Speech Analytics Gets Real (Time)
with large volumes of calls and transcribing all words and phrases," Ziv says. "You're focused on identifying trends and root causes across multiple calls from multiple agents and customers, and you're not interfering with the customer or the agent, so the risk with post-call is much lower than in real time."
Real-Time Analytics Delivers Live Results
While buzz about real-time speech analytics seems to be everywhere, it is still very immature technology. It works in much the same way as post-call analytics, but listens to contact center calls or snippets of conversation at the same time that they are happening.
"Post-call speech analytics is a convenience, while real-time is tactical," Fluss says.
While a call is happening in real time, speech analytics is scanning for predefined keywords, extracted mainly from post-call analytics. If a keyword is said, it might trigger an action, such as alerting an agent, possibly a contact center supervisor or a retention team worker, who could then offer instant guidance so that the agent is able to take the next best step.
Jeff Gallino, founder and CTO of CallMiner, says that among the factors influencing his company to get into the real-time space were improvements in the underlying hardware and software that made it feasible and affordable. Gallino also says that, thanks to stricter legal compliance rules, there has been a surge in business. "In legal compliance [matters], we were having customers tell us that they were happy that the software told them after calls that there were violations, but it would be even better if they were told during the call."
Context is a major part of CallMiner's real-time speech analytics solution. While keywords are part of the equation, Gallino says, "you don't ever want to focus on single keywords, which may have a different context than what you think will happen in a situation.
"We're not doing keyword spotting, because we're looking at every word said, including the acoustics of how it was said, the underlying sentiment," he continues. "We'll accurately tell you when someone needs escalation, even if they don't say the words."
Gallino also maintains that providing call center agents with real-time, next-best guidance and knowledge-based software can "level the playing field for agents" and help with anything from product queries to competitive messages to upselling.
"Agents rarely know everything, and the turnover rate in the contact center space is very, very high," he says. "We can help next-best action, and that can also mean less agent training and getting them on the floor faster. After we make suggestions [for agents] a few times, people tend to get it."
Nexidia also offers agent guidance on the spot with its most recent speech analytics solution release, Interaction Analytics 11.0.
Pellet agrees that compliance is a key sector for employing real-time speech analytics, and adds that agent focus is also, in his opinion, another important category.
"It's essential that real-time makes the agent's life easier while making the customer interaction better," he says. "Contact center interactions are getting more and more complex. Being able to take action as the call is happening is critically important. Customers don't want to make two calls; they want the problem figured out the first time."
Listen Before You Leap to Real-Time
Organizations may think that real-time speech analytics sounds wonderful, and it can be, but industry experts advise companies not to put the cart before the horse. In other words, companies need to see what has happened in a contact center after the call is finished, and that still holds true.
"Most of our customers start with post-call and add real-time afterward," says Jason Napierski, product marketing manager at CallMiner. "They want to use post-call to get their house in order and then use real-time to get their actions in order."
Ziv advises companies to start the speech analytics selection process with post-call information to understand their environment.
Related Articles
Enhancements provide visual data analysis and improved tools for review and workflow for the legal market.
02 Feb 2015
Nexidia AudioFinder will help the law firm and collections agency ensure compliance.
27 Jan 2015
Increased targeting and precision make an upgrade worthwhile.
10 Jan 2015
Real-Time Speech Analytics can provide agents and supervisors immediate feedback during the call.
17 Oct 2014