If customers believe that your on-line knowledge base makes it easy for them to locate answers to most of their questions about your company, then they get into the habit of consulting your knowledge base as their primary source of support. You can similarly encourage customers to rely on a speech recognition system to automate the way they retrieve answers.

But there will always be customers who prefer to hear answers directly from people. Simply offering these customers an on-line knowledge base or an on-line trouble ticketing system may not be enough to convince them that there are alternatives to calling support reps for help. That's not a bad thing; for many customers, what matters is whether your company offers live alternatives to automated assistance.

Nor is it a bad outcome when customers seek help from agents after viewing your knowledge base. Customers' traversals through your knowledge base reveal a lot. They indicate how customers communicate their support requests and where in your knowledge base customers expect to find answers.

If you let agents see what areas of your knowledge base a customer has viewed, you enable agents – and your company – to learn from the customer whether the knowledge base can convey certain information more clearly. You also enable agents to identify specific circumstances when customers are more receptive to answers that come from people rather than from machines.

Automation is most effective if it complements, rather than replaces, live service and support. By presenting customers with on-line knowledge bases and trouble ticketing systems, you're not automating service, but rather the delivery of answers to questions and requests for help. Customers only consider the automation of these processes to be beneficial if they think that the answers they receive are helpful, or that agents are sufficiently responsive. When agents make your call center look – or sound – good, you increase the likelihood that the technology you deploy can, too.