Web Enabled Decision Support Systems

By Donald Bodwell and Darell J. Herbst, PT Consulting Partners

The convergence of two technologies: the web and mobile computing has created an astounding opportunity: The ability to provide expert knowledge on demand-anytime/anywhere. Just imagine--your somewhat experienced troubleshooter is at a customer site and runs into something very unusual. She pulls out her iPad, logs into your company's web-based expert system that begins to pose questions about her situation. As she answers questions the system rules out the obvious causes, collects additional information about the problem by posing additional questions, and ultimately arrives at a diagnosis. The system then offers practical advice for resolving the problem. Your troubleshooter then applies the recommended solution and seeing that the problem is resolved, signs off, packs up and heads for the next customer site.

Initial steps in providing expert advice anytime, anywhere are being made by companies that are dealing with vast numbers of inexperienced, remote customers who are struggling with technology problems: Dell, Lenovo, and Hewlett-Packard all offer rudimentary assistance in solving installation, set-up, and update problems. Most of these examples fall far short of what is technically possible today. That is because very little effort has been made to understand the principles involved in building decision tree expert systems and learning how the web can be used to accelerate their construction and implementation.

From the beginning, some 25 years ago, expert systems have promised fantastic benefits. But mining these benefits required progressing up steep learning curves. The LISP programming language offered a typical example. Not only did you need a specialized understanding of a highly scientific and cryptic programming language but you also needed to invest in powerful specialized computers as well. Not so today.

Now it's much easier and more practical to capture vast amounts of specialized knowledge and make this available over the Internet. That is because HTML is uniquely suited to building decision tree systems: Each page can provide both information and choices that allow the viewer to navigate to the subsequent pages that rule out the obvious problems and collect ever more detailed information about the viewer's situation. And since HTML can handle diagrams, photographs, drawings, sound, and even video clips in many ways it can be more effective than your very best physically present human expert.

We built an expert system in the 80's that evaluated medical symptoms and provided expert computerized guidance to the user. Interestingly, the medical doctor who provided the protocols and subject matter expertise remarked that the system was actually better than he was, even though he wrote it. In turns out that this is so because in the process of coding the system, the doctor was forced to think through every possibility and figure out how to rule out the most obvious causes first. In real life he could never be that efficient or effective.

Beyond the medical and computer installation fields, the possibilities to profitably apply expert systems are endless: Anywhere large numbers of people are on the phones with customers taking orders, expert systems can provide the "best" or most efficient way of posing questions and getting the order right. In many instances, the human order taker can be virtually eliminated from the loop altogether. Amazon.com and Travelocity.com provide prominent examples. Anytime expertise is held in the minds of a few key people but needs to be made more readily available to others. Anytime people need to be provided with advice on how to do it, or what to do when the usual solution doesn't work, web-based expert systems should be deployed. Anytime organizational success is dependent on the ability to get the expert on the phone or get the expertise to the problem location, web based expert systems are the way to go. Those companies and organizations that want to keep such capability in-house can do so by restricting access to those who need it and make it available through the organization's Intranet.

Now mobile computing offers the ability to provide needed expertise anyplace a cellular phone can connect. From this point forward things will begin to be different. The dynamics of knowlege and power are shifting. Those companies that seize the opportunity to deploy expert systems will achieve true competitive advantage. We are just at the surface of a very deep pool of practical applications. We will witness a time when companies and organizations will operate much differently. Today most people are reluctant to consult an expert in their own field. Wouldn't that be an admission of lack of expertise on the caller's part? And thus performance is so much less than it could be. But the anonymity of the web protects the would-be caller and the day is coming when the first question of a sub-standard performer will be: Did you consult the expert system?

To learn more about how this technology can be applied in your company: Contact Us at PT Consulting Partners.