Good morning. I want to talk about what I call the "digital nervous system." When I say nervous system of a business, what I mean is the way that a business deals with events, planned events like yearly budgeting or sales results or unplanned events -- competitive activities, unhappy customers. And there's all the different systems -- the meetings, the paperwork, the way the information workers are organized, the way that information about customers is stored, the budgeting system, the coordination system. All of those are the nervous system of the company. And no matter what business you're in, it's my claim that the excellence of that nervous system determines your competitiveness.

Now, in the past, companies were virtually all alike in terms of how well they managed these things. The tools were well known, and so there wasn't much differentiation. But my thesis is that with the incredible advances in technology, it's now possible to have dramatically more responsive nervous systems. And I term this the digital nervous system.

Technological change has been pretty unbelievable. When Microsoft got started, computing was a million times more expensive than it is today. And computing was simply a central tool, a tool for keeping track of business accounts or reservations. And only the largest organizations could have this tool. In fact, it was a clear advantage for very large companies.

That factor of a million did more than simply drop the price. It changed the very character of how we think about computing. It changed it to be a tool of communication, and it enabled people to individually think about how they manage information so that everybody has a machine and great flexibility in dealing with that information.

Over the next 20 years, the cost of computing will fall again by a factor of a million. Now, the good news here is that the demand elasticity for computing is quite amazing. Other markets where you find efficiency anywhere near this, you have a tendency to shrink the market. When radial tires came along, they lasted longer, and so they were quite a bit cheaper [long term]. But people didn't respond by driving around a lot more, and so in fact the total revenue of the market shrank.

When people looked at computing and the exponential impact of Moore's law, the doubling in performance every two years, they were really worried about what would happen to the computer industry. Would it also shrink? In fact, the computer industry has grown because the ways that we can use this power have proliferated even faster than the performance improvements. And my belief is again that this next factor of a million will be absorbed in the same way. So it's very different than any other industry on the planet.

As we use these devices for communications, we need to think about more than just the speed of computation or even the size of the memory or the size of the disk. We need to think about how these machines are connected together. And if there's anything that will slow down the pace of revolution, it’s if we don't get these high-speed, low-cost connections in place.

I'm very optimistic about the demand elasticity for high-speed connections. I believe that the phone company investments in ISDN and ADSL will pay off in a major way, even more than they expect to. I think that PC cable modems coming in and providing users additional choice there is a very positive thing that will spur more competitiveness in this area. And so this is a fundamental element, and unfortunately it's one that's subject to a great deal of government regulation.

Inherently, networks have a problem of needing to be shared to be economic. And so you will see a lot of differentiation from country to country in terms of how this is managed. Does a country create an atmosphere where these investments are encouraged? And because the quality of these nervous systems depend on that, and because the very competitiveness of the firms in the countries depend on that, it will make a huge difference in terms of who leads in world competitiveness.

Now, when we take the improvement in speed and the improvements in communications and put that together with what the software industry is doing, I'd really say that there's an opportunity to do almost anything businesses are interested in doing with information. If you can imagine how information would flow in your company, how it would automatically be analyzed and compared and summarized and brought to the attention of the people who need to know about it, I'd say that building a system that fulfills your wildest dreams there is possible today.

And it's possible with development costs that are really quite modest. We no longer talk in terms of many-year projects -- two, three, four years of development or a huge backlog of activity. Today, if the right infrastructure is in place, people can build applications in four months or six months or eight months.And the beauty of these applications is they get piloted in a few days, and so the user is very involved in saying how they want it to work and how they want it to tie in.

Now, this does assume a basic infrastructure for the knowledge worker. It assumes that every knowledge worker in the company has a personal computer that's reasonably up-to-date -- say, about a three-year hardware cycle. It assumes they have common productivity tools for spreadsheets and documents and presentations and that those are shared not only within the company, but also with the key partners, with the law firm, the accounting firm, the key vendors that you work with.

It assumes there's electronic mail where your email, faxes, and voice mail all come into one place. And it's an extremely reliable email system, so people can count on it working not only for simple text messages but also for sending rich compound documents. And, of course, it assumes that these machines are connected to the Internet, so that not only the knowledge and information inside the company are available, but it also connects them out to customers -- and the entire world of information being built on the Internet is there.

To put this basic infrastructure together is a very serious investment. And to set an example that it really should be used and become a primary means of communication requires leadership from the top of the company. But all these other applications, all these other improvements, cannot flow until you're at that point.

What these standards represent is the fact that the PC [business] model really has won out. The computer industry has had many debates over the years about things. But the idea of separating hardware from software, that's proven to be a model that allows hardware innovation to go at full speed and software innovation to also go up full speed, independently. Knowing that there's a standard that allows for the kind of volume that really allows great software to be developed and yet priced at very low prices. And five years ago you could have said there was a debate. Should hardware and software be separate? Today it's not possible to say that.

There are really only two questions left for the PC model: What can we do about manageability? And what can we do about scalability? People are seeing now that if you take the cost of this infrastructure and you really look at everything -- server costs, communication costs, application development costs -- it is very, very expensive. And it's excellent that there's this incredible focus on that, because, in fact, there's an opportunity to reduce those costs rather substantially.

Manageability has become the top priority not only for Microsoft, but many, many of the key companies in the industry. We want to get this lower-cost manageability without giving up the flexibility of the machine, of the ability to go and get the latest and greatest software, the ability to have a portable machine you can take with you to the customer location. Over a third of the PCs sold today are portable machines, and that's increasing rather rapidly.

We don't want a solution that gives us manageability while recentralizing things in a way that unempowers the worker or simply shifts costs back to the center, where cycles are more expensive and people can't count on responsiveness. It's a tricky problem, but the magic of software is going to allow us to do very, very well here.

One great example is what Westinghouse did when they went out and looked at the variety of electronic mail systems they had and different Unix systems they had, and they decided to be very rigid in imposing one productivity suite, one email standard, and get all the hardware right up to date, so that everything was common. They were able to reduce their number of support centers from 35 down to one and really pool the expertise there. The investment paid back in less than a year. And so having clear standards, particularly in electronic mail, is very, very important. I'd say at this point it is quite basic.

Even if one should get to that point, though, there's the question of what creativity you bring to using these systems. I think that is where the attention is going to shift now, particularly as we get the manageability thing very much under control.

Scalability is the other issue that people still bring up about the PC structure. Can the PC take on the most demanding tasks, the highest transaction rate? Can you cluster the machines together so you never would run out of power? That, too, is something that in the next year I think will be answered with a resounding yes. In fact, the transaction rates and the storage capability, as well as the robustness of PC architecture systems, will go well beyond what mainframe or Unix systems have been able to offer.

If we think of what the computer system will look like 10 years from now, it will be radically different, because the way we interface with the machine will have changed. It won't just be the keyboard and the mouse. We'll be talking to our computer. We'll be typing in sentences. If you want to find things on the Internet, you won't use keywords, where you get back thousands of things that you have to wade through. Instead you'll say the same thing to your computer, either by typing or speaking, that you would to a human assistant. And all the richness of linguistics and common sense will be applied in helping perform that operation.

And so we'll be using that next factor of a million in extra performance to change the way that we interact with the system. The operating system of the future, most of the code, over 90 percent of it, will relate to speaking, listening, recognizing handwriting and understanding linguistics. And the things we think about in the operating system today will still be there -- graphics, menu management -- but they'll actually be the smaller part of the code. Now, the engineering challenge to put that together is one that is very exciting because it's a larger-scale software project than has ever been done before.

Well, let's talk about excellence in these nervous systems. How are you to know whether you're managing this in the best possible way? There are lots of questions you can ask that get at this issue. First is the use of paper. How much paper is flowing around the company? I did an exercise recently where I had every paper form at Microsoft brought to me. And I looked at them and said, you know, "Why do we have this paper form? Everybody's got a PC. We're really connected up. We're ready to use electronic mail." And very quickly we were able to eliminate the majority of them. Within a year, we will have gotten rid of virtually all of them.

For example, for our 401(k) plan, we had eight different forms -- to enter and exit [the plan], to change things. If an employee wanted to know what their status was, what their option was, they had to go and see someone. So very quickly, on top of the infrastructure, in a few months we've created something that has better accuracy, better accessibility, and those paper forms are simply gone.

Another bad symptom of the nervous system is where you're printing a lot of information out, where you have these printed sales reports. When you have data on a piece of paper, you can't delve into it. You look at it, and let's say it's a sales number that really bothers you. The sales aren't high enough. What do you want to do? Do you want to call somebody up and talk to them about it and have them find the same piece of paper and look at it and have them say, "Whoops, we'll dig into that"? Or would you like to just point to it and click on it and see it broken down by geography, by time period, by product type, and really try to get at the bottom of what's going on? Is it a volume change? Is it a price change? And as you go in and see something that catches your attention, then you want to be able to send electronic mail to the right people in the organization and enclose exactly the view of the data you have. So everybody's sharing the same thing.

A lot of companies talk about executive information systems. Although the idea behind those is probably good, the notion of having an information system that only executives look at couldn't be more inappropriate. The last thing you want is for you to be looking at a system and call the people down the organization and have them looking into a different system and seeing different numbers.

You want to invest in one system that goes down to the ultimate level of detail, and everybody can traverse the hierarchy of summarized information in a very, very flexible fashion. In fact, CEOs are the last people who need executive information systems, because there are a lot of people out there helping to manipulate and find things in the data for them. It's really the organization as a whole that needs it, and needs it in a very common way.

Five years ago, the idea of having your sales results in a pivot table, where you could just sit and see it in these different ways, it was unheard of. Today it's a mistake for every company not to have the information in that form as the only form. When we sit down and do reviews of product groups or country results, we bring the pivot table up onto the screen, and the person doing the presentation has highlighted the numbers that are of particular note -- where they've done well, where they've done poorly. And we key off of that in order to have the discussion, including the ability to interactively sit there and see the information in different ways.

Another good example of doing this well is what [the insurance firm] AIG did in terms of interfacing with its brokers and agents. They used to use paper systems to send out reports and send out pricing and commission structures, and it would take a lot of time. And over 25 percent of the time was spent correcting errors and just dealing with the overhead of getting the information back and forth.

When that was moved to be purely an electronic system up on the Internet, not only did that 25 percent go away, and so the cost of running the system was much lower, but also the ability to target information and get it to the broker is improved. The information is completely up-to-date in the middle of the day, and they can analyze which brokers were doing things particularly well and make recommendations to them. It was a whole new capability that the paper system never would have allowed to operate. We will be moving away from these paper systems.

One system that may be a good benchmark is the budgeting process in a company. Microsoft used to have a budgeting process that took three months. It was our yearly budgeting process. And it was really painful, because during the three months, new things would happen. People would have to go back and make changes. And one country would have a really optimistic forecast, and then late in the cycle they'd see that their forecasts were so much more aggressive than some other countries. So they'd start cutting theirs down, and the other guys would start increasing theirs. You could never tell what was going on. It was always kind of a mess. And even after the budget cycle was over, people were still trying to get in there and change a few things, argue for a few more headcount. It was not a fun process. It took us away from thinking about products and customers, and it was just a mess to keep everything accurate.

When we put that into an electronic form and really had the templates driven in a digital way, we were able to reduce the budgeting cycle down to three weeks. It's a very intense three-week period, but we're much more up-to-date about what's going on. And during that process, everybody can see immediately what the other people are doing. We never get the situation of being surprised.

When you fill out your budget, you get the comparison to the other geographies. The minute you type in the number, you get to see if, relative to your market size, you're being particularly aggressive or particularly wimpy in terms of what you're calling for there. Everybody can see this in the form they want to. The product groups see it by product. The geographies see it by geography. The finance people see it in a P&L basis. But it all drives off of a single database that was pretty simple to put together.

Perhaps the most important system in this nervous system concept is customer information. Where do you store customer knowledge? Can you go to one place and see everything about a customer? I think that's an imperative. If you don't have it all in one place, how can somebody make a decision very quickly for that customer? How can they know whether a great job is being done there?

Let's say your company finds out that you're about to lose a customer. How long does it take to put the people together who can relate to that? They may be all over the globe, so you have to have them brought together electronically. They have to see a common basis of what's going on. This customer database is fundamental. It's the summaries of the customer database that really give you the sales analysis and make decisions. It's dealing with the individual records that decide how responsive you are.

Companies have allowed systems to go up that are really stovepipes, where information about the customers is captured based on the product or the division. Even the taxonomy of who the customer is and tracking the key individuals there, and tracking the contacts or events for that customer, that's very diffuse within a company. Driving to have a single system and all customer information brought together is very important, particularly as you bring in electronic interaction.

My claim is that there's almost no unique investment required to connect the customers through the Internet, because even if there wasn't an Internet, you want to bring this information together in an easy-to-understand fashion. Then what you want to present through the Internet to the customer, to let them bank online, to let them order online, is simply a subset of what you have to have internally for your employees in dealing with that customer. There are not two investments here, one for the internal, Intranet customer information and another to finally get out there and have a Web site where you can connect up and do transactions. That's all one effort that brings the customer database together.

Another test of a nervous system is how it deals with surprises. The human nervous system has evolved to be very sophisticated, even processing data locally when you need super-quick response times, and really integrating everything that's necessary. A good test for Microsoft was when the Internet phenomenon came along. It was a little bit of a surprise to us, how use was going up and who was using it. And the symptoms were seen by different parts of the company. Some of our sales people would see something going on. Some of our engineers would go to a conference and see something going on. It was only by having use of rich electronic mail and discussion groups that this came to a fever pitch very rapidly, and it was recognized as a crisis that required quick change.

Then, by using the digital nervous system, we were able to, without long meetings or anything, quickly communicate to people that the project plans are going to change, that new versions of software have been scheduled. We decided we were going to put more Internet standards into key products, and the schedule was very different, we were going to price things very differently. All of this got compressed into about a 45-day period. That couldn't have happened if we'd had a classic nervous system.

Another great example, I think, is the way that Cisco is able to acquire new companies. It's very complex to bring those employees in and get everybody working together. And unless you have good systems, that becomes very, very difficult. Companies that do this well are able to use acquisition as part of their strategy. Companies that don't have these systems often do acquisitions that just never come together. The people never seem to be getting the synergies that were imagined at the high levels when it was put together. In the world of technology, there are many examples of mergers that were not handled properly. I think it's more than just a business strategy question. I think it comes down to the mundane specifics of these information systems.

I want to give a few examples of how in the digital world the way this all works can be a lot better. One thing we've got here is a site called CarPoint that shows how you can actually tap a lot of rich information. You can let somebody get in a car and look around. You can take information about the person and make recommendations to them. The interactive nature of the information is very, very important. It's not just reading things that are out there. It's letting people contribute in and bring in all of their work together.

Another example is what Dell has done with their Web site. They're selling over a million dollars of equipment a day. That number is not really the thing that gets me excited.It's very easy to take orders that used to be sent in by paper and move them to the Internet and get some big number and call up the press and say, "Okay, we've got a billion dollars of electronic commerce here." If you're just shifting the paper form to electronic, that's not interesting. What's interesting is two things. First is a much richer interaction where the person can configure their systems in a better way and get better advice. Dell's done this where they create unique pages for their corporate customers, and so the configurations that you buy are up there. They've also taken the approval process that you have in your company and built that into their logic. So if an arbitrary employee comes along and orders, it understands who to send electronic mail to and who has to approve that before the order gets processed and sent directly to that person. You bring together the individual's desires, the corporate standards and the approval process without a single piece of paper, all just happening in a matter of minutes. A transaction can be placed, approved and on its way in a very efficient system. So as you ship the orders, you can make them a lot better.

A second thing you can do is match buyers and sellers that never would have found each other before. And that's very exciting, the idea that somebody who's searching on the Internet can see goods and services that are out there that might apply to them. It becomes a new distribution system, not just a way of sending in the order information.

Another pioneer in using technology is Columbia/HCA. Here in health care, they're using the fact that the Internet lets you send video and screen information very inexpensively. With NetShow, they have a doctor who does an operation, and talks about the right way to do it. And that's available to all of their medical centers. They can watch it when it's being presented and even come in and ask questions. Or it's stored away so if they're not available until later, they can call it up and play it across the Internet.

They also use what's called Net Meeting, where you can take and share a document, share a spreadsheet or a presentation. No matter where the two people are who are working on that document, because they're connected to the Internet, they can see each other's changes. This even works across corporate boundaries, so if you have your law firm and you're working on a contract, you don't need a face-to-face meeting. You can sit there and edit together and make sure it comes together. And that's just free software that works easily on the Internet today.

Another great thing one can do is redefine what your business is by having electronic systems. The Irish post office, called An Post, put up an electronic kiosk in every one of their points of presence, and they were able to expand to not only process stamp orders but medical benefits and even become the biggest retailer in all of Ireland. And now in the United Kingdom they're taking the concept of this system and even going further with it in every post office in the UK.

A final example is what Ernst & Young has done, take all their information about tax advice and put that into an SQL database. They scan things. They type things in. And they make it easily accessible so that around the world, all of that tax information can be called up by the people who care about it. There are people who are getting access who wouldn't have had any way to come in and see that information.

Let me close by making a few predictions here. These are not just predictions aimed at getting people anxious about are they doing the right thing, but real beliefs I have about where this is going. The first point is that I think there's a real tendency to overestimate how much things will change in the next two years; but also, and dangerously, a tendency to underestimate how much things will change in 10 years.

When somebody sits down to write an article, they want to say it's going to be different tomorrow. When your consultant says, "Well, it's good you hired me, because there's imminent disaster here," they want to write up that it's going to change very quickly. And then when it doesn't change, when you have that prediction that it would happen overnight and it doesn't change, you can be lulled into a sense of security, thinking, "Well, nobody's ever going to use the Internet. Geez, what was all that all about?"

In fact, year by year, the factors are driving it into the mainstream. The user interface will be dramatically better. The bandwidth will be much better. More and more people will be used to it. Kids who go to college in the United States today live what I call a Web lifestyle. That is, for any major thing they want to do -- plan a trip, make a purchase, coordinate things with their friends -- they're going to use the Internet as part of the process.

Even if they don't actually place the car order through the Internet, they'll know before they go into that car dealership exactly what the dealer paid for that car, so they'll certainly be better at negotiating what has to go on there. And so as more and more people like that move into the workforce, as we get the ease of use and manageability to be far better, slowly but surely electronic mail and all these information systems will become the mainstream.

I think it's safe to say that within 10 years the majority of all adults will be using electronic mail and living a form of that Web lifestyle. And they'll begin to ask, "Why can't I file my taxes that way?" Well, of course, you will be able to. "Why can't I ask my doctor a question or schedule an appointment easily electronically?" Some professionals will jump at that opportunity and make that possible.

Another prediction is that the whole boundary of what you do inside your company and what you do outside your company needs to be rethought once you have these digital nervous systems. I think the average size of a company will be a fair bit less than it is today. There'll be more things you decide can be outside the company than things you decide that are outside today that need to move in. And, of course, the boundaries won't be as hard as they have been. Because you'll be able to reach out to somebody who is far away and not only find them when they're far away, but also do work with them when they're far away, your flexibility will be much greater.

I was just in India a month ago, and I was stunned at the number of entrepreneurial companies there that see the Internet as a way for them to take the educated workforce that is there and get world-level wages for those PhDs and software developers. The idea of going to India and doing software development, it's been in the past a very difficult thing. Would you find the right firm? Would the information flow back and forth very well? Now, with the Internet, you'll be able to go out there, see the firms that have skills, look at their references, see what people say about the work they did, see what their availability and their hourly rate is, their expertise, and conclude an arrangement to get great software work done, purely electronically, and draw on that resource.

That kind of creativity, thinking through where the skills are outside your company anywhere around the world, can make a huge difference in terms of getting things done. It's sort of an extension of friction-free capitalism, taking the very mechanism of capitalism, which is the matching of buyers and sellers, and bringing that out onto the Internet.

There's some software technology still to be put into this. Today, if you sit down at the Internet and you say, "I want to find other people who are interested in this topic or somebody who has a product of a certain type," it's very messy to find your way around. But as we bring linguistics into the interface, as we bring richer data structures, all of which are not that far off, the idea of using it as a matching tool will come to a new level.

Another prediction is that many companies will think of their customer assets as their primary asset. And the world of the Internet has a lot of things where the barriers tend to come down. Distribution systems for lots of products are not as complex, not as much of a barrier. The thing people have is a customer relationship. It gives them their scale. It gives them their brand identity. And the way they process that information to make custom offers to those customers will be very important.

At Microsoft, we're measuring how many customers for whom we have their electronic mail address, and how many have given us permission to mail to them on a monthly basis information we think they might be interested in. Electronic mail, where you just take email names and send people mail that they're not interested in receiving, that will end up being blocked out. It's not worth a person's time to just receive arbitrary mail. And so their mail client will be set up to indicate who they're interested in -- their colleagues, their friends. Mail that's not in a predesignated category will end up in a folder you never have time to read.

For most people, the only way they'll read it is if it comes with a monetary reward; that is, if it's marked as saying, "If you read this piece of mail, we will pay you 10 cents or a dollar." It's advertising without the intermediary. Instead of paying the money to the content creator, who then creates something that's worth your time to watch so you're willing to watch the ad, it goes to the consumer direct.

That mechanism will be built into the software. If a car dealer sold you a car three or four years ago, you might be ready to buy the new model. It's probably worth quite a bit for them to send you an expensive communication and let you know what the latest choices are. That will happen. That interaction with the customer will define how effective a business is.

One of the predictions from the past that Xerox made about the paperless office -- I think it was right. Of course, they made it way too soon, and the use of paper has risen incredibly. If you go to my office, you wouldn't realize the paperless office is around the corner. But the fact that we can scan and find these documents, the fact that screens are going to get a lot better -- if you combine these things, it means it was absolutely right.

You have to think about different types of paper documents. The ones that will disappear first are forms, where you're filling in information. Where you have lots and lots of text, many pages of text, those will be the last to go. In fact, you'll have a hybrid where you'll use electronic distribution to mail those documents out to people. And then, because of low-cost, high-quality printers, you'll just print them out and be able to use them to read off of the paper.

A final prediction is that I think companies need to think about the environment they create for their knowledge workers. Knowledge workers are not like factory equipment. They're not cogs in a process. When you give them empowering tools, when you let them see the information, when you let them have the context of knowing what your company is planning to do and how their job relates to that, it can make an incredible difference.

People who talk about taking away the PCs from the people, who are not continuing to invest in giving them great flexible systems, I think that's a mistake. There will be a shortage of good knowledge workers. Getting them jobs that are interesting, because these tools make it that way, will be important both for recruiting and for getting the most value out of the worker.

It's an exciting time. It's a time of great change. And I'm certainly looking forward to how this unfolds over the next 10 years. Thank you. [Applause.]

Question & Answer Session with Bill Gates
View Vice President, Al Gore's transcript
View Steven Forbes' transcript