece or I want to capture something or say, “Give me a to-do,” all of that, right?

So you just sort of—it takes part of that day-to-day doing it in order to be able to embed it and to learn how it works for you in that sense. Because I think at the end of the day, it is going to work differently, right, for different individuals.

Samuel:
So just start. Yeah, that’s what I’m telling my customers. Just do it.

Erin Chapple:
Sure, just do it.

Samuel:
My last question—and I know it’s not an easy one, and that’s why I’m asking it, because I think the answer is always interesting. Looking ahead 10 years—I know it’s a long time in AI and technology—how do you think AI will change the way we live and work?

Erin Chapple:
Look, this one—I don’t think you should be allowed to ask this question, Samuel. I think about it now, like, I have a 10-year-old now, and I think about the year he was born and where we were in the evolution of technology. And now, I’m not sure I could have predicted, in some sense—particularly in the last 12 to 24 months—it just feels like things have taken sort of an exponential shift in where they are.

Samuel:
Yeah.

Erin Chapple:
I’m going to tell you what I hope is true. What I hope is true is that as a society, as humanity, we are able to focus more on the things that drive benefit to our lived experience because we’re applying more of our time on things that are uniquely—right—at the core of how we as humans think through and problem-solve.

And we’re able to do that because we’ve found a way to move to this world of Frontier Firms, where we have this expanded workforce because we are working in conjunction with agents and AI. And that is helping to both take the manual, repetitive tasks that are the least favorite portions, I think, of both your or my role, and that we’re able to spend more time on the things that are higher output, right?

You know, rudimentary—I would go back to, like, I was part of the team that—I wasn’t part of the initial creator—but I was part of the team that brought PowerShell, which is this automation scripting language, into the masses. Like, what was that really? You know, it really helped IT be more productive and spend more time working on the things that were higher impact to the business because they could automate and use PowerShell, right, to get rid of that.

Look, this is that to the nth degree. And I hope that we look at the ways in which this creates capacity for us—both leveraging AI, but also leveraging human capabilities—for us to actually focus on the things that really improve the quality, I would say, of lived experience across the planet.

It’s big, but you know, we gotta aim big.

Samuel:
I’m glad I asked the question. It’s a beautiful answer with a super positive outlook. Erin, thank you so much for joining me. This was insightful, it was interesting. I loved your energy. So thank you so much for taking some time out of your very busy schedule to join us on the show today. It was a pleasure.

Erin Chapple:
No, it was great. I’ll come back any time. So I hope we get to continue the conversation.

Samuel:
Awesome. Thank you, Erin. Have a wonderful day.

Erin Chapple:
Thank you.