for joining. My name is Brent Thill, with the research group at Jefferies. Jared is with us today. He started, the stock was $24, and his, during his tenure, the stock's up a modest 1,200%. He started in SharePoint, then was a GM of Office, then in charge of M365, and now in charge of Modern Work and Business Apps. Jared, thanks for being here today.
My pleasure, to be here.
Maybe just to kick off, give us a sense of your role in creating this modern environment, what you're spending the bulk of your time on at Microsoft today?
Well, right now I'm responsible for two of what we call our customer solution areas: Modern Work, which is all our productivity business, and then, Business Applications, which essentially is Power Platform and Dynamics. Most of my time, Brent, is spent right now on AI. We introduced a new set of products that go under the moniker of Copilot, and man, demand is kind of, I've never seen anything like it in my career, so very busy with that right now.
You brought it up, so we'll keep pulling the thread on Copilot. 99.5% of the questions I get are on this, and everyone asks, you know, I know you haven't availed everything, maybe not all the pricing's out, but what does this look like in five years in terms of, does everything have a Copilot? Are we gonna have a Copilot across every application, and it's gonna be all fueled by AI? Well, I mean, just share your vision of what this looks like.
You wind me up, and here I go. Get ready. I've got a lot to say. First, we're in early days. I'll start there. You know, this is very early, but if I take a step back and paint a big picture of what we're seeing right now, there are basically two broad scenarios that we see within my portfolio for AI. One is what we would call In-app assistance, and this is the one that people understand the best. This is a Copilot in Word, a Copilot in Excel, a Copilot in Teams. It's the one when I introduce the product that everyone kind of gravitates toward because it's easy to explain, and people can see an immediate benefit in terms of productivity for them personally and often in the terms of a group. That's very interesting to us. We think it has long-term potential.
We do think that most applications will end up having a Copilot. Over time, many applications, we think, will start to have natural language as one of their primary user experiences, and that's important. While that's getting a lot of the attention, the second really big scenario for us is something that we call Cross-app intelligence. It's just a moniker that we would use to say, "Hey, independent of any particular app, you can essentially have a ChatGPT-like experience for your business." You can ask this experience, "What were Q4 sales? What are the trends on this product line? What are my competitors doing? How do I compare versus competitors in this category or sector?" This is a brand-new set of scenarios. Over the long term, I'd say, Brent, this is the place that we actually think will be most disruptive because it's brand new.
It's the hardest one for me to explain to customers. I typically have to demo it. I have to bring the product up and say, "Watch this." When they see the concept of ChatGPT against their company data, against their financial data, against their, you know, graph data, then they start to realize, "Wow, that could be a very big inflection point for how we think about productivity." That's the one that we're most excited about, so we are gonna pursue both. Again, I think that first one's easy to sell, it's easy to build, it's easy for people to understand, but that second one probably has the most disruptive opportunity over time.
Back to when you haven't seen anything quite like this.
Yeah, sure.
How do you keep up? Like, how are you guys keeping up with this? Are you having to hire a lot more people? What, how do you stay in front of it? Or is it, "Hey, get in line, and there's seven bodyguards, and there's a line out the door"? How are you?
It's busy. You know, it's busy. I'd say that. It's a strange time at Microsoft, but I mean strange in a good way, in that, on the one hand, macroeconomically, you have all this pressure from our customers of what we would call an initiative of do more with less, that we do very well. You know, when there's consolidation in the industry, we do very well. In this productivity space, everybody has 1 million vendors, you know, for everything from video conferencing to chat. All of a sudden, Teams becomes really interesting. We already had this foundation that we were operating in. I would say probably the biggest change is the cycle time. Like, it's just faster right now.
This Copilot thing did not exist at the end of January, you know, in terms of us having it be a product by any stretch. We made the product, came up with the concepts, came up with kind of the basic ideas over the course of end of January until we announced it on March 16th. That's how fast we're moving, and then we've continued to be fast since then, you know, adding new capabilities to it. How do we keep up? I would say, you know, kind of radical prioritization. There's only so much we can do, and more than anything, I'm spending a lot of time with customers. That's where I'm focused. I don't, I don't personally build the product any longer.
We have teams that do that, but I wanna stay plugged into customers and what they're thinking, and that's where I'm spending most of my time is in those discussions.
For the rest of the year, I would assume we're gonna see a lot more announcements soon.
Yeah, we have a lot coming, I'd say, for sure. I would also say that just getting the world to digest what we have announced is a big piece of work, and we're very focused on that. We just opened the preview program for this Copilot to 600 enterprise customers, so we're very focused on those customers. We've been way oversubscribed on that, so we just opened the list and had as many customers as wanted to sign up, and we're just kind of starting to chop through that list and get them using the product. Both kind of helping the market digest what we announced and at the same time continuing to innovate. That's what we're up to.
The monetization strategy, we can see it in GitHub.
Mm-hmm.
It pretty nice price uplift. I mean, when you think about kind of what's baked in at the core versus what's the incremental add-on, how do you think about this? I know you can't unveil all the pricing and all that right now, but how do you think about this? How should we think about it as we're all gonna be probably evaluating this soon, you know, as customers?
Everyone wants me to utter things that will help you build models. I'm sure of it. It's early days. Here, let me look at the competitors for a moment, we can talk about public things. Competitors typically do a per user, per month in this space, a productivity space, 10 between anywhere of $10, $70, $80 per user per month. We don't know where we'll fall. What we're really focused on is trying to understand kind of a couple different dimensions. Number one, where's the value? Where do people really see value? We have home runs in the product right now in AI, in the context of Teams. Every time I demo that, like, actual product, people are like, "Wow!" We've got really solid technology in the context of things like writing in Word or processing email, Outlook. Those are quite strong as well.
We've got other places where we're still developing, we're trying to figure out the value. I'd also say that when you think about this technology, it is expensive to run. You know, this runs on GPUs, as you're seeing the coverage right now. That was something that people are starting to understand, there are costs associated with it for sure. We're trying to understand the economics of usage. You know, what do usage patterns look like? What does that cost us, and what can we do to drive costs down? The reason I keep saying it's early is the value and what people perceive the value to be, we're still understanding that, our own ability to run this efficiently, we're still getting the hang of that.
Right now, you know, it's not like I'm hiding some sort of like monetization model that I'm just not telling the room. We're still working on it. You know, we look at a couple of different things. We also have to look at things like making sure that we don't open the door to unlimited usage. We do that, and you've seen this from OpenAI, you have, again, a comparable that's public right now, where with their GPT-4 model, they limit the number of essentially prompts that you can give it over an amount of time. We're playing with all those types of variables right now. Again, you just have to understand the cost model and understand the value model, and somewhere in there, we'll get to a monetization model.
I'm not saying you're doubling the price per seat, per user, per month, but if you did, it would be because of value, productivity, and then the cost to run. You, it's just. We've seen.
All those factors. Yep, that's right.
We've seen 2x lift on pricing on some of the original products, so that's...
GitHub is a really good example, if I probe into that for a moment, of us, I think, doing the right thing where we're able to run studies on productivity uplifts so that we can tell you, for instance, that developers who are using Copilot say they have just about 50% of their code now written by the Copilot. We can do calculations and say: What's that look like, and what would we charge you? We're, you know, you can imagine we're doing the same types of things now. As we move to these 600 customers, we're looking at, "Okay, talk to me about your average day, talk to me about time saved, talk to me about where you put those minutes. You know, what percentage of writing do you do differently? Analysis do you do different?" All those types of things.
We're using the same methodology we did with GitHub as a way of evaluating value here.
One of the biggest questions we get is, if you're successful with this, then inevitably, are there just less actual users?
Mm. Mm-hmm.
Your pricing goes higher. I mean, when you think about the volume of users that can use versus pricing, I think most of our clients are defaulting to, well, does XY assistant need to be here? You're seeing a lot of companies saying: We're reducing jobs because of AI. How do you address that? What do you think happens there?
Yeah, let's take a step back and just talk about my space for a moment. A couple of quarters ago, I think it was a couple of quarters ago, we announced 382 million users of the Microsoft 365 constellation of things. In the grand scheme of things, we actually think there are over 600 million information workers in the world, and on top of that, another 2 billion frontline workers. When we look at, you know, "Hmm, do we have opportunity to continue to grow seats for the business?" We would say, "Yeah, we feel pretty good about that." Will there be some consolidation of jobs over the mid to long term? We say yes. Over the short term, I think most CEOs are using AI as an excuse to get leaner, to be honest with you.
I can't imagine they're really using it in ways that are cutting jobs. I'm not too worried in the short term about it reducing the number of sockets that I have to sell into. Over the mid to long term, yeah, I think there will be some changes, but generally speaking, our position is we're very tech optimist. We don't think that this all of a sudden automates people out of jobs. We think they go on to do higher value jobs within the functions that they run. I'd also say that it gives us an opportunity with AI that we're very excited about, with some products that are public, to also start to move from just a plain per user type of charge to more consumption-based charges. That combination could be pretty interesting for the productivity space. I'll give one concrete example.
We have a product out now called Syntex. Syntex essentially is AI applied to content management in bulk. Of course, you can do that on an individual basis, but you can do this on a repository basis, so on contracts at large or on those types of kind of documents. Here, the industry already is charging on essentially a consumption basis, and that's pretty interesting to us because on top of what we can do on a per user basis, we can add a different economic model. Right now, you know, my biggest problem as I talk to the board or to our senior leadership team is not, "Oh my gosh, I think AI is gonna reduce the number of seats that I can sell to." My biggest issue is, hey, there's a lot of demand. How do I turn that demand into money?
How do I continue to grow the value I deliver through that demand?
We don't have a demo booth here, but when you said there was a wow factor in Teams, I'm a huge Teams user.
Yeah.
My team hates me using Teams as much ’cause I can always find them. We use it a lot. Can you explain the wow factor? Like, what is, with AI infused, what’s different about us using Teams today that it’s gonna look different at the end of the year or early next year when you launch this?
Yeah, there is a wow factor. One of the first demos I will often do is go back to a recorded meeting and use Copilot to essentially query the meeting. Was my name mentioned? Were action items given to me? Were decisions made? If you think,
I wanna pay Microsoft $10 million that showed up at the highlight.
Yeah, exactly. I mean, it's, it is amazing to not attend a meeting and feel like in 5 minutes of prompting a GPT-based engine, that you can understand exactly what happened. We've done a nice job of doing things like recordings and transcripts, but nobody goes to dig through those things. It takes as long to dig through them as if you just attended the meeting. Not attending a meeting, but getting something out, there's a wow factor there. It is immediate. GPT, like these. You know what it looks like 'cause you do ChatGPT, very sophisticated. That's not all.
Typically, when I demo to customers, I'll ask them if I can transcribe the meeting, and I'll start down, I'll let it go for about 15 minutes, and then I'll say, "Hey, can I show you something?" I pull up Copilot, and I say, "Ask any question about what we've talked about. Any question." First, they start typically with tentative, kind of basic questions: Oh, what did Jared ask? You know, what did Bob answer? They start to realize that the thing is very, very good at analyzing, you know, sentiment.
You can ask. The other day, I was in a meeting, and my boss asked me a question, and I answered it, then I thought, "Oh, this is a great chance." So I asked Copilot, "Did Jared know what he was talking about when he answered Takeshi?" It came back and said, "It didn't seem like Jared really did know what he was..." I was making the answer up, so, and my boss knew it. He was like, "Wow, that's very impressive." You can ask it, "Did anybody tell a joke?" It can say, "Yes, there were three jokes told. How good were they? And rate them in order of, you know, how good they were. Who laughed?" I mean, its ability to understand the human dynamics of a situation is incredible.
During a very complex meeting, for instance, you can ask it things like, "Hey, parse out for me the different arguments people are making, and who agrees with what sides? Tell me what is still left to be decided. Tell me what we did decide." I mean, its ability to just do those types of things is amazing, and it changes human interaction in real time, and that's kind of one of those wow factor demos.
You love all your apps equally, but if there's kinda one that you're really excited about in the next three years, what would that one be?
We talked about Copilot, so I'll put that to the side, unless you want to ask me more questions about that in a moment. There are two that are top of mind. One won't surprise you, and one may. The one that won't surprise you would be Teams. We have over 300 million monthly active users now. We've been able to get to a point where it has a really nice position within the communications infrastructure of enterprises and increasingly smaller down-market businesses, and the opportunity we have is to monetize that. You know, the questions you will ask me is: "Great, Jared, what are you gonna do with that 300 million? Can you further monetize?" We see an incredible opportunity. We already see Phone taking off and becoming a very big opportunity for us. That is really growing very nicely.
We see Rooms growing very nicely. We see something that we call Teams Premium growing very nicely, even though we just introduced it. I think Teams is very interesting for us as a platform for monetization and serving up more value. The other one that I'll flag for you that I think you'll understand a little bit better if you tuned into our Build conference, is Windows. I actually think Windows has a moment of revival coming, as AI really changes what a operating system can be in personal computing. I'm happy to talk more about that if you'd like to, or we can just skim over it. Windows.
Yes, please.
I think, is really something that you should watch because the combination, kind of the confluence of tech trends right now, could make the operating system very interesting again.
Why is that?
Well, let me take the two big ones that are on my mind right now as I'm really trying to drive the team. The first is the move to the cloud. We have seen tremendous success with a product that we introduced called Windows 365. It essentially has taken some of the back-end virtualization technology, wrapped it up, and turned it into a software-as-a-service offering. You can spin up a Windows kind of environment in the cloud within literally a matter of minutes, and it becomes your personal environment. You can stream that down to any device. People who are starting to use this, the signal we're getting from customers is very strong. They first will pilot it, and they'll start with 5, 10, 20. Now we see people going to hundreds, thousands, and tens of thousands of Windows 365 licenses.
We've got really good signal there that this kind of Windows in the cloud is very interesting. I see people starting to now use it on their iPads. I use it on my iPad. It's transformative as you see, oh, Windows can be anywhere. It really can be my most productive environment. There's one trend there, and then the second trend that we're very excited about is Windows Copilot. If you think about what Windows Copilot can do, we demoed it at our conference called Build last week. Windows Copilot essentially can really bring together a, an agent and assistant that understands everything that you're doing in your digital life. It can watch where you're browsing on the web.
It can see the applications that you're using way outside the scope of M365, which is what the M365 Copilot is limited to, and it can bring all those things together in what, again, would be a very personal assistant type of scenario. It knows every keystroke. We think there's really something there to make personal computing more personal than it's ever been, to follow you across any device, big or small, and to essentially call that the evolution of Windows. As companies are starting to catch that vision, they're very excited about what it could do to end-user computing. We think there's something there.
Everyone's been super surprised by Office. I mean, to put up 17% to 21% constant currency growth consecutively, it's like Jimmy Butler or Steph Curry, or you go through, like, it's just the consistency is insane. Like, how do you do that?
I mean, first, I won't take credit for it. I mean, there are kind of two components to it. I'd say we have a very good product team, and under Satya's leadership, they have not become complacent. They have tried to understand customers and say: "What's the core value that we're providing? How do we improve that?" You see that in both improvements to the core apps, where something like Excel is more loved than it's ever been. You know, like, despite the competition, Excel still is incredibly loved. Then with things like Teams that have kinda breathed new life into the suite. It starts with product. We think from our perspective, value is number one, but it's not just product.
I love the people who work on my go-to-market team, who help to turn, you know, what we have into a machine, a machine of execution. In fact, if you look at our Q3 results, you know, it's not that the macroeconomic kind of conditions changed, it's just that the go-to-market team really executed for us. We have really worked hard and been very focused on two aspects of the business, both driving the Q and the P. We both drive, kind of, seek growth, and we try to drive the price up, and we have some really interesting ways of doing that as we move through, you know, E3, E5, kind of our SKU lineup.
I think you mentioned Power Apps earlier.
Yeah.
We wrote a deep dive on Azure way back when, everyone's like: "AWS has already taken off. No, no room, and here we go." Kind of an interesting opportunity. Do you think Power Apps could be the next Azure?
I'm very bullish on Power, and Power for us is Power Apps, Power Automate, Power BI, so it's all of those. We actually now have Power Pages and Power Virtual Agents. They're kind of are some up-and-coming shoots of growth as well. Yes, if I were to cut to the chase, we would simply say the world is starting to move so fast that this idea of low-code solutions or easy automation, man, it's incredibly attractive to the market. We're just getting to a really big business milestone for us. We would say we're starting to see Power tip from this early part of the market into what we'd call a majority buyer.
What used to be BDM-type buyers, where it'd be like we do pockets with an enterprise, now that buying habit has switched over very clearly for us over into IT. That's good, because IT typically will do these broad wall-to-wall type purchases. We're definitely seeing kind of an opportunity right now as the technology is maturing, and the market itself is maturing across all those different categories right now. It's a space to watch, for sure.
Dynamics?
Yeah, Dynamics. I'm very excited. People kinda give this one up for dead, but let me tell you, I've been in the midst of it. I took it over last quarter, well, two quarters ago, and I've been in the midst of it, kind of really trying to understand what's going on. First off, the products have gotten good. It's hard to take someone like a Salesforce straight on, head-on in what you'd call a static market, where the structure of the market is set. They're so far ahead, you know, that could feel like a slog. The products have gotten good. There's a lot of dissat, quite frankly, on some of the leaders, like Salesforce.
We're doing a lot of work in trying to understand the market, and there's a lot of unhappy Salesforce customers out there in terms of how they feel about the commercials and the product. The big game changer goes back to, Brent, AI. We think AI is a tremendous game changer, and whether you're doing kind of like Salesforce enablement, so you're helping your sellers be more productive, or like service centers, so call centers, et cetera, man, AI really changes the game. It has been growing nicely for us, but we really see an opportunity here for us to kinda pounce on what we think will be a new structure of the market. As the game changes, as the playing field tilts, we see an opportunity.
Do you envision, I can walk into my desk and say, "Show me the last 10 customers that bought Microsoft stock on our trading desk," and pull up their name, and I can hit an automated call feature, and I can just start calling people or send it on an email? It just seems like there's so much
Absolutely.
we at Salesforce, and it doesn't give me anything back. It sucks everything in.
Absolutely.
which is okay.
Yep.
That's where we're at with it. This, this ability to then give you back your time, is this seems like the direction this is going.
I'll go back to the beginning and say the most strategic thing you should take out of what we're doing is that Cross-app intelligence scenario. At the end of the day, your ability to walk into essentially a ChatGPT interface and say whatever question you have, "Hey, tell me the last, you know, 10 customers that bought Microsoft. Pull up their profile, tell me what else they bought, and tell me what you recommend I should do next." That nothing exists like that today. What we see essentially is people bopping around from application to application, essentially kind of trying to piece together really what are the meta goals of their job, which is: How do I serve my customers better? What should I do next? What's the next most high-value action I should take?
We really think that these natural language interfaces can change the game. We do anticipate that people will spend less time in biz app type scenarios, interfaces. We think that people will spend less time, frankly, in other apps, even Outlook or Excel as well. The race is on, we think, for this opportunity to have this interface that is an interface to your business, not to any particular app.
I mean, I'm hearing a theme of faster innovation.
That's for sure. Yes.
I mean, with AI, it feels like we may get it to a more efficient R&D organization, faster dev times, like, the innovation comes...
We definitely see that. It's not like these things come for free, but I'll use an anecdote to give you a sense of what that looks like. One of the reasons we could move so fast on this Microsoft 365 Copilot is we had been working behind the scenes on a scripting language for all the apps. Think of a scripting language that allow you to do anything you can do in the ribbon in Excel or PowerPoint or Word. The coolest thing happened in January, essentially, when we took the GPT foundational model, and we exposed it to that scripting language. Now, remember, a GPT model, what it can do is it can understand the syntax, the vocabulary, the structure of a language, and a scripting language is nothing more than a language. It learned how to speak the scripting language for those apps.
The reason that we could introduce Copilot, and it could command the apps, do this in bold, add a slide, you know, take a picture of this and put it here, was because it simply learned how to speak Microsoft App, and it did so within a matter of about 10 days. Is that fast R&D? We think so. Will everything go that fast? No, but we do think it's kind of a sea change in what's happening.
Jared, exciting times. Thanks for joining.
Yeah, my pleasure. Great to be here. Thanks for having me.