The Real-Time Analytics Matching Program (RAMP) is an attempt at improving the traditional call center experience, where calls are routed via few menu options, such as which language the caller speaks and the type of product involved, but don't necessarily end up getting handled by an agent capable of resolving the problem.

The RAMP system's real-time analytics engine weights factors such as a caller's average wait time, payment history or spending patterns. It then looks for an ideal agent to take the call, based on their past performance and skill set. Prediction algorithms figure out when an optimal agent will be free, and the system sends them the call.

RAMP's capabilities could also be expanded by adding more types of data, such as "psychographic" details of a caller's historical demeanor, said Toby Cook, practice leader, IBM Analytics Solutions Team, IBM Global Business Services. "If they're traditionally cranky, and [the company has] that data available as a score input, you could match that to optimal agents."

RAMP helps a company determine which agents are the best, and also "takes this next tier of agent and tends to move them up to the top, because success garners success," he said. "Because you're matching them [with the most appropriate caller] they tend to do well. Even your lower performers tend to do better."