Last year, Scotts Miracle-Gro Co. noticed a higher than expected number of customers were canceling its lawn-fertilizer service.
Befuddled, the lawn-and-garden company sent out surveys to its customers, asking them to rate their satisfaction and provide feedback. Normally, Scotts would have tabulated the ratings and ignored most of the qualitative answers. Instead, it crammed thousands of surveys into Web-based software called Luminoso and made a surprising discovery: People were canceling because they wanted better customer service.
Though many people didn’t explicitly mention “customer service,” the software was able to associate seemingly unrelated phases like “listen” or “not responsive.”
“It digs a little deeper into what consumers are actually saying,” said David Erdman, a senior analyst at Scotts Miracle-Gro, who ran the program. According to Mr. Erdman, the company is now reevaluating its customer-service programs and planning to expand its use of Luminoso so that all divisions have access to it.
Luminoso Technologies Inc., a five-year-old startup spun out of the Massachusetts Institute of Technology Media Lab, is part of a larger effort to use artificial intelligence to dig deeper into the minds of consumers by analyzing focus groups and surveys, online forums and social media. Luminoso’s software draws from a large database of common knowledge and relationships (such as, “the sun is hot”) to understand how words and phrases relate to each other.