Multi-Content Search and Rapid ROI

Keeping up with customer support issues is critical to Inference’s long term strategy. In recent years, industry trends relative to customer contact center activity have pointed to some interesting developments.

Table 1 shows that a high volume of calls concerning common, everyday problems occur in the customer contact center. These common problems are resolved either in a self-service format involving highly structured, case-based solutions or through junior, call-center agents applying similar methods. The point here is that common problems require a narrow range of solutions, are handled cost-effectively, and represent the vast majority of calls to the contact center.

By contrast, more complex problems require a much broader range of solutions and must be handled by well-paid, senior customer support agents who work in a less structured environment to diagnose problems and offer interpreted solutions.

Table 1. Customer Contact Center Activity

This scenario represented a twofold challenge:

• How do you provide a multi-content search that offers both a structured solution for common problems (self-service end-users) and a less structured solution for more complex problems.
• How do you deploy this application with a rapid return on investment (ROI).

To meet this challenge, Inference developed the k-Commerce Support Enterprise product featuring the Knowledge Gateway. The Knowledge Gateway meets both the requirements with a hybrid search system that gives you a multi-content search and fast ROI.

The hybrid search system takes an enhanced CBR product that covers the self-service issues and combines it with the new Knowledge Gateway. The Knowledge Gateway taps into the enormous investment most companies have made in their document archives located on intranets, extranets, LANs and WANs across the enterprise.

Rapid ROI

Building highly structured cases can be time consuming and require a major commitment of resources. When Inference built the Enterprise product from the ground up, reducing this commitment of resources was a prime consideration. Inference responded with the Knowledge Gateway to get customers up and running faster. This means that, once the Knowledge Gateway is installed—in two or three weeks—companies can go forward with a case-base, self-service solution when time and personnel permit. In some situations, companies using the Knowledge Gateway can avoid building cases completely. The following suggests an action plan:

1. Choose your company’s three most important information sources: For example: customer call logs, tech notes, and email.
2. Use Site Server or Lotus Notes to go out to these information sources and index them into related collections.
3. Using the k-Commerce Support Enterprise authoring tool, build a concept taxonomy based on the words and phrases pulled from the document collections.
4. Start using the k-Commerce Support Enterprise Knowledge Gateway to search them.

This proposal has some compelling advantages:

• Return on investment is almost immediate. In two to three weeks, you can build a concept taxonomy and you’re in business.
• The cost of building structured cases is reduced because you only need to build cases for the most frequently recurring problems.
• The cost of building cases for these common problems is reasonable because the subject is well-known and less complex.
• Using the gateway search instead of structured cases to solve uncommon problems amounts to very significant cost savings.

Taxonomy Building

As concept taxonomies are key to the question generation process and to other steps along that way, customers want to know the tools with which to build one.

• Taxonomy construction requires the k-Commerce Support Enterprise authoring tool. Inference offers a taxonomy building tool that does an analysis of your searched documents and proposes potential taxonomies from the words and phrases in the documents. This utility is, however, a semi-automatic means of preparing a taxonomy and requires the authoring tool to supplement it.
• One of the important features of K-Commerce Support Enterprise is its ability to share one taxonomy across all components that need one. For example, a single taxonomy accommodates as many databases as the organization may want to add.