Creating a Speech Application to Solve a Business Problem

 < Day Day Up > 

A Remote Business Problem

Slugger Sports produces a wide variety of sports equipment for T-ball and youth baseball games. It employs dozens of regional sales representatives who work almost entirely on the road. Eighteen months ago management had the information technology department implement a remote scheduling and ordering application. The company also purchased laptops and PocketPC devices for all of its sales representatives.

Every morning, the representatives access the application with their laptops to retrieve potential sales leads and build their daily schedule. Once the contact list is downloaded, they transfer the data to their PocketPC devices via ActiveSync. The system works great as long as everything goes exactly as planned.

Unfortunately, sales reps often encounter unforeseen glitches. Appointments may be rescheduled at the last minute or traffic jams make them miss opportunities. Existing customers sometimes ask about the availability of items they have never ordered. Quite often a representative is tipped off to a new lead during the day and feels the opportunity requires a schedule adjustment. Representatives may miss good information when new leads are added to the database after they have retrieved their daily schedules.

The representative often has to call back to the home office for information. In some cases the rep wastes time looking for an Internet connection in order to reconnect the laptop to the customer-scheduling application. The company is not willing to go back to the drawing board for a new solution, but wishes to supplement the existing one.

Case Study: GMAC Commercial Mortgage

GMAC Commercial Mortgage, one of the nation's largest loan originators, recently replaced a speech-based application used by its executives. The original application was used to access e-mails, contacts, and schedules from Microsoft Outlook.

The application was a success initially, since remote executives could use it to listen to and respond to e-mails using their cell phones. Some executives were able to clear out their in-boxes while driving to work.

The first problem was that the application was limited to twelve analog lines, so it did not scale well for the number of users who required it. Second, it was hard to debug and required that the developer make an actual call into the production server. Finally, it limited the dialog between user and application. Users were asked to speak one piece of information at a time.

The new speech-based application was written with the Speech Application SDK and runs on a Microsoft Speech Server. The new version still allows remote executives to access their information from Microsoft Outlook. However, it can now handle a 24-channel digital T1 connection.

The new application was developed quickly because GMAC's IT people were already experienced with Microsoft .NET. Debugging is no longer an issue since the developers can use the emulators to recreate errors and then step through the code till they locate the problem. Developers can resolve problems using their own development machines.

GMAC likes the new application because it fits well with the company's current environment and allows it to take advantage of emerging technologies. Busy executives like the application because they can manage their e-mails and schedules no matter where they are, saving them valuable time. The new application has improved the dialog between user and application since it can now manage phrases with multiple commands.

For more information about this case study, visit the case studies section on the Microsoft Speech Server Web site at

A Telephony Solution

The Microsoft Speech SDK (SASDK) 1.0, part of Microsoft Speech Server, enables developers to create telephony applications. These are applications in which the user will access data over a phone. This chapter will examine the source code for a telephony application built with the SASDK. The application should help Slugger Sports resolve problems with unexpected scheduling changes. It will allow sales representatives to dial in using their cell phones and to access product and opportunity information with voice commands. Because it is entirely voice based, the application can be accessed while the salesperson is driving or walking between meetings.

Since Slugger Sports is already utilizing Microsoft .NET, and the Speech SDK is free, the software costs will be minimal. Of course, there will be development costs. However, for experienced .NET developers, the time needed to develop a speech application should be lower than if a proprietary speech-development tool was utilized. Hardware costs are also small because each salesperson already owns a cell phone. The largest cost may come with setting up the telephony server.


This chapter examines a telephony or voice-only solution. Keep in mind that a multimodal solution, as discussed in Chapter 2 and featured in Chapter 4, could also work well for Slugger Sports. The company's sales representatives could use the PocketPC devices they already own to access the solution.

     < Day Day Up > 

    Building Intelligent  .NET Applications(c) Agents, Data Mining, Rule-Based Systems, and Speech Processing
    Building Intelligent .NET Applications(c) Agents, Data Mining, Rule-Based Systems, and Speech Processing
    ISBN: N/A
    EAN: N/A
    Year: 2005
    Pages: 123 © 2008-2017.
    If you may any questions please contact us: