Think about the last time you called in to a customer support line for help. Do you remember what your experience was like? I sure do. I bought an Amazon Firestick during Prime Day (aka Christmas in July). I tried installing it a few days after it arrived, but something wasn’t working. Long story short: It took five separate calls to Amazon’s customer support team over a three-day period to get a new one.
Even if consumers have bad experiences with contact centers, most still prefer to speak with a human representative – 70%, to be exact – according to a study conducted by the Sitel Group.
With that being said, contact centers play an influential role in customer retention strategies. In fact, research conducted by Forrester suggests a contact center can hold its own against aggressive ad campaigns, by noting that 3,000 agents averaging 50 calls per day amounts to more than 54 million contact points with the customer base in a year.
Within call data lies rich insights about customer satisfaction, customer churn, competitive intelligence, service issues, agent performance, upsell opportunities, campaign effectiveness, and more. So, how are companies supposed to get all that information from call transcripts?
Speech analysis is essential in terms of understanding customer feedback given during contact center calls, as well as sentiment overall. Stratifyd’s speech analytics functionality solves the constantly expanding need for enterprise clients to transcribe and analyze voice data.
How Does It Work?
Stratifyd takes dual-channel recordings of contact center calls and applies AI modeling, Natural Language Understanding, and acoustic modeling to develop transcriptions and analyze voice metrics, such as call length, talk rate, overtalk, silence, energy level, and more.
The transcriptions and voice metrics are combined with all other data – chat transcripts, surveys, star ratings, CRM, etc. – to deliver a holistic view of the customer in minutes.
Here are three ways companies can leverage speech analytics to be innovative in their industries.
Silence Versus Overtalk
Have you ever been on the phone with a customer service representative when they ask for specific information? After giving said information, what happens? Most of the time, nothing happens (at least for a few seconds). Contact center gurus refer to these instances, when callers experience a pause for more than four seconds, as a silent event.
And if a caller and representative continuously interrupt one another, it’s called overtalk. This metric is tricky because it could indicate excitement or displeasure depending on the nature of the call, so we look at the frequency of the two individuals’ voices and choice of language, too.
Our platform tracks instances when there were long pauses during a call, how often the pauses occurred, and how long the caller and agent spoke uninterrupted. By combining these metrics with other data sources, companies can determine whether an agent was recording information, following along with or deviating from a script, or if an argument ensued, leaving the agent frazzled and unable to calmly proceed with the call.
These valuable insights help to train agents to make sure they’re better prepared to handle calls, no matter the customer’s mood or experience.
Tone of Voice
One’s tone of voice allows us to determine how a person is feeling in an instant. And conversations between customers and representatives must be tracked and properly measured for quality assurance.
Our speech analytics technology is complete with acoustic modeling to gauge energy levels and indicate feelings of excitement or anger based on tone of voice.
One of our clients is a person-to-person insurance company that sells policies through telemarketing. On average, it takes four rounds of calls before an agent can make a sale. The company utilized our speech analytics functionality to create an AI model that determines an individual's interest and indicated whether a follow-up call was warranted. After one week, the agents using the AI models improved sales efficiency by 50% on average.
Sentiment With Speech and Text
Speech analytics helps enterprises paint a full, vibrant picture of customer sentiment.
Stratifyd focuses on the omni-channel view of customer analytics, and speech analytics plays a big part in that framework. By combining the customer voice with omni-channel structured and unstructured data sources, companies can better address customer pain points to improve customer retention, acquisition, and experiences.
Speech Analytics Completes a CX Masterpiece
Customer sentiment is a lot like pointillism art, composed of thousands of other images to make a larger, more captivating piece. Without the smaller, individual metrics like voice and textual data, customer sentiment is never fully complete.
Our AI-powered platform uncovers insights and delivers real-time feedback that can be used to facilitate the ultimate CX across all channels. To learn more about how your company can benefit from speech analytics, click here.