Morgan Stanley TMT Conference

Keith Weiss: Excellent. Thank you everyone for joining us. My name is Keith Weiss. I run the U.S. Software Research franchise here at Morgan Stanley. And I'm very pleased to have with us from Microsoft, Jared Spataro, Corporate Vice President of Modern Work and Business Apps. Jared, thank you so much for joining us.

Jared Spataro: My pleasure. My pleasure.

Keith Weiss: So you've been with Microsoft over 18 years now. Currently responsible for Copilot, Microsoft 365 business applications. I think we're going to go into it head on. I think the biggest topic of conversation right now is Microsoft 365 and Microsoft 365 Copilot. Before we get into that, let's just talk about the franchise that has been Microsoft 365, or specifically, the Office 365 commercial base. This is a business that we estimate is around a $40 billion run rate but still growing really well. Still seeing double digit growth, both in terms of seats as well as pricing, on the pricing side of the equation. So maybe we can start with the queues, the seat side of the equation. Can you talk to us about where that's coming from? All of us use Office 365. From our perspective, it's prevalent everywhere. Where do you guys see the additional potential for seats?

Jared Spataro: Sure thing. Well, just to level set for everybody, we hit a really important milestone recently where we announced over 400 million licenses, and that was really important. As we look at where there's additional seat growth, and we definitely see additional seat growth in the future, it comes from a couple of places. First, we see SMB as a great opportunity for us. So going down market, essentially. Second is our frontline workers where we do a nice job with information workers, but frontline workers, people who don't carry PCs or work at a desk, we see lots of opportunity. There are over 2 billion of those in the world. And then finally in emerging markets, we still see a lot of opportunity. So in the developed world, our penetration is pretty high. But in emerging markets, we still see tons of opportunity.

Keith Weiss: Got it. And when we think about that frontline worker opportunity, I think that's the part that I get the most investor questions about. What is the functionality? What is the use case for a frontline worker, somebody working in a warehouse, in a retail operation? What are they using the platform for?

Jared Spataro: Well, there are a couple different places the frontline worker really has a value prop that resonates with customers. The first, I'll just start with the baseline. And that baseline ends up coming from security, in fact. So there are a lot of folks who carry devices. You'll see them in stores like Target and other places and warehouses, and they need to secure those devices. And our front line allows us to do device security and management. So that's the baseline.

After that, we typically get to communications, and companies want to be able to push out communications to those workers. And so Teams becomes kind of next place that they go. And related to Teams, then they quickly get to information access or collaboration, and there's a lot of collaboration that happens. That's why you see people with headsets in different types of organizations who are frontline workers. And Teams can be used, in fact, as a communication, both real-time and chat-based communication.

Increasingly, we are also starting to see some pull for Copilot. So this idea of having a chat-based or voice-based interface that would allow people to get access to information real time in a very dynamic way. So those -- all three of those give us great opportunities with front line.

Keith Weiss: So basically connecting these workers into the digital infrastructure --

Jared Spataro: Exactly.

Keith Weiss: -- and all the goodness that it enables in terms of information retrieval, process as well as collaboration.

Jared Spataro: That's right. Yes.

Keith Weiss: The other side of the equation is going to be the ASPs or the price side of the equation. For the past several years, it's been about the E1 to E3, E3 to E5 migration. Can you talk to us about where we are just generally in terms of the migration? Where does the base shake out in terms of that progress?

Jared Spataro: You bet. So again, we think of this as average revenue per user and SKU sales kind of moving people up the stack for us. And we would still say that there's a lot of room, early days. So our CFO, Amy Hood, disclosed, I think it's been about 18 months ago that we were 12% penetrated with E5 in the Office 365 base. And if you think about 12%, man, there's still a lot of opportunity there for us.

We just see tremendous opportunity. We tend to think of our ability to first secure the license, and then from the license to upsell to increase the average revenue per user. And that's a motion that we do very well. By the way, we do very well at that -- sounds like a dog whistle there. Do very well at that both in the E SKUs as well as the F SKUs, those frontline worker SKUs. So it's not just in the information worker space. Wedo it alsowith the frontline worker space.

Keith Weiss: Got it. And so the frontline workers have the potential of getting an expanded functionality. Maybe focusing on the E SKUs for now. What is the core functionalities that are driving people up that stack of going -- and particularly the E3 to the E5, what's compelling that upsell?

Jared Spataro: The big driver for us is security, and that won't surprise anyone out there. But from E3 to E5, security is really the big driver. But there are other components that make it economically make sense for customers. So those include things like analytics. So we have BI capabilities. Those include compliance capabilities that are there. And those include Teams Phone, what we call telephony capability. So it's a pretty screaming deal as you move up the stack. But the real driver that's just there every time for us is security.

Keith Weiss: Got it. Got it. So there is a component that you're bundling functionality, there's good price performance for the customer. And in this spending environment, that's probably even more of a good tailwind for you guys.

Jared Spataro: That's right. We call it do more with less, and it's our ability to go in and say, why pay multiple vendors what you could pay one vendor for at kind of a discounted price.

Keith Weiss: Okay. I'm very proud of myself. I've restrained myself for seven minutes before turning to Copilot. But Copilot is going to be part of that ARPU story on a go-forward basis. Maybe just to start out with. Copilot is an add-on sale onto sort of existing E3 and E5 users, if not E7. So the pricing strategy on Copilot of not just being an additional tier but being a separate add-on, can you walk us through that decision?

Jared Spataro: Sure. This one, it is interesting. I've had a lot of customers and other folks in the market ask why did you do it. It's very simple. We wanted to be able to reach into the base very broadly. If you think about that 12% number I just gave you, that means we're kind of moving people from E1 to E3, finally to E5. That means only 12% of my base is on E5.

We wanted to give people who are all the way down, not just on the M SKUs but even the E SKUs, we wanted to give them the opportunity to buy, certainly try and then buy and then expand their usage of generative AI. So for us, it was just a numbers game about making sure that we had broad coverage and could drive scale. And that's -- over time, we continue to always work with our packaging, but it made a lot of sense for us to make sure that we could get to the base in a broad way.

Keith Weiss: Got it. So with the add-on, you get a broader set of a base day one. You don't have to wait for people to go through the journey.

Jared Spataro: That's right.

Keith Weiss: And can you remind us the Copilot -- Microsoft 365 Copilot is available to which SKUs today?

Jared Spataro: You bet. This one is important. So it's available today. One of the changes we made in January was to the E SKUs. So OE1, OE3, OE5 as well as the M SKUs, M1, M3, M5. And so it's just available broadly to the base. And then if we go down market, it's available on two SKUs. It would be business standard and what we call business premium. So those two.

Keith Weiss: Got it. Got it. This is my number, it's not your number, but we estimate that that applies to -- again, that 400 million total base, about 190 million of users that are applicable to potentially pick up the Microsoft 365 Copilot. I think the biggest question for investors is pace of adoption. And it's a big base that you're penetrating into. Obviously, it's going to take some time. But maybe you could give us some of the initial customer feedback, what you've been hearing since the general availability back -- or early last year.

Jared Spataro: Yeah, it hasn't been that long, but let me just kind of recap where we are right now. Very, very exciting time. At the individual level, we've done quite a bit of work that you may have seen published by Microsoft that people are definitely seeing time savings. So our best users are saving over 10 hours a month. We find that about 70% of people who are using Copilot tell us that it's helping them to be more productive. When we run through a battery of tests, we find that on average in what we would call typical information worker task people are almost 30% faster in their work. So we call that more productive. It's 29%. We find that 77% of people who use it for three weeks or more tell us they don't want to work without it. So at that level, we feel like we have been able to establish quite a bit of value where people say, hey, this is really value.

Now, we are just at the moment, though, in adoption where people say to us, yeah, but now we're not quite sure what to do with that at an organizational level. So let's say everybody saves 10 hours a month. Does that mean I should allocate more budget to this? Or how do I think about an IT budget that is fairly fixed or not growing by leaps and bounds? So building that business case is right where we are.

In terms of deployment, what we find is there are certain patterns that are starting to really make themselves manifest right now. We call it a land and expand. So people trial the product. Not with a large number of licenses. They typically do it with a small number of licenses, and then plot out kind of a deployment plan that would allow them to expand out over time. That's pretty straightforward in the way it's working. That's a little bit different from Office where Office is such established value at this point that we don't do much land and expand. You don't start who should get email and maybe we should roll that out to other people. That's just in a different part of the adoption curve.

Keith Weiss: Got it. It's a different kind of unit economics versus what you see in Office. Within those uses cases, is there -- part of -- there's a broader engine of capabilities of what Microsoft 365 Copilot can do. It can do generation of emails. It can do summarization. It can do information retrieval. Any initial trends you're seeing of like what parts of that that capability set's being most used? Or maybe go a step further into use cases or sort of the most common use cases that you're hearing?

Jared Spataro: Sure thing. The first one that we see that is probably the biggest game changer for people understanding it as a valuable kind of addition to their workflow would be the chat experience. This where we have chat as a part of Microsoft 365 that goes across all of your data estate. So it's looking at your emails, your calendar, your documents, your meeting transcripts. Here we just have no competitor. There's nothing that does anything like what chat can do.

And the use cases here are often about two things. One is sophisticated information retrieval. And then the second is what we would call kind of sophisticated task completion. So something that people can do with it that they can't do with anything else is, hey, Brad missed our standup meeting a week ago. Will you find the transcript for that? And then will you write an email to him summarizing the meeting and pointing out the action items that we gave them in the meeting? That's something that typically takes about 20 to 30 seconds for it to return. Very concrete. And you can't do that with anything else except for your own little hands. So there's something about that task completion that is really magical for people.

Then we move over into Teams. Copilot and Teams never disappoint. It's this thing that is amazing because it just grounds on the transcript. You can use it during a meeting. You can use it after a meeting. Again, not really a comparable out there, and it's something that really saves people time. We hear them saying, hey, I don't go to as many meetings because I typically go to these types of meetings just to listen to what's happening. I don't have to do that any longer.

And then the third one that really is catching for us is Copilot in Outlook, which is email. Once you really know how to use it -- it takes a little while to figure out how to use it in your workflow. But you just love -- I love to hit that summarize this long email thread, and it does a great job. Summarization is a very strong use case. And you can also use it to write emails, especially ones that you have to kind of have to use a lot of brainpower to write, and it can help you with triage of your entire inbox. We also see people using it in Word,Exceland PowerPoint, but those are coming along. Those aren't as strong as those first three.

Keith Weiss: Got it. So there was a recent Wall Street Journal article that got a lot of attention amongst investors that it was highlighting some pushback on the current state of the functionality versus the price being paid. And basically making the argument, it's kind of expensive for what you get. Do you find verity in that article? Is that something that you hear a lot from customers? And maybe taking it from a different perspective, how do you think about the price? How do you justify theprice, ifyou will?

Jared Spataro: Sure thing. So yeah, they ground you on where we are right now. Across the portfolio, Copilot in those different scenarios, those different applications, it's in different stages of development. So if you look at Copilot and Excel, like we all can't wait until an Excel jockey can do a lot. But it's coming. It's learning. And people have very high expectations. The same is true of PowerPoint. And that was a little bit of what they focused in on in the article. They were like, man, come on. I want this thing to be my financial analyst. And it's going to disappoint there because we're still learning the commanding surface of Excel. And so that's pretty straightforward.

We are able I think to justify the price point quite well when we move over into those scenarios that I listed as our top three scenarios. But typically, people don't go there first. So we do have to spend some time with our customers, showing them and then helping them realize that value.

Now in terms of what this looks like, it's worthwhile just to look at the list prices for the other things that we sell. $36 for ME3. $57 for ME5. The $30, that's a substantial price stack, for sure. So people are definitely trying to understand, who should I get this for? Is this for everyone in my organization? Is this for a certain segment or population? And mostly where we are right now is working with customers to help them build out that business case. And it's just based on hard facts. Are we saving enough in terms of time? Are we generating enough value?

And right now we're very optimistic, but for this group, I would say my job here is to temper your expectations. Over the long term, we think this will be a great moneymaker for us. But don't think that it's breaking free of the normal deployment cycle, the normal evaluation cycle, the normal budget cycle associated with IT. And you'll kind of see all those things come into play. So it's an interesting thing to think about the short term impact versus the longer term impact.

Keith Weiss: Got it. So a couple things I want to dig into there. When we think about software and sort of the value-add, it's the productivity enhancement that you get on any kind of given asset. So if I take what you're telling me, the average information worker makes $75,000 a year. You're saving them 10 hours a week. That's probably 20% to 25% of their work week. It would seem $30 against that productivity savings is a pretty good value proposition or a very high return on investment. Is that the calculation that your customers are doing? Or do you have to sort of prove it out a different way?

Jared Spataro: First, I'd say you're hired as a sales person. That was wonderfully, wonderfully done. That's exactly what we're trying to do is to get people to understand they're fully loaded costs for their FTEs. You spent $75,000, $80,000, $90,000, depending on the market, and let's do 12 times $30 and talk about the time savings.

But here's where we are that is so fascinating to me. What I see -- I'm just down in the trenches with them -- they say, Jared, at this point we wouldn't dispute that that's saved. We're just not sure exactly what to do with that information. My budget is what my budget is. How do I break out of this? Our most successful customers are breaking out of just IT spend and realizing that they can tap into essentially people spend that is all about a fully allocated cost. But that's kind of advanced finance math for a lot of people because that's not the way they budget. That's not the way they've thought about kind of the asset and what the asset's doing.

The number two point that is very fascinating. What people do with that saved time is really important and kind of systematically what customers do. We certainly have customers who say to us, Jared, that's awesome. But if those 10 hours are going to my employees' health and wellbeing, I don't know, man. Like is that how I want to spend that? And we try to help them understand, well, some of it goes to work-life balance. That's not bad. And some of it goes to the company.

But what an interesting place to be in where, again, we would say we don't have a lot of pushback on this is saving them time. It's now, what do I do with the time that's saved? So that's just, for us, it's such an opportune time to be here talking with you, talking with the market. Very fascinating to see how it's all playing out.

Keith Weiss: Got it. You mentioned an anticipation of doing more around Excel and PowerPoint on a go-forward basis. You recently announced Copilot for financial services. Can you talk to us through about what that specifies down to? Like what's the added functionality? And is that kind of the unlock that's going to make it more interesting for like Excel users and my associates?

Jared Spataro: You bet. Yes, I think that over the mid to long term, it will. Let me tell you a little bit about our strategy here. You should think of Copilot for Microsoft 365 as being designed for all workers and all the information worker task that we all do. So whether it's Teams meetings or writing emails. But there are a lot of things that people do within the function, the domain that they work in that we essentially have to skill up Copilot to understand. You have to give it the context of understanding what is a variance report? What does that look like? How do you think about the basic finance task or sales task or support task? And what we're finding is we can do that work, but we would like to monetize what we do there, what we would call those extra skills in a different way.

So we have priced those skills at a premium. The package is $50, and it's a combination of the underlying M 365 Copilot and these additional skills. So you can basically think the imputed value of the skills is $20. And we're just going to go really work hard on the core tasks. And what the team is doing -- it's actually our Biz Apps team that's doing this. They just go into the functions and are mapping out a core task that each of those functions are really focused on and teaching Copilot the context of those tasks, how to do those tasks.

Keith Weiss: Got it. So you mentioned before that -- we do have to bear in mind, this is enterprise software. And one of the difficulties in enterprise software in gaining adoption of new technologies is how do you get this into somebody's workflow. We come into the office in the morning, and I have plenty of stuff I could do with my five core apps all day long. How do I get something else into that workflow?

So the question would be is Microsoft 365 for Copilot, the financial services edition, is that part of like more directly addressing the workflow? And outside of that, how do you do that? How do you break into somebody's day to day?

Jared Spataro: Yeah, let's talk about two vectors we're really focused on. Vector number one is getting them the data that they need. So that is an important part of these for sales, for service, for finance Copilot. It's constructing the very sophisticated data layer that connects into those systems, whether they be SAP or Salesforce or whatever it happens to be. That's incredibly important. Today, out of the box, Microsoft Copilot is able to connect into the Microsoft data and do so very well. But we're providing those connections to the other data. So that's number one. When people can access their own data, the stuff that they need to use every day, it's super valuable.

And then number two, interestingly enough, we just found through some research kind of a pattern that gives us some sense for what it looks like to work into the workflow. We call it 11 x 11. And basically what we find is if someone has used Copilot for 11 weeks or more, and during that time, they would say -- and we do this in a survey-based way -- that they believe they are saving 11 minutes or more, just so happens that's what the cutoff is, they work it into their workflow. They really feel like, man, this is valuable enough as a tool that it's going to become a part of my arsenal. I'm going to use it every day going forward. It is the clearest indication for us, that 11 x 11, of when it has become a daily habit and it's starting to become a part of the way they do work, which is great. And we see a lot of that. It's not like we see a big drop off. We do see you have to teach people how to do, though.

Keith Weiss: Another aspect of being enterprise software is there's security, there's compliance. There's types of concerns that the end customer will have. And the one that we hear about most often is data security. You're giving these employees now an ability to go out and do a really good search against a much broader set of data. And the companies are worried, do we have the controls in place that are going to keep Keith from seeing stuff that he shouldn't be seeing? How does Microsoft work with the customer to get over that fear and sort of ease that deployment?

Jared Spataro: Yeah. Most customers do not. And this is definitely something that comes up kind of as an initial objection. Gosh, Jared, I love this, but our early tests, our early trials tell us that we are doing security by obfuscation, meaning we don't have our permissions set correctly. We don't have the groups all set up. And now all of a sudden, you can ask Copilot some pretty detailed questions that can bring back lots of information.

So you can imagine, we have been really digging into this. And mostly this is about giving people the tools they need to quickly get to a place that they feel more confident about. So we have a great product out there called Purview that allows you to do essentially, also interestingly enough, AI-based labeling where you can go out and find your content, label it, identify it as highly confidential, business important, et cetera. So that gives you some sense for what essentially kind of can be your data estate.

The second thing that we've done is we have been working with our customers on a way of not whitelisting, but instead being able to create this list of places that you want to go to look for the information while you cut out everything else. And what we find customers do is they say, you know what, these three sources of information, we feel really good about the permissions, but the others we're not so sure. And this let's them get started really quickly.

So it's mostly about building confidence, about getting to a place where the data that they're focused on, they feel good about what it can do.

Keith Weiss: Got it. And Microsoft also is bringing tooling to help --

Jared Spataro: Correct, yes. We're inventing this as we speak with the customers.

Keith Weiss: Got it. So we've been asking -- we do a quarterly CIO survey. We talk to 100 CIOs. And we've asked about a CIO's expectations for adopting Microsoft 365 Copilot. In our most recent survey, 38% of CIOs expected to adopt Microsoft Copilot -- 365 Copilot within the next 12 months. Does that surprise you how high that number is?

Jared Spataro: I don't know if I'll comment on surprised or not surprised, but here's what I'll say to you. We believe that over time that the heart of Microsoft 365 will be Copilot. That the soul of M 365 will be Copilot. That nobody's going to want to use this suite of applications and capabilities without Copilot. So we're really trying to develop it so that it is so compelling, so that it feels so differentiated, feels so valuable that you're just going to not want to work without it. That's the focus.

So 38? I love 38. 38's better than 37. But I'll tell you what. When we look at where we're headed, we'd like this to be 100% of people feel like, gosh, that's a tool that I really want to use. We would look back and say anytime that there's a discontinuous change, this would be going -- sometimes I tell my team, going from shovels to the steam shoveler to the backhoe. You're really trying to convince people that almost every job would be better done with augmentation of human capability as opposed to just what humans can do on their own. And that's where we think we are now for information work, not for physical work.

Keith Weiss: Got it. Two questions that stem off of that. One, one of the questions that I've gotten from a lot of investors post -- when you guys first announced pricing on Microsoft 365 Copilot was is Microsoft pricing this for value or for ubiquity? And historically, we've always thought about Microsoft as almost pricing everything for ubiquity, and it's still a debate out there. It sounds to me that you guys think this is the right price point for ubiquity.

Jared Spataro: Yes, becausewe think it's that valuable for us. There are constraints here that are a little bit different than some of the other things that we sell. These GPU things you've heard a little bit about, like they cost real money, and running them costs real money. And so we do have a cost structure associated with it. And we feel like we need to, on behalf of our investors and shareholders, make money. So we're watching all those types of things just like we did in the cloud, looking at margins, looking at the cost structure, looking at what we can do with all of that.

But if I really pause for a second and say, okay, we're going to deliver a personal system that over time is going to get to know you better and better, that can do things that no human can do, that allows you to reach across all your information, and as it gets better within the apps, will take all the drudgery, all that toil out of the work you do, how much is that worth, given your fully burdened costs? Gosh, we do not think that $30 per user per month is an exorbitant cost. And so we really did try to price in a way that was like, we feel like it is equal to the value that we aspire to, and we hope everybody would be willing to pay that price point.

Keith Weiss: Got it. And the other question that it spurs, thinking back two years ago when Amy Hood was in that same seat, we weren't talking about Microsoft 365 Copilot at all. And the pace of innovation that we've seen on how fast this has come out and how quickly you're innovating on these technologies has been pretty amazing. What should investors be looking for over the next 12 months, over the next 18 months? Like where does the Microsoft 365 Copilot story go?

Jared Spataro: First I would just say as a direct response to that, please don't think that we've hit some sort of apex in innovation. Now we're just about grinding it out to make it better. We certainly are going to grind, but there's a lot still to come out.

If I think about the system that is the Copilot system, you basically should think that you have kind of four big components. There's Copilot as an orchestrator, and it orchestrates the data that you're going to grab. It's going toorchestratesthe underlying LLM. And it orchestrates the specialized skills. Every aspect of that little system right there is continuing to get better in a non-linear way. So the orchestrator itself is getting better. The data, what we have access to and all the sophisticated work that we're doing, getting much better. The LLMs and the underlying models, you're seeing those get better. And then the skilling, you will continue to see it get better at Excel, better at image generation, it will get better. All of those things will get better, and it will be non-linear. So I think that you should expect that it will continue to feel pretty magical, is the word that I'd use.

Keith Weiss: Got it. The evolution a lot of people are talking about in terms of like where Copilot goes or sort of where the functionality goes is you start to add more workflow to this and you're starting to talk about intelligent agents. So what's your perspective or the Microsoft perspective on intelligent agents? And is that where the roadmap is leading?

Jared Spataro: It is. We don't want you to get the sense that the Copilot itself is going to necessarily become an intelligent agent. What I mean by that is it'll be intelligent. It's an agent. But the domain, the way we think about it is that Copilot will be a personal assistant. It'll get more proactive. It'll get better at understanding your needs, like all of that will be true.

But when we talk about intelligent agents, we think that there will emerge a class of AI tools that are no longer just bound to a single person as the personal assistant. They will be bound essentially to business entities. So they'll be bound to meetings or processes or workflows. And there they'll be shared by many people. And they'll be very focused on driving efficiency in those processes. And we absolutely think it's headed there.

If I were to just give you an example of kind of the types of things we're seeing, finance has a bunch of these types of processes. Things that are fairly repetitive, but they require today human brains to do. We certainly see support be another place that you see that. Even sales in terms of like the repetitive work of selling to consumers and to commercial customers. So there are a bunch of places that we think these agents, semi-autonomous agents could do great work in the coming years.

Keith Weiss: Got it. And can you -- just an example of a use case there. Within HR, is it -- when I want to reallocate my 401(k), it's the intelligent agent that's helping me through that process versus having to get on the phone with a person.

Jared Spataro: That's exactly right. A little bit earlier, I was at an HR conference and we were talking about these types of things. You essentially have help desk internally, and those help desks help you to do things that are very real for employees. Often times you end up having a person who picks up the phone and helps you do that work or works you through it. And we believe a lot of that can be replaced with agents.

But I'll go down into kind of the common business functions we see out there. In finance, as an example, closing a quarter or giving a quarterly or a monthly report on my business takes a bunch of people to date to do. But they pull data from systems. They look at that data. They munge it together in very predictable ways. Those types of things we think will be done by agents very well in the coming months and years.

Keith Weiss: Got it. Shifting gears a little bit. I want to talk about the partnership with OpenAI. It's been a tremendous success I think for both sides of the equation. The recent almost leadership change at OpenAI late last year really brought up a question for investors of Microsoft's reliance on the GPT model to power the AI offerings and power your solutions. Can you talk -- maybe address that question directly. How reliant is all of this on the GPT model? How reliant is it on the OpenAI relationship?

Jared Spataro: Sure thing. Yeah, let's start with OpenAI. A very important strategic relationship for us. I don't have to explain that to you. But it is worth saying, why is it complementary? We just love what OpenAI has done. They pushed far beyond what our research have been able to do when it comes to large language models and the results from those. At the same time, they have not had the scale, the structure, even the skill set associated with scaling the information to run those models. So we think it's a very complementary relationship right now. We've got a really good thing going.

At the same time, when you look at what Copilot is today, it doesn't. Today, it already does not rely on a single model. Certainly the one that is held up would be GPT. And increasingly, GPT-4 Turbo is the one you see that in Pro and you're seeing that in Copilot for Microsoft 365. But we have other models that we use. Some of them are homegrown models, and it's engineered to be multi-model type of product for us.

And if you look broadly at what we're doing with Azure, that's the place that you'll see us ensuring that we can provide a model of the service, essentially. So you see the investment that we've made in Mistral. You're seeing what we're doing with Llama. You see what we're doing with other providers out there. All of that is important to us. We think of ourselves as a platform company, and that platform-ness of what we do isn't just kind of in the services. It's also in the finished products.

Keith Weiss: Got it. And moving more broadly across the solution portfolio. If I think about the fundamental capabilities of what you're doing with Copilot, and then the vision of what you guys were doing with Power Platform, giving more people access to information with Power BI. Giving more people access to workflows. Giving more people access to almost become a citizen developer. The two seem to go together very well. So how does Copilot potentially accelerate what you guys are doing with Power Platform?

Jared Spataro: Gosh, peanut butter and jelly. Really what you should look for in the future, we think kind of future trends would be this combination of AI and automation. There's so much that can be done with AI and automation. And the way to think about that, or at least what we're doing right now, starts with a product that we call Copilot Studio. We created Copilot Studio based on the Power Platform. It used to be what we called our Power virtual agents product. We rebranded it, changed it, and tucked it into that $30 SKU that we have.

So it's a part of what you buy with the $30. And it allows you to kind of get started with the Power Platform without even knowing it. As you use Copilot Studio to do a little bit of automation, it has basic workflow and automation in it, and a plugin to additional data sources, you are starting essentially, without knowing it, to use the Power Platform. So we think of this as a perfect way, with the packing that we have done, to start to drag through Power Platform and get people exposed.

Right now on Power what I'd say is, man, it is a really underdeveloped asset for us. And it's just hitting the point where we have enough market momentum, enough customer usage that we feel great about its capabilities, and now the next chapter is scale. And when I look at it I think, man, Copilot could help us scale in a way that we've always dreamed of with Power.

Keith Weiss: Got it. And I think from an investor perspective, I'll be honest with you. I don't talk about Power Platform very often. And I think it's a really interesting opportunity. Part of it is investors are pattern matchers, and they're tying to understand kind of where this fits into the landscape. Power BI was easier. Tableau versus Power BI was dynamic. When we think about flow and automating the broader platform, how do you think about that competitive dynamic? What is this replacing if it's replacing anything? Or sort of where does this come into the budget?

Jared Spataro: Yeah, just to provide some context for you. We've announced publicly, I can't remember what quarter it was, but you'll find it in earnings, Power Platform all together is over $2 billion. So it's a big business for us. And as you said, the Power BI side of things has kind of a more established set of players.

For us, what we're seeing is the emergence of low-code as a category, and we're one of the leaders in that category. We see ServiceNow and others competing in that category. Typically what we're doing is not necessarily replacing budget that was spent on building as much as it was spent on applications. And that's a lot of what we're thinking about is how can we get out of this like kind of pre-created application space to the place where we're helping you create exactly the solution that you need, but in a low-code, low-cost type of way. So we're essentially, it's a bit of a paradigm switch for us and we think giving flexibility to the IT portfolio. But man, lots of opportunity in Power. If there's a part of the portfolio that I'd say is under-known, it would be Power and its ability to monetize $2 billion and growing very, very aggressively right now.

Keith Weiss: Got it. If I was going to add a second one to the underappreciated assets within the Microsoft portfolio, and it falls under your portfolio as well, would be Dynamics 365. So I think it's been a number of years since you guys had re-platformed Dynamics 365 to all run on top of the Azure platform. How has this changed the competitive dynamic?

Jared Spataro: I had the chance to start to manage Dynamics, I don't know if it's been 18 months ago. And I have just been amazed at the price/value ratio we have with Dynamics. If you look at the price point versus the competition like Salesforce, we have an incredible steal of a deal. And I do believe it's a lot of like under known. We announced just recently that FY '23, we just passed the $5 billion mark. We are very focused on three categories: sales,serviceand ERP. In all of those categories, we have some down market momentum, but in particular in ERP, we have a real winner of a product down market. So there's a lot of assets for us.

And what I would say is we see two opportunities. Number one, we're doing some competitive takeouts where we're seeing momentum like we've never seen it before. So we're really starting to see people gravitate especially, as you would imagine, as they get up to those anniversary dates on our competitors, and comparing the price point versus the value, we see a really compelling deal. But the second place -- I don't want to over say it or oversell it, but Copilot gives us an opportunity to really have people rethink their overall business application strategy. So as they see what Copilot can do, they recognize, wow, maybe we should be thinking about a change here.

Keith Weiss: Got it. And the Purview of Dynamics 365, I've gotten chided several times by Microsoft executives for pigeon-holing it as an SMB offering. How applicable is it up market?

Jared Spataro: Oh, very applicable. Yeah, for sure. We do a lot of business in enterprise. It's very competitive in enterprise, as you mentioned. We've been working really hard to ensure that we're competitive. And at this point, I think we're being incredibly innovative. For instance, those Copilot capabilities that we talked about kind of on the M 365 side of the house, we've tucked right into the core dynamics offerings themselves. So as you buy a Dynamics 365 for sales, you're getting incredible value with Copilot right there besides all of your users. So there's a lot there. We feel really good about it not just down market, but also up in enterprise now.

Keith Weiss: One of the things that Microsoft has talked about, and I've actually heard from other CIOs is that the Dynamics CRM Copilot is not just for your Dynamics CRM. You can use it against other data sets and other CRM systems as well. That extensibility of Copilot, is that something only in Dynamics 365, or would you see that more broadly?

Jared Spataro: Great question. We've clarified that recently. So we started out as we were innovating with exactly the model you just said, we have now broken those two things apart. So we have this Copilot for sales. And Copilot for sales can be used with Dynamics, but it's designed intentionally to also be used with other data sources and other line of business systems. So that would be, for instance, with any other CRM system, including Salesforce, and we definitely see customers who are using it. Mostly where that's coming from is nobody wants to spend more time in their CRM system. They want to be in the flow of their work. And that's what Copilot is meant to do is pull that data up right within your Teams calls, your email, to places that as a seller, you're already getting your work done.

Keith Weiss: Got it. We're running to the end of our time. I just wanted to squeeze in one last question. As you think about the next 12 months, before we have you here again at the TMT conference, what are you most excited about? There's a big swath of innovation going on at Microsoft, but what are you recommending is the one thing that we should really be focusing on?

Jared Spataro: I would say this. When I look at Copilot in particular, the thing I'm most excited about is its ability to become a proactive personalized assistant that knows your job like nothing else out there, that is very proactive and starts to identify the patterns as you do your job. And helps you literally automate away the drudgery that keeps you from doing the most highly value-added tasks that you can do. And you're going to see us really investing in that space. I'm excited to see what we can do with it.

Keith Weiss: Outstanding. Jared, thank you so much of joining us.

Jared Spataro: Thank you.

Keith Weiss: Great conversation.

Jared Spataro: Thanks for having me.