A Question of Customer Service

Opinion: Conversagent's Automated Service Agent is an example of how a self-service app can provide the first level of support interaction to direct people who want support to appropriate answers.

Providing good customer support is one of those things that is so difficult to do, yet so essential to controlling costs, that its no surprise companies almost universally do a poor job.

Most support organizations feel the pressure to keep customer interactions to a minimum to keep costs down, but at the same time, to use customer interactions to create revenue opportunities.

Self-directed help is an area where companies could save substantial amounts of money if they could do an excellent job providing information. The tough part is directing users through options without making it a time-consuming activity.

The problem comes from an over-reliance on FAQs and undirected search solutions. It sounds like a good idea to answer users frequently asked questions, but most companies do FAQs poorly. The FAQ section gets bumped to a hidden location shortly after a product introduction, answers questions people dont ask, or includes far too many questions and answers.

Faced with this lack of visible information on a companys Web site, users might try to find what they need through an Internet-wide hunt—namely, a Google search. But in some ways, Google is one of the worst things that happened to search. In the days before Google, directed search, while not common, was a methodology that had a following and could, if properly done, yield good results.

For the most part, directed natural language search has given way to best-guess searching—with a couple of notable exceptions.

Conversagent Inc., for example, provides a service called ASA (Automated Service Agent), which uses natural language through a chat interface to direct people who want support to appropriate answers. One nice aspect of Conversagent ASA is that, when done right, it gets past the problem of vernacular. For example, Conversagent ASA provides the first level of support interaction for Comcast on Comcast.net. To see it for yourself, click the Ask Comcast link on the homepage.

In that instance, the expectation is that customers will interchange product names and product types fairly frequently when asking questions about e-mail, for example.

Conversagent ASA does a good job directing the customer to make a choice about an e-mail client through interactive response lists to narrow the list through likely answers to the right answer. Another nice aspect of the way Comcast has put Conversagent ASA to work is that it will connect customers and support agents when it senses escalation is required, either because of the vagueness of the question or an inability to find an answer.

/zimages/2/28571.gifMicrosoft will apply ICE to VOIP. Click here to read more.

Although a less interesting use of ASA, Panasonic uses the service to provide information about Panasonic plasma TVs to potential buyers. This is a mixed bag from a consumer standpoint, because the selling is heavy-handed, but it certainly provides better answers than the buyer would get in a retail scenario. I also found that the ASA service provides information more quickly than searching on Panasonics Web site would. Of course, its also less expensive than having to staff a customer support center as well. Check it out on the Panasonic home page by clicking the Ask Panasonic link.

Technical Analyst Michael Caton can be reached at michael_caton@ziffdavis.com.