The growing field of social media monitoring has a relatively new entrant called In the Chat, which announced a new product Thursday.
The company's enterprise solution works with Twitter, Facebook, forums, blogs and Google+. It analyzes and categorizes millions of social media posts to find mentions of a client's brand and products. Then it picks through the posts to find clues to customer service problems, prospects who might be interested in a company's products, and customers about to leave. The software then provides this insight to customer service agents, who can follow up with customers privately, through a chat tool or phone call.
"A lot of conversations that happen in the financial world are private, you wouldn't share private data over the web," points out John Heuhn, CEO and founder.
Call centers are one of the most expensive line items for banks, he notes. This software could potentially turn those customer service agents into salespeople.
Human beings do the categorization for In the Chat. Those categorizations are then fed into an engine that applies keywords and such automatically. "The first time through, they do a terrible job," Huehn says. "If there's something new that shows up, the algorithms will shoot the post back to the experts to apply categories." For instance, the use of the word "gingerbread" didn't exist within the telecom world until Google introduced the Android Gingerbread operating system. Humans noted that and added it to the taxonomy. "TD" in a tweet could stand for "touchdown" or TD Bank. The algorithms are designed to parse out bank references from game results.
The new product In the Chat is bringing out today is a cloud-based version of the software called
"We're trying to make it easy for customers to dip their toe in the sales and service spot," Heuhn says. At first, the cloud-delivered software will work only with Twitter, analyzing tweets to determine who's planning to leave a company and who might buy from a company. Later, it will expand to other social media networks and additional use cases.
