Real Time Speech Analytics

Do your call center agents make customers happy? Do they solve customer problems? Do they upsell or renew when possible? Do they ever get frustrated with customers?

And how about your customers? Do they threaten to leave you for a competitor? What are they complaining about? Do they end the communication satisfied or frustrated?

When it comes to agent-customer interaction, one communication that goes awry is one too many. There is a way around it, but many companies consider it too expensive, too complicated, or too telling.

I’m talking about real time speech analytics — technology from companies like CallMiner, Aspect, Nexidia and Salesforce that analyze voice calls in real time. Speech analytics systems can detect tone and sentiment of voice and talk/silence patterns to gauge emotion and satisfaction and tie detection of user-defined phrases to specific agent actions—in short, identify and prioritize what needs fixing, and then contribute to the resolution.

If you monitor agents’ interactions with customers, for example, you can easily ferret out which agents aren’t being proactive, or which are not succeeding in satisfying customers. You can then pull that call center agent aside for education or training in order to improve agent performance. If you identify from your speech analysis that certain types of calls are difficult for agents to handle, you can segment those calls and implement specific business processes, or otherwise treat them differently.

You can even set up the speech analytics software to listen for specific phrases and, based on those phrases, immediately prompt the agent with next best actions or proactive response alternatives. Call center managers can also monitor the system for the amount of times customers mention a competitor or say "Thank you". You can even monitor how many times customers are getting angry by their vocabulary and pitch of the customer’s voice during the call.

What’s more, you can use real time speech analytics to monitor tone and silence patterns, which can help ferret out when customers are frustrated or about to become angry. Tone also can signify age, which can be used to determine how effective a marketing campaign is on specific age segments.

Despite clear evidence that it makes a big dent in productivity and customer satisfaction, relatively few contact centers actively use speech analytics solutions. According to Ventana Research, about 20% of contact centers have deployed this technology, and another 16% plan to deploy it this year.

One reason for the slow technology adoption is cost and justifying ROI. Unlike the Session Initiation Protocol (SIP), which is an IT expense because it used throughout the company, speech analytics is seen as a contact center expense, which makes it an operations expense, and therefore harder to justify. But that's just not a good enough reason; according to research by DMG Consulting, which shows speech analytics has a payback of less than a year. How many other call center or IT software projects achieve a payback of less than one year?

It’s also partly the fault of speech analytics vendors, who have made the technology seem unnecessarily complicated. It really isn't; it's mature and definitely ready for prime time. It's also easy to integrate with CRM software and call center productivity systems - and it works.

So maybe it's the fear of the unknown. If you don't know exactly why customers are dissatisfied and leaving, you can't be blamed, right? Sorry, that's just not good enough anymore. Not with technology like this available.