“That’s bad!” means something different to people of different generations and in different contexts. Sometimes it means something is bad, but in other uses it’s a statement of positivity. One of the key areas of natural language processing (NLP), a subset of artificial intelligence (AI) is sentiment analysis, the ability to understand emotional tones in speech and print. It is an area that is the focus for a number of different functional applications.
The most obvious need for sentiment analysis is in customer service. In a few recent articles, I’ve covered chatbots. In their basic version, NLP systems understand the basics of a question and then respond with a canned answer – with a basic sentence filled in with keywords such as a customer or product name. However, many people get quickly frustrated with and angry at chatbots and automated call distribution (ACD) systems that don’t understand how upset the person is, providing the same answer pattern to all people.
Unfortunately, this is not a problem that is limited to automated systems. Many call centers have live support personnel who are limited to very strict call scripts, limiting their ability to address customer concerns and frustrations. That means both more call time and higher customer turnover. The ability to analysis calls can help call center management.