Search Our Catalog of Articles

Q:  Are automated agents viable for customer self-service?
 
Companies have been working with automated agent technology to deliver cost-effective customer service for years. What many users have found is that the agent either doesn't deliver the right answer, or they get presented with a list of potential links to try.
 
Any good customer service solution must first understand the needs of the user to meet their needs. If the agent only understands a portion of what the user wants, the user will not have their needs met. Current approaches such as keyword, or pattern matching may understand a segment of the request, and therefore fall short. For example, the statement, "I want to buy a used car" may be understood as "I want to buy", and then a wildcard, "used", followed by "car", where the solution makes no distinction of the wildcard. The agent response may be, "You've come to the right place, we sell used cars." The same response would then be made for the statement, "I want to buy a stolen car"; "You've come to the right place, we sell stolen cars."
 
Humans converse in inconsistent, incomplete, and incorrect ways, but we still can understand each other (most of the time anyway!). How? There are many reasons, but at the core is the ability to understand the intent of what someone is saying, based on the context of the entire conversation.
 
The latest approach is through computational linguistics, or by understanding natural language. By combining linguistics, with artificial intelligence and machine learning, an agent can be created to understand the intent of the user, in complete context with the entire conversation and knowledge base. Using knowledge of language to immediately parse user questions and interpret the sentence. Through this understanding, the agent could be built to respond back intelligently, even asking qualifying questions of the user. By engaging in a dialog with the user, the automated agent can understand the intent of a users' request and can therefore deliver on their needs.