Automated online self-service means more customer answers, less cost

Conversagent’s automated agent has cut attended service costs by 15% to 20% in other industries. With a consumer electronics maker about to go live, it moves toward retail.

Mary Wagner

Online customer self service is frequently offered in one of two forms: an FAQ, or a searchable knowledge base. Either option may supply a long list of results that force a customer to work to get his or her question answered, prompting a phone call at a cost of $4 to $5 to the company – or an e-mail. But automated online customer service provider Conversagent Inc. says its system, a natural language processing search engine populated with a company-specific knowledge base, has reduced the load on attended services, whether calls, e-mail or live chat, by 15% to 20% on average where it’s been implemented.

For some customers, says CEO Steve Klein, that’s meant a savings of $1 million a month on an annual $750,000 investment. Conversagent, with clients in the telecommunications and financial services industries, will shortly announce its first customer in the retail arena, a consumer electronics maker and marketer that will use the system both to support the purchase process and for post-sale customer service, Klein says.

Klein, who notes the system is particularly suited to marketers with a large customer base and a complex product or products, says those attributes are what helped drive the company’s expansion into the area of consumer electronics. “We’ve had numerous conversations with online retailers, but there is so much on their corporate shopping list in terms of merchandising or shopping cart technology that customer service is slightly overlooked. That’s why consumer electronics retailers have become great territory for us. They are using 10-year-old technology to solve the problem,” he says.

The system loads in a company’s own content--likely questions from customers about products or services, and their answers--and automatically overlays that with natural language processing technology that recognizes syntax and spelling variations. One component, for example, is an industry-specific library licensed to system users that reconciles synonyms. That allows the customer question, "How do I phone China” to call up the same correct answer as “How do I ring China” or “How do I call China,” for instance, without the company having to figure out those variations and enter it into the system itself.

“Customers that don’t get their questions answered will end up picking up the phone,” says Klein. “We help companies avoid the cost of those calls and e-mails. Just as importantly, we empower those customers to get answers to questions on their own without being slowed down by a phone call or e-mail.”


Customer intelligence, customer service, ECRM, Electronic commerce, Online shopping