Remarks by Bill Gates, Chairman and Chief Software Architect, Microsoft Corporation
CEO Summit 2003
Redmond, Washington
May 21, 2003

BILL GATES: Good morning. I hope the Tablets are working, but I also hope they won't be too distracting.

We're going to start out with a topic that we've touched on at every CEO conference going back to the very first one in 1987, and this is talking about IT and its role in corporate competitiveness: What are some of the key issues, and what are some of the key opportunities?

Microsoft's view on this has been pretty constant throughout. When it became over-hyped, we were a little concerned about the promises that were being made during those times. At this stage, in a sense, you could say it's almost under-hyped, and a good example of that is that there are various articles that have come out. The Economist said "Paradise Lost." Even IBM, the other largest company in our industry, talks about the post-technological period. The most extreme was probably the Harvard Business Review sort of suggesting that railroads and IT had a certain similarity, and now that the tracks have been laid, there was no competitive advantage to be had from having better IT systems. The New York Times said, "Has technology lost its special status?"

Well, our view on this is that IT long ago moved away from being simply about back-office activities, simply about printing checks and keeping the account books. And, over these last seven years, it's moved to become the tool that determines whether your information workers can do their job effectively. Do they know what's going on with customer satisfaction? Are they engineering new models in a very effective way? Are they finding partners to work with in a strong fashion?

And, although a lot of that is very difficult to measure, it has been a very big challenge for IT departments to step up to these new things. Historically, the IT department knew that its equipment was all in the glass house, and understood how to deal with that. Today, it's cell phones that people are carrying around and downloading information to. It's portable devices, it's spreadsheets that people have on different desktops, and in a sense, the scope of their responsibility, and how much they should invest in making those people more effective, is something that a lot of companies have had a hard time seeing exactly what that level should look like.

I think the good news is that the advances in the technology are strong enough that, literally for the kind of investment people have been making, they can get best-in-class, very exciting advances. And so, we don't need to be even at the peak levels that existed during some of the more ebullient years, and yet, cleverly applied - being on the forefront and getting a lot more out of the very substantial investment that's made in the workers themselves - then that's achievable.

Our industry, of course, benefits from the so-called Moore's Law, the doubling of power every two years. The key investments that drive that forward have not slowed down, despite what's going on with the economy or IT spending. If you ask how fast will the chips be over the next six or seven years, it's that same exponential increase. How large will the disk be that's connected up to these machines that are now typically 20 or 40 gigabytes? Those will continue to grow in size, at the same time that, actually, the cost of those are coming down.

There's been an interesting crossover point that we've been looking forward to for quite some time, and that's the point at which the high-volume, so-called industry standard, machines that use Intel or Intel-compatible chips: When will they have the performance of the more expensive, proprietary type systems, which have been lower volume? Classic mainframes are very high-end UNIX type systems.

We've passed that in terms of price performance a long, long time ago. In fact, that's sort of the business model of the industry standard offering has implicitly because the R&D cost is spread across so many units, millions of servers and over 100 million desktop machines. Whether it's chip R&D or software R&D, it's really a very different world.

The one footnote to that, though, is that if you wanted the absolute highest levels of performance, there were still some things that these devices couldn't do. A few years ago, for all but the applications that had to run on a single system, we reached the crossover. That is, in any application, like a Web front-end where you could use multiple machines and split the task up, the industry standard price performance and absolute performance was in the lead.

More recently, as part of the launch of our new Windows Server 2003 product, we - together with HP - showed through a wide range of industry benchmarks, transaction type benchmarks that actually, even in the most demanding case where the database or another application has to run on a single system, we have passed over.

So, this intersection point that I'm showing here on this timeline - we are now just past that crossover. That's a nice thing, because it means in terms of simplicity, what you have to have - development models, development tools and just plain the price of the equipment involved - now, it's not just some applications that can ride this curve, but it's everything that you're doing.

So, that's a very interesting milestone. In fact, one of our top researchers, Jim Gray, who writes about transaction systems and the great things we're doing, wrote a paper saying that it's a zero-dollar business over time, because as the hardware gets more powerful, the hardware prices go down. It is a true map mathematically that at least the hardware piece, that the price approaches zero. It's not at zero yet, but that kind of performance really opens your mind to thinking about using these transactions and using these rich systems -- that the trade-offs will be very different than they've ever been.

Let's look at the different eras that we been through, going back really to the Internet explosion. By the time of the Internet explosion, which was '97 - '98, PCs were in place by and large, but they weren't all that well connected together. It was really this period, '98 to 2000, where we saw the explosion of a number of things, and, coupled with this, of course, we saw IT spending in absolute, as a percentage of capital spending, achieve record levels.

The Internet came along, and there was a lot of discussion of what that meant, some of it in retrospect overly optimistic, but it did mean information accessibility. It did mean that companies needed to drive to have Web sites to do things very rapidly. It meant that some of the ways business relationships were handled were becoming very different.

It was during this era that the number of devices, PCs and servers really exploded, and it was just necessary that if you had a PC for every information worker, if you had servers for all these tasks, you were going to have these large numbers. E-mail exploded, almost became for most companies the standard way that people share information. The number of vendors of software was very large, and because people were in a rush to get the best pieces put together, they often found themselves with literally dozens of software suppliers. And integrating those things together, understanding how to work across those different things, this was the era where it became common sense that enterprise applications - particularly SAP's R3, but also some of the others - that everyone would move to use those as the foundation piece, instead of writing in-house software to do those things.

And so, there were huge benefits. This was an era where people were moving very rapidly, and certainly well over half of what people expected to get out of these things came to pass.

We've gone through a period that I'd say started March 2001, if you use stock price as a nice demarcation, to when the mentality kind of shifted. We've gone through a period where the view of looking at this, the glass that's half full, and seeing what are the harsh realities of what's really not there, and what should be done better: that's been the dominant theme.

So the Internet bust, you can measure that not only in financial terms, but think about what people said about B2B exchanges, where there'd be these middlemen organizations, and all the business would flow through those things. I think of the several hundred of those that were put together, only two or three have found enough value-add that they still exist today.

That vision of e-commerce is still very important, but it will be done without hubs. It will be done with each business essentially being its own hub, and being able to reach out to its suppliers and customers using software to get all the benefits that the hub was supposed to provide - the ability to find everybody out there, the ability to map their data formats into your data formats - and so there really doesn't have to be any friction, any middlemen, in that.

But those dreams were strong. People have looked at the complexity of IT management, all these different packages. You know, do they know if their systems are up to date? Do they know that they can say there won't be any security problems that come along in the systems? Even e-mail, as wonderful as it is, some of these things like spam are now at pretty unbelievable levels. And it's not just spam against consumer e-mail accounts like Hotmail or AOL. I'm sure many of you have noticed that spam has come into corporate e-mail domains as well. And e-mail can often be a distraction unless the tool is very good, and people are smart about using it. In some ways, the time benefits you get, a lot of that can be thrown back away unless it's done in a very clever way.

And finally, the idea of those enterprise applications, as good as they've been: the panacea that you'd be able to really dive into data in a way that's meaningful to manage, or get down to any level of information and run complex processes like sales analysis and forecasting in the most effective fashion. That dream has not been realized.

And so, many people looking at these harsh realities sometimes say, well, this IT stuff, it's messy. Let's outsource all of this. Let's get somebody else to do it. They can get the benefit of Moore's Law, and we'll just sign a five-year or 10-year contract that drives that outside.

Certainly, for some parts of IT, which are very measurable, repeatable type things, there is validity in that. But we're from the camp that says when it comes to defining new applications and thinking about business processes, IT is so central to the way work gets done and the quality of that work, and there are so many opportunities to do that better, that staying in control of this to have it as part of the overall business strategy is very, very important.

Now, the industry saw these harsh realities not just in these last few months. The issues of reliability, cost of ownership, security - those go back three or four years. And so, what we're seeing in the products that are coming out this year and next year is how the industry has responded by making investments to deal with those things. And this is where I think in some ways people are really underestimating what can be done. It's kind of natural if you overestimate what an industry can deliver, and then that you cycle back to where you underestimate those things. But I think we're on the verge of particular software advances that really address these harsh realities.

For example, e-commerce. Many of you have heard about XML Web services, and although that's a fairly technical thing, it's a fundamental thing, because it's the infrastructure that allows companies to exchange information for buying and selling and collaborating without the two IT departments having to build special applications that only relate to that one particular relationship. It's a general way that no matter whose software you use - whether you use IBM's or Microsoft's or someone else's - as long as that software adheres to these Web services standards, that ability to buy, sell, collaborate is available. And not just in the trivial sense of taking the paper invoice and sticking it in e-mail, but in the deepest sense of how you find those partners, how you do secure exchanges with them, how you track the workflows, so that if something is delayed or something comes in that's not right, the electronic path is the effective way of dealing with those issues.

As people move to partially do e-commerce, in some ways it was even more complex because you had the straightforward information passing electronically, but all the exceptions would result in phone calls and faxes and e-mail, and having the understandings that were created in parallel in that knowledge worker side, and getting the backend systems to understand that sometimes the impedance and the mismatch there even took away the benefit of having a piece of it be electronic.

The idea that when you have these Web services, that you can capture the full richness of what's going on with complete visibility to the knowledge workers, to update those things and be notified appropriately of things, that is where you get real benefit of saying that the paper approach really is completely obsolete.

Web services are something the industry started on several years ago. We bet our company on it in the year 2000, calling it our .NET strategy, and there's been great progress on that, in fact, lots of interoperability between the different software stacks, the key specifications all put together, and now we have pioneering customers doing very well with that.

The second issue is about managing all these things. The IT department generally measures its difficulty in handling things as proportional to the number of servers or proportional to the number of desktops. That really shouldn't be the case. I mean, the software should automatically keep things up-to-date, and if you say apply this new version for all of these workers the software, the network should make that happen. So things shouldn't be proportional to those large numbers.

Well, that takes a new generation of software. It didn't exist on those earlier systems because the numbers just weren't large, and so having manual involvement was not impossible. Here, it is impossible. The vision has been articulated by many companies in the industry. Sun talks about M1. IBM talks about autonomous computing. Microsoft calls it the Dynamic Data Center. And it's really the same thing; it's software tracking the system, and IT only having to see the exceptional cases that can't be solved through that software intervention, so none of the efforts having to be proportional to the large number of systems out there.

Another key thing is this idea of Trustworthy Computing. How can we make these systems so that virus attacks or software holes are extremely rare, and that recovering from those things, even when they do happen, is very straightforward, where you understand exactly what needs to be done to reset so those aren't a problem. This has required a lot of invention by the industry. It's required a new way of doing software development, new testing tools, and the progress here has been even faster than I would have expected.

This is not a completely solved problem, but the rate of improvement, the degree of improvement, the monitoring tools, the ability to audit and say, is somebody following the best practices on these things, has improved very dramatically in the last few years, partly because this became the top priority for the industry. We took products like the Windows Server I mentioned earlier and we actually stuck an extra six months in the schedule because we knew this was such an important thing.

E-mail: People love to complain about e-mail, but if you say to them, OK, we're going to take your e-mail away, then they'll complain even more, so it's that love-hate relationship, and the key is how do you take and do things that get the love part to be preserved and take away all those things about the distraction. And I'm going to show you in this next generation of e-mail products how we think that's been achieved.

Another big advance is the understanding what's going on with these systems. Microsoft four years ago had no visibility when people were using PCs and things wouldn't fit together. If a device driver didn't work, if it was hard to install, if software conflicted with each other, we didn't have a real picture of where those mismatches were.

Now the software has built into it reporting-back facilities, and so for anybody who's willing to let those reports flow back, which 70 percent do, we see those things, and that drives our priorities in terms of who's got to work with these driver makers, who's got to work to make these systems more resilient, having a total picture of what people are frustrated about, what's not working for them, which we call our Watson Initiative. It's been night and day in terms of being able to understand exactly how engineering resources should be applied, and being very numerical about what the PC experience or PC server experience is like.

And finally this issue with these silos of information, serendipitously it turned out that the architecture, this Web services approach that is necessary for e-commerce, also is necessary for moving information even within the company, and so that one approach, we're getting massive payoff for that. In fact, we're using it for systems management, we're using it as the basic connecting architecture for everything that goes on. And so the leverage there, because now we have tools there, we have broad industry understanding, it means information flow within a company, across company boundaries, and then the specific IT issues are dramatically advanced because of that one architecture.

So for any company, our view is you have to think about how software, the software strategy, information sharing, the rich applications that you license and that you create, how those affect the four things I'm showing here: How the people work together, how the processes work, and even simple processes like expense report management or the human resources review type process, is all the information really there in the simplest, most effective form? As you talk to somebody about their headcount budget, is it easy for them to immediately go and see how that's changed and where things are versus forecast? Very basic processes that, I'd say, the state of achievement in most companies is way, way less than even commonsense would dictate that it should be. Information about the market and what's going on; there also people who haven't gotten things that really would help them drive their processes.

Web sites today, although they have lots of information, they're not gathering as much customer information as they should, and that information, the way it flows inside the company, is not a simple as it should be.

Creating relationships where you think about key partners and how you do work with them, that's another thing where partly because the security infrastructure hasn't been in place, partly because the Web services things weren't there, people are only getting a little bit of benefit once you've crossed that corporate boundary.

So all of these things tie back into whatever the core business goals are, and really require a strategy focus.

One of the accounts I work with set goals for their IT budget, and unfortunately it was decoupled from the idea -- this is an auto company -- was decoupled from the key initiatives in their engineering group to get the design cycles down to be 30 percent shorter than they were. And when you get that kind of decoupling, of course the IT people aren't going to do the things, some of which are very modest in expense, you know, the basic expense of the PCs and the network are there, the idea of building these sharing Web sites and getting people able to use them, it's about a 5 percent increment on top of just keeping that normal infrastructure in place. And yet, because the investment decision and that strategic goal were not tied together, the engineering initiative thing was delayed, and may not happen in the timeframe that it should.

Jack Welch talked when he was at the CEO conference about how do you get competitive advantage. The people you're hiring come out of the same pool, the techniques that you're using are not that proprietary, and so he talked about this information advantage being the key thing: understanding customers, understanding competition, moving at a speed that you take that information and translate it into action.

And so every software system should be benchmarked by exactly that kind of criteria. And so when somebody says, to take the extreme quote from the Harvard Business Review article, they say IT doesn't matter, they must be saying that with all this information flow, we've either achieved a limit where it's just perfect, everybody sees exactly what they want, or we've gotten to a point where it simply can't be improved -- and that's where we'd object very strenuously.

We have seen in terms of new technology a lot of waves come along: the original PC, where you had the development and you had big adoption; then we had Windows, which was a wave like that. I talked about the Internet wave; really it was the biggest, because it moved out from just the individual to the way that organizations worked. And we have some of these waves that we're just at the early cusp on, things like wireless networks. Of course, these Tablets are connected up to a wireless network. Everywhere at Microsoft, that wireless network is in place. And so when people go to meetings they take their PC, they have the latest information, they can write their notes, share those with other people, check information, so we just kind of take it for granted. Wireless networking: one of the beauties of this Wi-Fi is that it doesn't have any ongoing costs. Once you get it into place, other than maintaining it, it's there, it's simply using spectrum that's available to everyone, and so as an increment to the base network cost that you have it's not that significant.

So that's just catching on, catching on in the home, catching on in the places that people travel, so-called hot spots, catching on inside corporations. I'd say about 20 percent of corporations have done what we've done and made it something that's pervasive in terms of what they do.

If you look at the investments that are made in IT, of course it breaks down into some very significant buckets. Licensed software tends to be, even if you take all the different applications, on the order of 5 percent or so. The biggest expense -- and the ones that really new software advances will be measured against how well they do -- is how they can help with these other pieces, letting you take advantage of low-cost hardware. I'd give the industry a very high grade for doing that. Letting you rationalize your network cost so that you can benefit from the price declines that are taking place there; a good example of that is that now we allow people to take their mail servers and put them all in one place, you only have the mail servers in one location for a global enterprise. That makes it a lot easier to pool those things and to have the IT expertise, as opposed to before where, because of network delays and sometimes reliability issues, the software didn't get around those problems, so you had to have the e-mail servers out on a very distributed basis.

Making the IT staff not having to visit desktops so that, as they diagnose issues, they come to them in a very high-level form, and making sure whatever IT services get used are not applied to writing essentially glue code that just connects applications together. I think that we'll see very substantial changes in how these investments are made, but the enabling factor is the advances that come in that platform and application software.

Because of the volume economics, literally billions, if you take the aggregate R&D of these things together, it's about $40 billion, of which Microsoft would be about $6 billion of that. The advances will come very rapidly here and allow for either cost reductions or more effective use of the other parts of those investments.

Well, let me now switch and talk about some of the particular things that drive our excitement and optimism about how people are going to work, and how that's going to change information flow inside a company.

Communications today is nowhere near the ideal. Your multiple mailboxes and phone numbers, and the different devices and people trying to get a hold of you, a lot of integration could be done here, and that means integration across the different devices, integration between the PC and the phone and allowing the PC to be more than just e-mail, letting you communicate in real time and share and do things together. We're investing very heavily in this because any improvement to communications has this very dramatic effect. Knowledge workers spend most of their time in communication activities.

The idea of collaboration, sharing information, this is another area where the choices have been pretty limited. Web sites are very hard to build. If somebody in the office says, OK, I want to make a new Web site, they have to go to IT and get it approved, they have to use complex tools, so they're not likely to share that way. Sharing files -- all you get is a list of files up there. And the final way of sharing, that's the most common right now, is just doing enclosures in e-mail, but that doesn't let people see the different versions, your e-mail gets flooded, you have different people working in parallel with documents that may be out of date. Really what you want is that Web site, but you want the Web site so that anybody can just sit down and create one without having to go to IT, without writing a line of code, and pick a template to choose for the Web site, and then easily customize what they want to create on that. This kind of sharing and collaborating is a big step forward for us. We call it using SharePoint.

So, just to give you an example of some of these communications tools, and how the PC will be viewed in a different way in this, I would like to ask one of the people who has really driven some of these products, JJ Cadiz, to come out and show us real time collaboration.

Good morning.

JJ CADIZ: Good morning. Thanks very much.

It's my pleasure today to be able to show you PlaceWare, which is a technology that you can get in your company today to drive a lot of improved productivity by enhancing telephone conference calls you may have in your company, but also to reduce costs by reducing the amount of travel that happens in your company to talk between distributed team members.

So, let's go and take a look at PlaceWare. What is PlaceWare? PlaceWare is a Web-conferencing application being used to basically take any Web browser and telephone to have meetings online. And when we talk about meetings online, we're talking about meetings with remote colleagues, with customers, and also with partners. And when I say meetings, I'm not talking about just the small one-on-one group meetings that are out there, but also huge group meetings. So we're talking about meetings of up to 2,000 people. So a lot of times those would be meetings of a whole division.

Now, I'm actually using PlaceWare to talk about PlaceWare here. So in the main screen, what you see here is PlaceWare, and we're actually logging into a room that I've created for this demo, and in the lower right hand corner you should see Bill is actually logged into the same room. That way we can give you the sense of the two different rooms, and how they look, and also how quickly things update.

Let's talk about the basics of PlaceWare. Now, as you think about meetings, there are kind of three major stages of meetings. There's the before-meeting, the during-meeting, and also the after-meeting. Now, the way that you can schedule meetings is very simple. You just need to go to a Web portal and schedule meeting, a PlaceWare meeting, when you schedule a meeting with Outlook, or you can also go and use the Outlook client to do that as well.

Now, when it's actually time to enter the meeting, you can do things with Messenger, you can click a link within Messenger to go in there, you can go into the same Web portal to do it, or you can just go to an e-mail invitation, which is probably the major way that most people get into PlaceWare meetings. Let me show you an example of one of those e-mail invitations. I have the ones here for the demo that I'm showing you here. And all any employee would have to do with a Web browser is just click this link, and they'd be able to get into that meeting.

Now, if I actually wanted to show someone an example of this e-mail, you'll notice that in the lower right-hand corner there, you aren't actually able to see this e-mail, all I would have to do is hit the snapshot tool here in PlaceWare, position it over the e-mail, and then just take a picture. Let's go ahead and do that here. Now, so we have that within the PlaceWare client, and then all we have to do is say, here's the link you need to click. You can just kind of ignore these other links down here. So you can do things like that.

Now if I actually wanted to show someone a live demo of something going on, let's say we want to have an Excel spreadsheet that we're working with, OK, I can actually just bring it up here, and then do an application-sharing slide that we can do with PlaceWare. I'm going to do that. So, I just position the frame, and then press Play, and this entire anything that's in the frame -- will be shown up in the lower left-hand corner. Now, the cool thing about that is that I can, of course, have my e-mail client over to the side here and be doing things, knowing that it's not actually being shared with anyone else.

I can also show live technology demos, saying here's the way you can modify Excel pie charts and such.

The other thing is, I can also give other people control of this spreadsheet. So, if I wanted to, and someone else had newer numbers up here, they could go there and I could give them control to modify the numbers within here.

Now, PlaceWare also has a bunch of other tools. And one of the kind of most fun tools we can do is ask questions of people who are in the room. So, let's say that I have all of you in this PlaceWare room here, and I went to ask you a question. All I have to do is press the polling slide, and I could ask you a question like, how about, what should Microsoft's next dividend be? So, I could say zero, I could say the same thing it was last time, or a little wild thing, what you do think, $1.50. You can vote here, and you get a live update of how people are voting. We can look at that person's vote, so we can see how people are voting from the presenter's view over here. So, polls are very easy to create. And the other cool thing is that all of these poll answers are kept for me later on. So, actually the presenter will get all the poll answers later on. So, you can imagine even being able to use this to grade students if you had them in a class in PlaceWare.

And I also get a list of all the attendees who showed up. So that way, if I knew 100 people were supposed to show up, and only 95 did, I could actually go back to those five people and say, hey, if you want to look at the recording, it would be great if you could do that, because with PlaceWare that's as easy as bringing up the recording control panel and pressing start here if you want to record any meetings. Now, the other nice thing is that the meeting rooms that you use within PlaceWare all persist. So, if you want to go back and visit them afterwards, you can do that.

So, PlaceWare is something that we're very excited about here at Microsoft. The last fiscal year we actually spent over $3 million on tools like PlaceWare to help our employees cooperate, and now that we have PlaceWare, we're using it in a variety of settings, including regular meetings between employees and field offices, and employees here at Redmond. We are using it to present to customers. And even at one of the last board meetings, where not all board members were able to attend, we actually had them attend remotely using PlaceWare, so there are a lot of great examples of methods that you can use PlaceWare for out there.

Now, what I would like to do is change gears and switch to a second demo that also has to do with real-time collaboration, and it has to do with a real-time collaboration tool that all of us are probably very familiar with, and that's this thing sitting right here, the common telephone that all of you probably have on your desk. Now, the telephone is a very powerful real-time collaboration tool, but it doesn't right now interact at all with the PCs that we also have sitting on our desks.

So, the enhanced telephony demo is all about how can we help bring together PCs and telephones to create a powerful user experience for people. So, I'm going to bring up the Enhanced Telephony client here. And the way we have this set up is that integrating with kind of existing systems, so we're not trying to turn the PC into a telephone or replace existing telephone systems, but rather integrate the systems that are currently out there.

So, let's get the music going here, that will be part of the demo. And when we talk about helping the PC and the telephone to be better together, what do I mean? Well, I mean kind of two specific things I'm going to show you, and the first thing is that it should be really easy to dial someone. Any time I see someone's name or phone number on the PC screen, I should be able to dial it. And, of course, it sure would help me when I'm receiving calls as well.

Let's go ahead and do the scenario where I want to call Bill. So, here's my favorite list of people to call, and so I'm going to call Bill. Now, what just happened there? First of all, notice that the music isn't playing, right, so the PC knows that I'm using the telephone, so it can mute the music any time that I'm using the telephone. We have a phone set up here to be me, and we have that phone set up to be Bill. So, when I click the link, what happened is that the phone went off hook, it actually went to speaker-phone mode, and then dialed that phone sitting over there.

Now, once I'm in the phone call here with Bill, you know, I can take notes here, you can imagine these being ink notes if it's on your Tablet PC over there. I can so do easy conference calls and transferring of calls because the interface on the phone right now sometimes kind of makes that difficult. So, when I transfer calls I forget how to use the transfer button on the phone here. I can share my screen with Bill using PlaceWare type technology. And, then once we're ready to hang up, we can actually just press the hang up button here, and then we go back to where we started from.

The other cool thing about merging together phones and PCs is I can allow people to search across the entire corporate directory as well as their personal directories for phone numbers. So, if I went to look for Laura C., not only would it show the Laura Cadiz, who is in my personal address book and all her phone numbers, but also all the Laura Cs in the Microsoft Corporation, OK, and I can dial them just by clicking them.

That's how we make things easy to dial, and how to make people easy to dial. So, again, imagine anywhere in the operating system where you see a person or a phone number, you can just click it and dial it.

Now, let's talk about the incoming calls here, all right. What happens if Bill calls me. Let's go ahead and do that. Okay. So, there are two different things that are going on there. One is that you might have heard faintly in the background that the PC can actually act as the ringer. The other thing is, notice that I didn't actually answer that call. And because I didn't answer that call, ET automatically sent me an e-mail that says, hey, Bill Gates just tried to call you. That way, if I wasn't sitting in front of the exact PC where I was running ET, I could tell that people were trying to call me.

The other cool thing is that, notice that the only thing coming out of the phone here is a normal phone cord. There are no other actual wires. That means I can actually receive notifications of people trying to call me no matter where I am. So, let's say I have the common cell phone sitting here with me, and I'm sitting in a Florida hotel room like I was a few weeks ago, and I get a notification of an incoming call. I can actually set the transfer button here, I can attach this to my cell phone, and that way even though I'm in a completely other state, and I'm not next to my phone I can still take calls. Now, because ET can also programmatically transfer phone calls you can imagine setting up a variety of rich rules, for example, things like any time I'm not at my computer forward all my calls to my cell phone during business hours, and during work hours rather during home hours, go ahead and forward them to my home phone, but only if one of my coworkers calls me, not if any other random person calls me.

Now, the cool thing about ET, if we could go back to the main slide, is that ET is not actually just a demo that we bring out on the stage every once in a while, it's actually something that's deployed throughout the entire Puget Sound Microsoft campus. We first made it available on January 9th, and only told about 150 people, and didn't do any advertising, and we relied on just word of mouth to spread the word about ET, and right now we have over 5,000 people who have installed ET on the Puget Sound Microsoft Campus. It's also an excellent example of a technology that can be available using .NET, over 90 percent of ET is built completely using .NET.

Now, the feedback on ET has also been very good from our users. We've had a few unsolicited feedback emails that people sent us, saying a variety of things, like just because ET is able to make me more available it saves me often on several big issues. And actually, I especially like the second quote where Alice talks about the fact that one time she was sitting in a meeting, not next to her phone, and saw that one of the managers tried to call her, and she wasn't able to excuse herself, but she was on e-mail, because of her laptop, and was able to resolve the manger's issue there immediately.

These are two examples of real-time collaboration tools that we're working on for you, and the enhanced telephony demo specifically, that's working right now within Microsoft. We're going to need a lot of help from companies represented here in this audience to make it a reality for everyone else, and we very much look forward to partnering with those companies.

Thank you very much.

(Applause.)

BILL GATES: It's interesting, that getting the phone and PC to work together; a lot of people thought about that as requiring you to change the whole telephony infrastructure to work across the Internet, so-called IP-based approach. But it turns out that that example is a traditional PBX, it's a fairly simple piece of software, that talks between the network, the computer network, and the PBX network. And so even without changing out any of the existing infrastructure you can start to get these benefits. Likewise with PlaceWare, what most people do, because the Internet telephony, also called voice over IP, isn't high enough quality yet, they're placing a traditional phone call in parallel with that screen connection. And so they're the best of both worlds, they're getting the screen interaction, and yet the voice quality and all that is the same. And yet, it all gets set up, when you click to join Net Meeting, in one simple step.

I'm going to quickly talk about some applications that our IT group has built. This is along the lines of... there really are those things that were exploding in the late '90s, there have been some really concrete benefits that have come out of those things, procurement benefits, purchasing all purchasing being done electronically, making that a paperless process, invoicing a paperless process, expense reports a paperless process, where you can see the history of what somebody has done on expense reports, and simply approve those things electronically. This is pretty straightforward stuff, you can see the time between starting the project and rolling it out is in these cases on the order of six months. The actual volume of these systems, because the servers have so much capacity, the idea of handling 400,000 orders, or a million invoices, or 200,000 expense reports, a single server is not overloaded running those applications, and that's a $20,000 piece of hardware to do it. So really the software is the key element here. I'd expect if we surveyed your companies, about half would have done what's described here, that is, created a very straightforward user interface, and moved these processes into a paperless fashion. These other two examples I don't think we'd see quite at that same level of usage at this point.

We have another thing that's been very important for us, which is taking all our information about customers and bringing it into one place, the Siebel information, the SAP information, the Ad Hoc information, and letting you navigate that in one way. This is a classic problem, and it's one where the Web services approach is very apt, because every one of those major applications lets you pull the data out, and then create the user interfaces that make sense for the various people. And depending on their role, that user interface is going to be very different, you want to be able to refine that, you don't want that locked in by somebody outside the company, you simply want the data sourcing out of their rich applications into the single user interface front end. And so I think, although only about 20 percent of companies have done this, this will become pretty standard stuff.

We've also taken most of the processes and said, OK, we're going to create a SharePoint portal for every one of these processes, the task activities where they found savings opportunities by seeing what could be done in different venues that were pretty dramatic, paying contingent staff, employee self service, we had a lot of paper forms related to these things. One way to find the opportunities is to go where the paper is, and although that's been said over the years, I think that's still really the best way to see where things can be done better.

Now, these are administrative systems. The actual biggest benefit to us, the collaboration software, is in our product development process. That's harder to benchmark, because we're doing unique products, it's not like there's some other company there. But, our use of Web services and sharing, the impact of these typical back-end systems I'm certain here is outweighed by more than an order of magnitude by how it relates to our basic product design, product testing, product collaboration, on those different product issues.

We mentioned that the Tablet is something that we've embraced pretty extensively. The Tablet launched late last year, I think the first unit shipped in November. Those of you who were here last year got a chance to use the early Acer prototypes; actually that machine is now moving to a second version that's even faster, and a really very fine piece of work. There's an HP Tablet that has a convertibility thing that's really great. There's a Toshiba, sort of a no-compromise in terms of its rich portability. There's a Gateway machine. There's quite a few of these with different design points in mind.

We've got over 5,000 now, and if you go to a meeting, it's typical now, people are taking their notes, sharing their notes, we don't even bring Power Point print outs to a lot of meetings, because it's a lot better if you have the PowerPoint where you can take sales figures, and dive in and say, what was that by country, or time period, or product, which if you have it on paper, if somebody wants more detail, maybe you anticipated that and slide 200 in the appendix, you say, aha, I knew you'd ask that, I have the Japanese expansion on slide 200, so you flip to the paper. So that means you have to hand out something that anticipated every possible question that the executive might ask. It's a lot better simply navigable in a live presentation that's connected up to that data. So there's no doubt we are being a pioneer on the Tablet. There is no other corporation that has 5,000 Tablets in use today. And yet we see this over the years to come. But, all portables work this way, and even more of these machines will be portable type devices. You've probably seen the speed in terms of reducing the thickness of the device, and making it an attractive reading device has been pretty substantial there.

I thought I'd do one last thing, which is just to give you a sense of how I work in a typical day, using these tools; what am I doing sitting in my office. I've got here a pretty nice system, this is a 23-inch LCD, you can see it's got an interesting aspect ratio to let you see lots of information across like this. This is still a pretty expensive display, they're about $2,500, they've just come out, it's a Sony display. These will come down in price over the next three years, we think, to about a third of that, to $700 or $800. So even though today, maybe only the executive staff should have these things, these are going to be commonplace. You've probably seen a rapid shift in your company from CRTs and desktops to LCDs because the 15- and 17-inch LCDs are already at that $600-700 price point. So we're reached crossover where for any new system the LCD is superior, partly because of the text readability, partly because it requires less desk space. But as those LCDs get larger you'll see a couple of cases where that extra screen area really is very helpful in terms of that productivity you get out of it.

So the computer is getting smaller, while the display is actually getting bigger. I'll just go into e-mail... I'd say that of my time sitting in my office, that is, time outside of meetings, which is a couple of hours, two-thirds of that is sitting in e-mail. E-mail is really my primary application, because that's where I'm getting notifications of new things, that's where I'm stirring up trouble by sending mail out to lots of different groups. So it's a fundamental application. And I think that's probably true for most knowledge workers, that the e-mail is the one they sit in the most. Inside those e-mails they get spreadsheets, they get Word documents, they get PowerPoints, so they navigate out to those things, but the center is e-mail.

So here you can see I've got a lot of different e-mails. I'm in my Inbox here, and you can see I've got a bunch these are different folders here, these are actually the e-mails, and I'm using this three-pane view, so that when I navigate to different e-mails I'm automatically seeing the e-mail there. And that works, because I have this big screen, so I'm not having to go in and out of the e-mail a lot, I can navigate around with a lot of speed. Now, let's say, I have an e-mail here, there's a classic question when you get e-mail, is it something you can just read and say, OK, I'm done with that, that's very satisfying, you just delete it, and never think about it again, but often it will be something where you'll send a reply, and expect a follow up, or you think, oh boy, I'd better read that more in-depth. What you want to do is just flag it, you'd want to just right-click and pick some kind of flag, I picked a blue flag there, and then have that indicated, and then be able to see all the mail that you've flagged. So what we have here is a thing called For Follow Up, and it sorts it according to which things I put into the different categories, the blue, the yellow, the green, and that's just a view on the e-mail. I didn't have to move things into different folders. Moving things into folders is a lot of trouble, because then you have to navigate all those folders to see what's in there. It's easier if they're all still available in one place, and yet you just do this simple flagging. Although you can also move them down into those different folders.

One thing that you often want to do is look for your unread e-mail. Typically, if I see something and I think, gosh, I don't have time for that, I remark it as unread, but then I want to navigate those things. Here we just have automatically a place where you can click and see all those unread items, and navigate through all the different things that are unread, and decide exactly what you want to do about those things, do you want to forward them off to someone, or what might make sense. And you can group it, you can do it by time, you can group it by the people involved, that's a pretty interesting thing. You can see the way it's displaying this e-mail, it's a lot more terse than it's been in the past, and the reason is that it understands that this is the mail from today. I hope the font is big enough to read that. So this group here is mail from today, so it doesn't go and put, take up a lot of space, describing that, it's just the mail from today, this is the mail from yesterday, this is the mail from Monday. And so when you squeeze these things down to use less space, it knows that it doesn't need to display that date information, because it's today, you just need the time information, and that's why you can see it in such a succinct fashion. And having things organized that way is pretty natural. I think, OK, that was yesterday, and if you go down here you'll see Monday, and then it starts to group it into bigger things, last week, so this is everything that's in that group. And then you can select all of those different things.

One of the problems you get with e-mail a lot is, there will be an issue that's a controversial issue, and so one person will send that mail to 20 people, saying, we should do it this way. And somebody will respond and say, no, we shouldn't do it this way, we should do it this other way, and then somebody will decide we should add some more people to this discussion, and people start disagreeing. And say you've been gone in a meeting for three or four hours, you can come back and your e-mail box can have 20 messages that are all on this one thread, where people are pingponging back and forth. Maybe people try not to expose the CEOs to that too much, but that is life in e-mail, that you get those things going back and forth. So sometimes what you want to do is collapse things so that it's mail that all relates to each other, you just want to see it as one item, and see a conversation. And so you can just use this view button here, which is what you use to control whether you have this right pane, you can just say, do I want to group conversations together? And so you can have different folders that you've set up to show things that way. So here we have people talking about a hot controversial topic, user interface discussion. And instead of seeing the individual mails you just have them so you see the most recent, and then you can go down and just see the ones you haven't read, or see all the different ones that relate to that. You can see this topic, people going back and forth, an immense amount, and yet you can just see the most recent, you can just see what's there. So having conversations as something that e-mail understands is very helpful.

Another thing that's painful, has been painful, is dealing with calendars. Here I can see my calendar, I can see that as seven days, five days, today's calendar, but I may want to see other people's calendars. Now you can just click on the different people and it will bring in their calendars, and you can see the different things they have going on. It uses different colors, so it's kind of nice. Now the system can automatically try and find free dates, but, in a way, you often want to see the calendar; assuming they've marked it, you can see their calendar. They can choose either to share nothing, they can choose to share busy, or they can choose, in this case, these are workers who have said, at least for this colleague, I'm willing to share the whole calendar. So you see that all there, side by side. And you can actually take events, like I can take the CEO calendar and publish it, so it will show up here as one of these things I can show up side by side, or you can take your regular executive meeting calendar, and have that be one of these calendars, so you can always see that, and either copy it onto your calendar, or have it there by itself.

Junk mail is a big deal. We do have a lot, but we're getting smarter at automatically moving things. And you see here, there's this junk mail folder; things are automatically put in there. Right now the junk mail people are getting smarter as we get smarter, but we have a few tricks that we don't think they'll have any counter tricks to get around. So that's something we can't defeat, even though it's at a pretty high level right now.

JJ talked about being able to call people, or connect up a PlaceWare. Any place I see a name, I can click on that name and choose whether I want to call them, send mail to them, do a PlaceWare with them, any one of those things is an option that I can get into. Another complaint that we have about mail is the fact that when you send mail out to people, it's very typical that somebody gets something, and thinks, well, this is kind of interesting, I think I'll forward this on to somebody else. And then they think, well, yes, it's really interesting. Let me forward it to someone. And sometimes when we send mail, we have the sense of, well, I might as well publish it in the newspaper the way this stuff gets out there. What we've done with mail is, when you create a piece of mail now, what you'll be able to do is pick a mail template. So, you'll have a mail template, say, for attorney-client mail, or mail that's only for people who see the early financial results, and so you pick one of those templates. And what it's doing is controlling that mail; you set up the templates to decide who can receive it, can they print it, and you can also make it expire. So, if you send information out that, say, gets out of date, you can tell it that, OK, after that date, the information is no longer viewable. And so people know to go and get the most recent information instead of the thing that's out of date.

So, this information-rights management, where you have a sense that you can control the spread of that e-mail, and the software is helping you do that, we think that's a fairly important thing, because otherwise people think of e-mail as something that's just simply got no boundaries. E-mail is the biggest thing.

The other thing that I do is I go to these shared Web sites. Each of our product groups or project groups will have a Web site, and at that Web site you have information about the documents those people work on together, the calendars that those people have, and so it's a way of working together. And it's a shared Web site that's got all the different information that's available up there. These are the things I talked about people being able to set up on a very straightforward basis. They typically this one here, you know, is a CEO thing, you can see the calendar is here. I used the standard template for this one, if I ask for the contacts, it's just going to show me all the different people who are attendees to the meeting, so that comes up pretty well. I can put so-called Web parts in here, where you have information from business systems. This is too small, I should grow the text up. So, here I have stock prices from various people here at the meeting, and it's live. I hope that doesn't shock anybody. And I've got live news coming in from MSNBC. And you can have these Web parts that, say, connect to your project management system and show the current schedule for things, or show the sales results, are they at forecast, below forecast, and different Web parts belong on different pages that people have.

I've got a document library, and people can check things out, and check things back in here. Let me just finish up with two quick documents that are in here. One is a little spreadsheet that I have, and I just created this, it's almost a kind of a humorous thing. But what I did is, I created a spreadsheet that connects out actually to Amazon, and looks at which books are important. Amazon actually has one of these Web services. And so you can go in, and for any book, you can ask how well it's selling. And so here and it didn't take long at all -- I created a spreadsheet that shows, for people who sell computer books, which Microsoft is one through Microsoft Press, exactly what the best-selling books are, and exactly what rank they are. And so because of one of these pivot-table things, I can just take this and I can restrict it to a different publisher, or I can just go back and show everybody, and sort this in different ways. And so it's a live view. Typically this is used for sales-type data. Here is the pivot table that just shows who's got the most top-sellers. I can click in to drill down on that.

I've also created a live connection where I can type in any author and see how well their books are selling. So, I can type in...my books are really old, so I'm sure they're not selling that well. I just click Go, and what it's doing is, it's going out to the Amazon web site, and seeing everything that's got my name associated with it. I just have the ones that I directly authored, and I see here, OK, I've got one that's the 138,000th most popular, 88,000th, oh, 400th, that's not too bad. And honestly I didn't go out to buy any this morning to try and make this thing work better.

And I can even do some comparative things here. Let's see how Michael Dell is doing as an author. And so this is live information that's up on the Web site. At any time you can go do this, and typically you could hook this into internal systems. I don't know if the Internet connection is defeating me here or what. Did I not hit Return? Oh, that was my mistake, sorry, I didn't hit Return.

So, I'm a little bit ahead, his is the 401st best seller. It looks like Michael and I need to come out with a new book to move up on the charts there.

Let me just show you one other document here and this is to talk about reading off the screen. If ,as soon as you print a document out then, of course, you've had to pick not only what sales information but you've also had to pick the fact that if you make comments on it, then getting those comments back to the person is fairly inefficient, and yet reading off the screen has been very painful, and it's been a Holy Grail for us, and you've heard us talk about this over the years, making screen reading something that's very comfortable, even for fairly long documents. And so here I have a document, and you can see it's noticed that I have a wide display. So, if I scroll through this, it's showing, side-by-side, two pages. At any time I could say, well, this text isn't big enough, and I can say, OK, I want to make the text bigger here, and just click this and make it as big as I want. So, if your eyesight is not as good as it used to be, and you like to read it a certain way, you're doing that. You're not affecting the document, you're not re-editing the document, it's just doing this dynamic layout, and it goes down in size.

Over here, I turn over what are called thumbnails, you can see the different pages and what they look like. And I can just click. And I click on a page, it's a thumbnail over here, then it navigates to that part of the document.

Now, before we had this, when you were reading online, you had to sit and scroll with the scroll button on the mouse, you had to scroll line by line. And all usability testing showed that was a really terrible way to read documents. You really want to read it page by page. And so here, as you roll the scroller, it's just going page by page. So you don't get things spread across the pages in a very strange way. It's a very nice way to see a document. You can see the document outline here, if you want, instead of the thumbnails. Now, we're using the quality of this LCD, and the fact that we have clear type capability in order to make this work very well. Another thing, since it's online reading, at any time you can point to a word and just click on it, and we go out and we find you information. Here I clicked on the word diagram, and so I'll see in the dictionary the definition, I'll see the thesaurus. I can pick any language and translate that word into Japanese, or I can pick French, and see what it is in French. I can also do this so that when I pick a term it goes out and looks at a web site. So, for example, if you subscribe to Dow Jones Factiva, you can take something like a company name and ask to go and do a search on that, and make sure that one of the sources it looks at is the Factiva news search. And so I picked NEC; I could have clicked on that in the document search, and it would go out and see all the latest news and information about NEC.

We often have code names for products, so I'll see documents that have these code names, and I don't even know what the code name refers to. So I just click on the document and I say for it to search, and I pick the Microsoft Web site, all of our different Web sites, and I can see exactly what's going on with that. And so the fact that you can take any of this text, and use it to navigate and get information, is another thing that you only get with online reading.

Hopefully that gives you a little bit of a sense of, during those hours I'm not in meetings, what I'm doing. Both with your direct Tablet experience -- and Jeff Raikes a little bit tomorrow is going to talk about when you're in the meeting and you've got the Tablet note taking -- how that changes that part of the job, which, I'm sure for a lot of us, sitting in meetings is actually even more hours than it is sitting in front of the terminal.

There are a lot of breakthroughs coming, I won't dive into these, the ease of developing applications, that's been very hard for corporations, because they're duplicating a lot of things, and the code is being written at a very low level. The idea of being able to navigate business data in terms that you understand, by division, to really have schemas that let you navigate not just at the cells of the spreadsheet, but the terminology that makes sense to you. There are some big software breakthroughs coming on that will turn this into sort of a low-volume specialized market, where every worker has these rich views of profitability and sales, and things like that. Speech is coming along; it's not mature like handwriting is, but it's within a few years of that, so that making voice annotations, of course, that's easy, it doesn't require recognition, but even giving your commands and navigating with speech will be possible. And the connection of the phone and the PC, we're going to see some dramatic things there, where the phone right now is a challenge for IT departments to manage, because people are downloading information onto it. By doing integration we can allow that scenario, but still have the connectivity that people expect.

So the conclusion for me, in terms of how you should think about IT investments, and where it's probably most effective, and making a difference, is it's important to see that although those harsh realities are there, how is the industry responding to that, software advances are the key breakthrough that turn back those realities. There's a lot that can be done in empowerment. And as long as these breakthroughs are coming along, it's worth it to give those people the best tools. And one great thing is that there are many best practices; for any one of these things that I've talked about, there are pioneering companies and Web services, and how they're using e-mail, how they're using SharePoint, they are doing pretty neat things. And so it's not like people are off on their own. They can move out with some certainty, seeing what the pioneers have done.

So we're pretty excited, obviously we must be, we're still increasing our R&D budget, up from the $5 billion level, and I think that will be fully justified. So we look forward to working with you on putting some of these things into practice.

Thank you.

(Applause.)