STEVE BALLMER: Well, thanks, and let me welcome everybody to this fifth CEO summit that we’ve done. It’s our pleasure to have a chance to host you today. We do this session as an opportunity for all of us to share feedback, to brainstorm, to prompt some ideas and hopefully leave you with some thoughts that are useful.

We already have a thought that is useful to us. If you’re going to put 150 people in a room with a device, you’d better not test with 75 devices the night before. You might not put enough bandwidth into the room. (Laughter.) And at least as I was providing help desk support to this part of the audience, I was learning that lesson very vividly.

It has certainly been a wild year since the last time we did this event. Last night at the cocktail party I was talking to a number of folks about observations on the economy and how wildly that has changed. Certainly in the information technology business there has been ongoing fervor, positive and negative, about e-commerce and e-business, about IT and capital spending. IT is over 50 percent of the capital spending now with businesses in the United States. And as capital spending has certainly suffered over the last year; a lot of discussion about what’s going on on that front. And certainly also a lot of continuing discussion about how to make employee populations more effective and essentially more empowered to make right decisions by putting information in front of people, right time, right place, et cetera.

We’re going to have a chance to hear from Bill Gates about that and then have a chance to share some communal thoughts. We have a number of panels in which attendees will share some of their experiences and thoughts on a wide variety of these issues.

Certainly from my perspective the adjustments that we’ve seen in the capital markets, in the labor markets have been amazing and they’ve been amazing in all the ways that people focus in on obviously, particularly in the capital markets, but I was particularly glad that Bob raised the idea of doing a panel this time essentially on the labor or talent market, because I have seen a dramatic change in the flow of people and talent over the course of the last year since we’ve seen the issues arise in this so-called dot-com world.

It is a pleasure to host this kind of event. We want these things to be very interactive, very lively. We have a pretty full agenda, but there’s still plenty of time for interaction, and that absolutely is the critical element.

And without further ado, I’m going to turn things over to Bill Gates. Bill is going to lead us off with a little bit of discussion and demonstration of some of the kinds of things we think will continue to impact and positively impact the work all of us do over the next several years.

Bill Gates.

(Applause.)

BILL GATES: Well, good morning. I titled my comments The Digital Decade because I want to talk about things that won't just happen in the next 12 months or 24 months. In fact, I think a lot of people overestimate the changes in the short run and then when they see that they've overestimated those changes, they underestimate what's going to happen over a five to 10-year period.

In terms of the digital underpinnings of how people work and how business is done, we are in the early stages of that, but the hardware and software technologies that are going to enable those things are just becoming available. And so I will speak as an optimist about the opportunities that are out there.

Every few decades, the way we think about handling information and using computing changes fairly dramatically, from the centralized mainframe era with databases there to the world of personal computing to the last decade where there's been a lot of Internet publishing, a modest amount of transactions -- but really just using the Internet to present screens.

We haven't taken the Internet and turned it into a data bus. We haven't taken it and made it so that you can see information from different sources and that it fits direct in to both your knowledge workers and your enterprise software. And the benefits that will come as we do that are pretty incredible.

The pieces that have to come together for this do take time. I want to comment on the idea of these marketplaces, which they came up last year, and I struck a little bit of a conservative note saying that the idea of an entity, an independent entity being the exchange for these companies was probably not the primary way this would happen, and there's been a lot of allusion to that in the last year. We want to show how we see that moving forward and some very clear actions that I think everybody here can be involved with.

So I think this next decade is the big one. This is the decade where your involvement with computing will be very pervasive -- from reading to meetings, in every way.

Now, given that bullishness, I think it's interesting to talk about what happens, you know, what about all these promises that weren't fulfilled; is this just another case where somebody is being overly optimistic; hasn't reality set in? Certainly one symptom of reality setting in is what's happening with the pricing of technology stocks. Here's the NASDAQ and if you go back to last May, it was about 4300, off from the peak of 5000 -- but now even with somewhat of an up-tick here, we're less than half of that peak.

So some of the mania about this really changed. People have stepped back and even in some cases wondered, "Is there anything magical here at all? What is this all about?" In the extreme, of course we all have our favorite examples of Web sites that we used to visit that are no longer there, and many of these companies, in retrospect, you look back and say, "What was that all about? Why was that IPO such a big IPO?"

People talk a lot about the new economy -- or they did talk a lot about the new economy. The thing that I think is enduring in all of this is that it is an information economy. That is, the role of information, in particular intellectual property, is far more dominant today than it was in the economy a decade ago. And overall, that's an incredibly positive thing, because the magic of intellectual property is you can invent the idea once and whether it's a song or a drug or a piece of software, that can be made available to millions of people at extremely low cost. And so having that be a big part of the economy means that the improvements in productivity that you get are much more rapid. And that's a very permanent thing.

There's tracking that's been done where you look at the capital investment by sector in the economy and the productivity increases are closely correlated with the sectors that made those investments into information technology.

I certainly don't think that the rules of profitability or cycles in the economy have been suspended. In fact, when you get into intellectual property, some of these things get tougher because the ability to predict which investments are going to pay off is actually much harder, and the impact of being wrong on some of these investments is far more dramatic. In a manufacturing-type economy, if you make too many, you can cut back 10 percent, change the pricing appropriately and the capital spending makes sense. In a world of intellectual property, particularly when you have an ebullience, where people overestimate what percentage of those things makes sense, you get too many entrants trying to share a market. Many of the investments that are made are literally worthless. There's no tuning or cutting back that makes sense in terms of giving any value.

So it is a very risky thing for capital; huge rewards, more outsized rewards for the cases that work than ever before, but on balance the return levels offset that with plenty of losers on the other side.

So consumers will get the benefits of the productivity. The things that have happened in the American economy are not something that just goes away as people come back to reality.

Now, intellectual property, the way you think about the business equation then, I want to discuss that because I think it's something that is at the center of this and yet hasn't been thought about a lot. When somebody like Microsoft sits down and says, "Okay, let's spend $500 million on interactive TV," what sort of uncertainty do we have there and how are decisions like that made? Well, somebody can say, "Well, it's the classic equation." I've got it at the bottom of the slide here. Why don't we take our costs, look at what we can sell it for, multiply it by the volume, look at the profits, take some discount rate, look at that over years and decide based on that. And that's absolutely right. That basic paradigm still exists.

But if you look at those pieces, if you look at the cost equation and the demand equation, they are very different. In classic economics -- this is Economics 10 basically -- you've got a supply curve that tilts up as you try and drive more capacity into the system and you've got a demand curve that intersects it and you have some pricing flexibility in there. If you're wrong about the demand curve, the total size of the market goes down pretty much in a linear way and that upper triangle, which is the profit, also declines in a fairly linear way.

Now, with intellectual property the upfront costs are what it's all about. And so the supply curve looks like this -- very different. In fact, it looks a lot like the demand curve. You know, this is an extreme case. Say a piece of software costs $10 million to create and the marginal costs because it's going to be distributed electronically are basically zero, and so you get exactly that curve as a result.

Now, once you've built one of these things, the only degrees of freedom you get are where to price it. And so if you see a demand curve that looks like that, you can price anywhere between those two intersection points in order to maximize profitability.

And the beauty if you underestimate demand, if it's actually higher than you expected, then every single additional unit is pure profit there. On the other hand, and this is very typical, and this has been happening in literally tens of thousands of cases out there, if demand looks like this because some hyped-up phenomenon you expected didn't take place overnight or because there are 10 companies competing for the ad rotation market, then you see a situation where it doesn't matter what price you pick, your fixed costs mean that there's no return that ever takes place there.

And so I want to go out and look at that in terms of companies and how that's been reflected. This is a chart -- I'm just out on the Web looking at various prices and valuations. This is a chart of Microsoft. And so when people first looked at the whole intellectual property thing, they said, "This is a magical business." Where have you ever seen something like this? Well, this is a case where the number of people who invested in building software like we did was very small. It wasn't heralded and so the over-investment situation that we've seen in the last few years didn't take place, and so the returns were really incredible. And there are many cases like this. Take somebody like Adobe, you know, the same sort of curve exactly. They got in, did desktop publishing software. They were uncertain about the market, but eventually they developed it and it was far larger than they expected.

And so when you get examples like this, capitalists respond. They say, "Hey, this is a great thing; we ought to get some venture capitals out here. We ought to get more money. We ought to fund more of these things."

Well, intellectual property is the backbone of our economy. There are a lot of issues about it. There are the piracy issues. Worldwide, you have that with music, you have it with drug knockoffs, so there are a lot of issues now about should people pay for intellectual property.

I think if we took a poll in this room, intellectual property would do quite well. (Laughter.) And you need to spread the word about that. But it does have some interesting challenges. Let me look at SAP, which is a wonderful company, but if we just zoom in on this, you can see a lot of fluctuation here. What's this about? Well, intellectual property has an interesting problem, which is that it lasts forever. And so it's not like Coca Cola where you say, "Okay, you liked your last Coke; now would you like another one," or razor blades that wear out. If people buy ERP systems, there's actually a serious question what the additional revenue will be.

Now, that's the kind of challenge that companies respond to with new versions, new capabilities, but a lot of the fluctuation that happens here in SAP came from the fact that people forgot that intellectual property lasts forever. Your own install base is serious competition. You have to always do better than that, which sort of guarantees that people are pushing forward in a very aggressive way.

Now, some of these markets, unlike ERP that aren't established, are a lot tougher. Let's take speech recognition software. There are a couple companies that do that up on the server. And here you can say this variation here, which is a factor of five or so, up and then down, it's not because they ran into some technical problem. In fact, technical progress throughout this period was every bit exactly as was expected. It's because that demand curve was incredibly uncertain to people. They had no idea if speech software like this was meaningful.

And, in fact, in intellectual property there's another company -- these are even small companies, not even taking IBM or Microsoft -- but there's another company in this market as well. And you do get situations where if somebody has a significantly superior product, your demand can literally almost drop to zero, and you don't get that when you have manufacturing constraints.

The whole psychology of trying to get ahead of seeing these intellectual property assets and then getting ahead of itself, it gets most extreme when you get somebody like an incubator -- this is CMGI -- where they're not just a company playing this game; they're sort of the exponential of this. They're a company that's creating companies that are playing this game, and so you get a variance off of that that's even higher.

Some of the people assume that capital would be available for them to put these things together and so you get situations like this one -- this is Teligent -- where if they've been given all that capital, it's the dream that stayed alive. If you never get into a situation that people are starting to wonder, then you can get to a break-even profit model. But if in the middle of that your cost of capital goes up fairly dramatically or the psychology changes and you're cut off, you don't even have a partial value, and that's been seen in some of these types of sell-offs.

And so you look at biotech volatility, software volatility, it's extremely high. So it's not just the dot-com phenomenon that changed here; it's sort of a coming of age of people thinking about intellectual property and understanding the risk levels and uncertainties that are in these investments.

Inside Microsoft, even though if you look at our sales and profit overall, it's a nice curve over the last 10 years, inside there are many projects where we've spent $100 million and it was a complete failure. And fortunately there were some other projects that we spent $100 million or less and did very well. So having that kind of portfolio is interesting.

What are the things you can control as you manage projects that have this intellectual property type aspect, because every company has -- almost every company has things of this nature? The things you really control are the talent of the people involved. You can also control saying to yourself, "What is the milestone that will let me understand that demand," for example, a milestone where you can see, "Is our product really better than other people's?" and some process by which you get an objective evaluation.

One thing that's clear about intellectual property projects is that the people who work on them fall in love with them. And that's good. You don't want to discourage people from that. But, you know, you rarely go to a meeting where people say, "You know, this intellectual property project we're working on, you really ought to shut it down. You know, the idea that you'd put another 100 million in and we might even be semi-decent by then, you know, we don't really want to do that." It's just not human nature.

And so keeping the talent level high and the objectives about what you're really trying to get done, that's a very top process. You see it in the entertainment industry. You see it -- pharmaceuticals are probably the closest map to what goes on.

And so the information economy will be an economy where judgments like this are very, very key and as an investor that's a very tough thing. As a consumer, the benefits of what come out of this are better than any economy there has ever been.

So the lessons are pretty straightforward. That equation about profitability still applies. Once you've got the project done, all you get to do is pick a price level and that's it. So most of your intelligence has to be exercised early on in the process in terms of letting that project proceed in how you direct it.

Technology enablement is not something that just startups have available. That was always one of the great paradoxes of the Internet. How could a technology that reduces barriers to entry allow the creation of companies whose value was greater than the ones who created assets in the real world where you build stores and there are real barriers to entry? You know, that paradox has now been resolved. There was no reason for that to take place. (Laughter.) But for a while it was interesting.

You know, the signals the market was sending were somewhat confusing. There are people who actually spun off their dot-com operations to monetize that, which is a very rational response to what the market was doing. Virtually all of those now have been spun back in.

So all this technology enablement, it's great for the companies, and in a way it allows you to magnify your excellence. You know, if you make the best product in some category, now your cost of getting customer feedback and distribution actually can be significantly reduced. If you were the number two in a market, the fact that the number one guy can get those distribution capabilities could be tough.

And a final point here is just that things take time. You know, if you take things like credit cards that today we take for granted, look at how long it took for people to just take that as common sense. Well, buying on the Internet is the same way. I have no doubt that Internet retail sales, consumer retail sales over the next five years will grow by about a factor of four or five. But they'll grow in a very linear way as the word of mouth, as the ease of use, as just the common sense that, "Hey, this is a good way to do things; it actually works," suffuses throughout the population and different age groups.

And so that's very predictable in terms of gross sales. In terms of where the profit is, that's tougher. There's no doubt in my mind that some companies like eBay, Amazon that were created partly as part of this Internet excitement, those are companies that are going to be here for the long run. You can argue about what exactly their profits will be, but in those cases, because they'll be a reliable place, you know what to expect for them, they've pioneered those businesses; they will be very successful companies that have come out of this.

But how impactful has the computer industry been? You know, you've got these reservation databases and you print checks and things. You've got office productivity. You've got e-mail. And now we've really added a fourth just in the last years -- getting information off the Web: news, sports. I imagine a lot of people in this room do that.

If you took the computers away, some things would change. These four things you'd lose. But the way that you get your movies, watch your movies, the way you read, the way you keep track of your photos, the basic way that people sit in meetings and talk about forecasts and plan things, it wouldn't be that dramatically different.

My view is that over this next 10 years, every one of those things will change dramatically. That is, the cost, the ease of use, the natural interface, basically the magic of the chips and software will come together to radically change each and every one of those things.

And even in the next year or two, you'll see much more of this. For example, take communications. Today, we have this split world. You pick up a phone and do a voice call. That works very well; it's very comfortable. You go to your computer screen and navigate documents and Web sites, spreadsheets, schedules and see things. That's very convenient. But they're split from each other. So if you're on the phone talking to someone and they want to send you a map or show you a plan or if you want to work on something together, that doesn't work.

Well, it turns out it's trivial for the systems to notice who you're calling and if they're near a screen and you're near a screen, even though the voice call still has the same channel, the same reliability, your screen and their screen are connected together, and so you can run an application that relates to customer service or just sit there and work on documents together. So this idea of real-time communication being a screen plus voice event, where you have that data sharing, that's going to become common sense because it just makes you so much more productive and it doesn't require a new infrastructure.

The software is not there yet either on the PC screens or on these new cell phones, but that's the kind of thing where the investment levels are still there.

We're going to talk a bit tomorrow morning -- Craig Mundie and Jeff Raikes -- a little bit about this new form factor for the PC. This is the tablet where reading starts to make sense and once you do your reading, the idea of annotating things and sending it around makes sense as well.

The key breakthrough on this device -- it's wonderful that the battery lasts all day and the screen is high resolution, but the biggest breakthrough is that you hold it. If you want to read for long periods, it turns out that because you don't like looking at a fixed position for a long period of time, you've got to hold it so that you can shift your position and be comfortable and be involved in what's called "immersive reading." And, you know, just that one boundary is a pretty dramatic boundary.

In the area of entertainment, we're just on the verge with both music and movies of making those things more convenient, making it so you can watch shows whenever you want. The cost of storage certainly is coming down to allow for that. In fact, I didn't put much in here about hardware advances because the summary of hardware advances, the microprocessors, the disks, the optic fiber disks, anything you want to do, there is no hardware limitation that will hold this back -- with one exception, and that's the cost of broadband communications primarily to the home and that's an area where progress continues to be very slow. Yes, cable modems are selling better; yes, DSL is starting to get together, but even if you took a reasonably optimistic forecast over the next four years, you'd still only see perhaps 20 percent of U.S. homes connected up to broadband and in very few other countries, maybe Korea or the UK would you see anything close to that level.

So overwhelmingly, even as these devices get so incredible, you're going to have a lot of people from home using the phone line type connection. So that's the one thing holding this back.

You wish there would be some breakthrough in that, but there won't be in the next three or four years. Over time the cost of the chip, hopefully the regulatory environment will make that pervasive, but not as quickly as we'd like.

This whole situation of the computer industry reminds me of how Winston Churchill felt in 1942. I know there's a very close analogy there. He said, "Now, this is not the end; it is not even the beginning of the end, but it is perhaps the end of the beginning." And what I mean by that is that these digital approaches, as much as they've been talked about, aren't all that pervasive, and yet this is the decade that they will become that way.

And so whenever people sort of bemoan things and say, "Well, the air is out of the bubble; is this is a post-something or other era," I say that's fascinating that people think that and in the short run, sure, you can see a lot of things that relate to the overhang -- used equipment that's very expensive, expectations that weren't met and things like that, but that actually creates a rationality in terms of the investment levels and the focus and things that actually build on this in terms of where we'll get.

So it's productivity, communications and leisure are the things where our ambition as an industry is to literally revolutionize every one of those things. You know, how you think of photo albums, how you think of documents; do you store them in file folders, are they easiest to get by calling them up on a screen? Do you take notes in a meeting simply by writing on a piece of paper? How do you manage your finances? Now, all of these productivity things, there are two ways they can go. They can go as a bottom-up phenomenon where individual employees come along and see the opportunity and get those things or they can come as a top-down phenomenon. And if they have to come top down, for example, like enabling a company e-commerce, getting rid of paperwork; you can't really do that on a bottom-up basis. Even electronic mail, which is so empowering and lets you get a directory of the user names in the company and you get sort of a culture that people are going to use it and you get an IT group that's chartered to have a reliable network that you just count on it being available; unless you do that, the whole e-mail thing can't happen.

And so to really drive these things inside companies, you want to take the few things that require top-down effort and unblock those, but then let people do things that are very bottom up. And if things really catch on, great, that's worth supporting, but if not, then the knowledge workers have spoken.

One thing that was a real problem was that people were having a hard time sharing information. And they were mailing it around as enclosures, but that's really a problem because then you get all these different versions out there, and they weren't able to collaborate very well.

Well, now that the Web site creation tools have been included in mainstream software like Microsoft Office, we're seeing a lot of interesting things take place there where all the IT group has to do is set up and say, "Okay, you can put Web Sites up on these servers." We have a product called Share Point that means IT doesn't get involved in individual Web sites, they just put up the storage and say, "Okay, you can create these things."

So here we have a Web site that was created when we said, "Okay, we've got the CEO Summit coming up; let's get organized." Now, in the past, we didn't just have this ad hoc Web site, bottom up, simple thing. If we had to get permission for a specific Web site, we probably wouldn't have done it and it would have been too hard to work on, so we would have just mailed a lot of things back and forth. And so the attendee list and the schedule and all those things, you would never know who had which version and your e-mail box would be just flooded with things, sometimes that you didn't care about, sometimes you did.

Now that you have this site and you can set it up to get notified when things change, you just have like the agenda up here and you can see exactly who's making changes. You know, I can go down here and say, "Okay, let's say, okay, there's no time for refreshments." (Laughter.) We make that. So, you know, it says "deleted" in refreshments and it says Bill Gates did that. And so if somebody set up to be notified when I make changes, as soon as I close this thing, they'll get a little -- depending on how they've chosen -- instant message or e-mail, or, if not, when they go back and look at the site, they can see exactly the edits that have taken place. So you just have a set of documents up here that, for example, you have the daily reports, the spreadsheet that shows exactly who's coming, the different slides that people have sent in, and there is no top-down involvement in this.

You do leverage the fact that there's a directory of names, so you have peoples names here and you can look them up and you're sure when you set up the permission, that the IT group is controlling those things. So you have bottom-up sharing taking place in a pretty profound way. And that's kind of a new thing that is really just real, really getting to be significant.

And I've picked a few Web sites and you can see they all have a little bit of a common look, just because the tool makes that really simple. People put up graphical data and make it so you can select any one of the different reports and see information and get that in a spreadsheet. And basically we've made this capability available just about six months ago and now we have over 2,000 of these different Web sites that have come up and people are using. And actually the amount of big enclosures and e-mail has gone down somewhat because that's how people are sort of trying to do this. So bottom-up sharing is a very powerful thing.

The Tablet I talked about a little bit. I would just make this one promise to you. When we have this event a year from now, we will have the Tablet PCs. We have the partners are working hard on that and those will actually be in production. So you'll get to see what the various roles of the different devices are. You know, for some applications, some of these medical applications and insurance claim applications, the mobility of the Tablet connected to the wireless network is exactly what you want. In some other applications, the small screen size, what's called the Personal Digital Assistant, or the phone-type form factor, and those are coming together, they won't be all that different. Some will be more optimized for the screen and some will be more optimized to be a phone, but it will be one product category. They both have an important role.

Now, some people went overboard and they thought that you would be doing all this commerce, like you would be in your car and decide to buy a TV set and you'd pick up your cell phone and you just couldn't wait, so you had to have that small screen and you just buy, buy, buy. (Laughter.) And now when you got home to the big screen where you could see more and everything, you'd already spent all your money. (Laughter.)

So the idea that there was going to be this whole branch called m-commerce, there was an element of truth to that where getting the map to your current location, seeing what's nearby or getting notification -- you know, if you set a trip up on your big screen and then let's say a flight is canceled or somebody's plans changed, and you're traveling and you just have your small screen, yes, you'd like to be notified. And so these devices have to be complementary. You won't have people who just do their homework on the small screen or just do their business planning on the small screen, but you will have that in a way that has to fit in. And, in fact, if we don't make it seamless where all the information moves between the devices without your thinking about it, the multi-device world will be completely impractical for people. And so that idea of using the Internet to move the information around, that's a key software advance.

Another key software advance is filtering out all these things that come to these devices so you only get the stuff that you care about. It's a non-trivial problem, because you don't want to have to force the person to be very explicit, but the last thing you want is to be interrupted with something unimportant. Even if you're sitting at your full-size screen and creating a document, your threshold to being interrupted is high, whereas if you're just browsing around the Web, your threshold for being notified of something is a bit lower. And having the software help with that and value your time is one of these great challenges that we're very focused on and certainly in this decade people will take that for granted that that's a solved problem.

Now, part of the progress of this industry has come through standards. The standard of the PC, that's now extending out to the PC server. The PC server has reached a milestone where both by having many of the servers to get lots of combined power and the fail-over capability, which is called "scale out," and by having individual servers that are extremely powerful, which is called "scale up," that PCs can take on the hardest tasks. And so you get the tools, you get the great price performance that comes from that environment.

In fact, here is this comment: Compaq and Unisys are two of the key partners who have really helped drive that forward. And so the costs of doing even a very high-scale Web site aren't what they would have been, and it makes the barrier to entry quite low.

Windows, of course, is a common interface, common way of exchanging data. I really want to focus on these last two, and I don't want to dive into the technology, but the Internet browser standard, the thing that allows you to go to these sites, all these different sites and see things, underneath that, what you never should have to look at directly, there's a thing called HTML. And for just viewing things, that's great. Now, just viewing things; and so there's a new standard that really got started three or four years ago when some people from Microsoft and some other companies really realized the limitations of HTML, and it's called XML. And what it is, it's about data, data describing itself. You're going to hear that term a lot, because the idea of Web Services -- different computers on the Internet talking to each other -- they need a standard to understand each other.

When you used to just write applications inside your company, you didn't need XML because the people who were writing the different applications knew each other, and they could say, "Okay, we're going to format the data this way," and it didn't have to be self-describing.

In the Internet, you don't get to coordinate with all the IT developers in all these different companies, so you need the information to be self-describing. And that's what XML is about. And one of the good things that's coming out of people thinking about these marketplaces is the idea of individual industries talking about what their XML data looks like -- healthcare records, banking records; in the case of retail, taking EDI and building on that, moving forward. It's a very important thing. And it is a standard that is completely agnostic to hardware, software or anything. It simply lets you pick the best hardware and software to do the applications.

And there is a checklist of things that we've come up to as you write an application that determines are you doing it in a way that supports XML flexibility. This is a new thing for developers. That knowledge is just starting to spread. But in terms of a priority for how applications are built, that's a very big thing.

Now, one of the scenarios this is key for is this idea of e-commerce. You know, I mentioned that a year ago that the way that e-commerce was going to happen was that there would be these independent marketplace companies and you'd run transactions through them and they would charge some type of fee. In a few cases that makes sense, but actually what makes a lot more sense and we'll be a huge part of it is individual companies saying, "Look, all my suppliers, I am going to control the information, but I want them to be able to talk to me."

And this is not just a problem of solving the computers talking to each other. That's a hard problem, but as you can see on this picture, if you really open the problem up, you've got lots of people inside, let's say company A and B are doing business with each other. You've got lots of communications taking place. You've got it not just between the computers of these companies, what's often called the back office, but also between the knowledge workers, sometimes called the front office. And a trivial transaction, where you just want to buy, say, a commodity-like item and it's going to be delivered on time, you don't need much front office activity, but anything beyond that, where there are surprises and changes and complexity, then humans, the knowledge workers get very involved.

Well, unfortunately, the way a lot of this software has been written, the human communication channel and the computer communication channel are isolated from each other. You can't see what's gone on between one and the other. You can't make changes. And so it's a big problem.

We decided that you've got to get into the office software itself the idea of seeing what's going on and tracking that. It's not only just supporting XML, but actually supporting the idea of these business processes.

And so the question we asked ourselves is: How can you very rapidly have a company connect up all its suppliers? You're not going to get your small and medium size suppliers to connect up if you just say to them, "It's your lucky day; you get to spend more money to set this up. Everybody else you sell to will have a different way they want you to do it. And the reason we're doing this is so I can get a better price from you. And isn't that a nice thing?" Or in the extreme case where you've got a marketplace where all of your offerings are going to be visible to everyone and the idea that you do special work, special design together is sort of taken out of that equation.

And so the Journal this week had a special reprise about B2B and it sort of started out saying, "Well, wasn't that one of these false promises," and actually as they went around and talked to people they realized that in a different form this thing really is happening. You know, B2B should not be a discredited thing because the transparency, the efficiency you get out of that is one of the great breakthroughs that given a few years here with the right software, something incredible will happen.

I decided to have a quick demo done to show you how this can be set up and what it looks like for a knowledge worker, and so I'd like to have Chris Lovett, who's an architect in our B2B group, to come up and to give you a glimpse of how it's a little bit different vision than just a pure marketplace approach.

Welcome, Chris.

CHRIS LOVETT: Thanks, Bill. What a privilege it is to be here today.

So Bill has been talking about XML. That is one of my favorite subjects, so I came to Microsoft to work on XML about four years ago, and I'm real excited that it's become as big as it has. And Microsoft has been busily building XML support into just about every product we have, including Microsoft Office.

So what I'd like to show you today is a quick demo that illustrates how this kind of deep XML integration into these products can make business-to-business automation really simple.

And what I've done is I've chosen to pick on the furniture industry -- not for any particular reason, but we've got all this furniture up on stage. XML integration applies to any industry.

How did this furniture get here? And typically the process starts with a great big catalogue like this one with lots and lots of product information aggregated from all the different manufacturers. And then telephone calls are made and faxes begin to flow back and forth. Sometimes you have to scribble all over the fax because you know the prices have changed, the catalogues are out of date. Some companies will even send you a CD-ROM and you get a catalogue online that way. And then paperwork flows back the other way with invoices and so forth.

With this kind of manual process we've had companies quote error rates as high as 30 percent, some even higher than that.

So despite the fact that the Web has been around for almost ten years and EDI has been around for 20 or more, most of us are still doing 10 times more telephone and fax than anything else. And we all know that the Internet is a vehicle to make this whole thing a lot more efficient.

So the question is why haven't we succeeded in doing that? And I think the reason is we just haven't made it simple enough yet, and so that's what I'm going to show you here is a way of doing it very simply.

Let's put ourselves in the shoes of a small company. Here I have a bunch of little retailers. Today, they have to connect to lots of different places, and each place has a different interface and a different way of user experience, a different way of doing things. How are they going to train their relatively low-paid staff with a high turnover on how to use all these different things?

Let's think about the larger companies. How are they going to bring thousands of these new business partners online? Bill actually talked about that problem. And we know that EDI, which was designed for more of a point-to-point solution, doesn't scale up like that to thousands of suppliers or distributors.

So the other thing is that B2B relationships, you're starting to share pretty sensitive information with your business partners. This is not like B2C where you just blast it out to the masses. This is a scenario where you need tight control over provisioning of the Web Services out to your trusted business partners, and how do you do that in a way that's scalable.

So I'd like to start with focusing on the small guy here, the retailer, and there's an employee there. The retailer I've chosen is Redmond Home Furnishings. It doesn't actually exist, but it would be a good company if it did. (Laughter.) An employee there named Jim wants to buy some new top-of-the-line ergonomic chairs, and he has a lot of Microsoft yuppies coming in and looking for them for their home office, so he wants to get stocked up.

And he already has a business relationship with this big distributor here and so he calls them up and he has a conversation with a gal there called Kathy. Now, this distributor is beginning to feel a little bit threatened by the Internet and they want to create a really strong customer satisfaction and loyalty, and so Kathy offers Jim a 5 percent discount if he'll try their new state-of-the-art automated B2B system for doing online business, where Jim could actually query their inventory online so he wouldn't have to make this phone call. And Jim says, "That sounds great," but he's really skeptical because what software does he have to install, who's going to support that software, how is he going to train his employees on this new way of doing things. And Kathy says, "It's really simple. Do you have Microsoft Office and an Internet connection? And Jim says, "Yeah." "That's all you need. Give me your e-mail address and I'll send you some Office templates that are already configured to do business with us online." Jim says, "That sounds great." He's already familiar with the Office products, so he'll give it a go.

So before Kathy is ready to send that e-mail to him, she needs to configure her system. Now, she has some sort of B2B gateway product, probably BizTalk Server in this scenario, that will control the flow of business transactions in through the firewall into their ERP system.

So Kathy pulls up the record for the Redmond Home Furnishing store, fills in the missing e-mail information. And then she gives him access to order management inventory and content, so now Jim can use these new Web Services to query the product catalogue online, to query the inventory online, see if they have it in stock, and then to send purchase orders, receive invoices and so forth. So she's ready to go ahead and configure that. She can go in here and customize this if she wants to, but just the defaults will work fine here.

And Jim will receive this e-mail from Kathy. You know, "Thanks, Jim, for giving this a go. I think it's going to be wonderful. This will make things a lot more efficient, et cetera, et cetera. And all you have to do is click on this button to install your Office templates that are already configured to work with our online systems."

So Jim goes ahead and does that, installs it and that takes him to this page. So now Jim can query the online catalogue or he can go straight to purchase orders. He already knows what he wants exactly and so he just goes straight into purchase order in this case. And that brings up this Office Excel spreadsheet.

Now, this is not an ordinary Office Excel spreadsheet. This one is wired up to some online Web Services, so he can type in Kathy's name here and click "find" and it brings back the exact correct information. And he can type in "10 chairs" here and look that up. And it comes back right away, so this did an inventory check first of all and said, "We only have nine of those." And what it also does is fills out the rest of the line item details here with the exactly correct SKU number. So now we have real-time information flowing between the distributor and the retailers and we're starting to eliminate some of those errors that can occur in the manual process.

Also we've put a link in here so that Jim can read up on the chair and be familiar with all the terminology. So let's go ahead and do that and read about the actual kind of material they use to make the chair and can be knowledgeable when his customers come in.

So Jim is now ready to send this purchase order. He can even put his own stamp of approval on it, you know, say, "Redmond Home Furnishings." This is just a local Excel spreadsheet. He can do whatever he wants.

And this one interface can be used to do business with anybody, and that way he can train his employees. He's got one stable thing that doesn't keep changing every week. And he's ready to go.

So when he clicks the "send" button, we're going to send some XML over the wire. And I'm actually going to show you what that looks like, since there's been so much hype about XML. This weird gobbledygook on the screen here is XML. And the reason why I'm showing it to you is not because a human would actually look at this stuff or deal with it, but for programmers like myself who need to wire this into ERP systems and Internet gateways and so forth, this is actually beautiful. (Laughter.) This is clean, pure structured data and I can consume this programmatically, and just about every platform, OS system on the planet now has support for XML, including the little iPAQs that you're holding right there.

And that now goes to the distributor and the distributor processes that purchase order, goes into their ERP system. They ship the goods. And then what they send back is another XML message for the invoice. And Jim will receive that, and he can check it and it's again another slightly different Excel template that takes the XML message off the wire and renders that in a nice invoice format. Don't look too closely; I didn't do too much work on that official invoice. And the other thing is we've put an approval button on there to just streamline getting right into the payment process once he's ready to approve that.

Now once Jim does this with a bunch of different suppliers, he's going to want to look at an aggregated view of all of his current business documents and what the status is, so we can imagine a page like this where he can track that. And he can do all of this on his local PC, so now he's starting to make better use of his local machine. Or he could use a hosted solution like bCentral where he can do order management out there in the cloud.

So one last thing is when companies have done this for a while, they're going to start putting together a history. So unlike the manual process where you kind of lose track of what happens over time, he can now start to look at trends, how many purchase orders he's submitting over a six-month period. And even the small guys can now do some of these kinds of business intelligent things using the right tools.

Now, the important point I want to make here is that most of what you're seeing today is already possible using existing software. I actually haven't got any beta bits or anything in this demo. And I know for a fact that there are several global 2000 companies that we've talked to that are actually planning to deploy this right now on a pretty large scale and already building the prototypes and so forth and getting ready to do that.

So what I think is .NET Web Services on the large company side, XML support in every product on the client side will do is make this kind of B2B automation really, really simple and actually scalable and doable.

So back to Bill.

(Applause.)

BILL GATES: The next time you see your CIO, tell him you came to Microsoft and you saw raw XML. (Laughter.) And that you liked it. (Laughter.) And I'll bet many CIOs haven't seen raw XML so you're out in front there.

Well, let me just close by talking about what I think the opportunities for leadership are -- you know, what are the kinds of things that companies can do to get more value out of the investments they're making.

You know, today, companies fundamentally are spending the money to have reasonable PCs connected to networks and Internet Web sites. The actual cost structure of what everybody is doing in this room on a per employee basis in terms of knowledge worker empowerment, it wouldn't be that different, but I would claim there's a very dramatic difference, perhaps as high as a factor of four or five, between the value you get out of it.

Now, knowledge worker improvements are things that are a little bit hard to measure. You know, do people get customer feedback more rapidly? You know, are they able to survey customers? Do they communicate with each other and partners in a better way?

But these things are fundamental to competition, and I would say that letting them build the individual Web sites, just like the Digital Dashboard things that we showed here a few years ago that have really caught on and become big things, I'd say that's the next big thing, making it easy for them to set up those shared Web sites. The IT department does have to provision the space, but just once and then everybody uses that, sets it up, decides who has access without any IT overhead whatsoever.

Another thing that we've found very magical is the wireless infrastructure. As was mentioned, the iPAQ is connected up to an 802.11 network. That's just the acronym. Actually putting that into all of your buildings so that people with portable PCs as they walk around and go to meetings have access to that data, it's rather inexpensive. And it's not like wide-area wireless where there's per minute charges involved; you just get the base stations and the spectrum is generally used spectrum and so it's just the capital cost of putting in the equipment and then everybody's connected up and they get those capabilities.

It's really changed meetings at Microsoft in terms of being able to get information right there in the meeting, discussing things, zooming in on the data and not a huge investment. So 802.11, which you will see not just in your offices, that you'll start to see in every hotel and homes and convention centers and things like that.

Finally, there's the whole idea of digital meetings. You'll see about that tomorrow. Jeff Raikes and Craig Mundie are going to actually show you what it means to have a digital meeting.

And then finally, and I'd say this is probably something that's very new, because the whole way this can bootstrap, come together easily and be simple for everybody involved, how digital commerce can get going, it's really only been in the last four or five months that we've seen how those pieces need to be assembled. And it's a different view than the marketplace has had certainly a year ago, but one that I think is pretty profound in terms of what it can provide.

So you're probably not surprised that I'm optimistic about what these things can do, but these are long-term investments and it's going to be a very amazing decade.

Thank you.