Hear that? It's the Evolution of Customer Service Calls

Like many people, the last thing I want to do is speak with a customer service agent when I have a question. Why call when the answer can most likely be found online in the FAQ section or through another self-service option?
Customer Strategy

Like many people, the last thing I want to do is speak with a customer service agent when I have a question. Why call when the answer can most likely be found online in the FAQ section or through another self-service option? At the same time, investments in speech analytics are rising.

The speech analytics market is forecast to grow from $589.2 million in 2015 to $1.6 billion by 2020, according to the research firm Markets and Markets. Companies are turning to speech analytics to make improvements in customer experience management, compliance management, and agent performance management.

So why invest in speech analytics when call volumes are dropping across industries? The answer is that while self-service solutions are improving rapidly, customers are saving their complicated questions for customer service agents. In fact, 83 percent of U.S. consumers prefer dealing with human beings over digital channels to solve detailed customer service issues, reports Accenture.

Anyone who has encountered an IVR system knows that these systems are ill-equipped to respond to questions that require more than a scripted response. And while the notion of A.I.-powered bots is promising, trials of Facebook Messenger bots demonstrated that the technology still has a long way to go before it's useful.

Balancing automated services with live agent support was a major topic at Verint Systems' Engage 2016 conference this week. One of the challenges is that people are inclined to expect only basic services from automated tools, noted Rick Merson, global operations director at Atento, a CRM provider. In fact, "if we automate too much people think the only way to get a genuine answer is to call," Merson said during a panel discussion.

Therefore, agents need more training and tools to adequately respond to the less predictable questions that customers are lobbing at them. Which brings us back to speech analytics. The speech analytics space is enjoying a rising demand as companies look for insights to boost efficiency and improve the call experience.

But implementing speech analytics requires a careful strategy, explained Alex Richwagen, corporate vice president of analytics at New York Life Insurance Company. At the conference, Richwagen shared lessons from his team's experience implementing speech analytics in a pilot program.

Similar to many companies, at New York Life, there was a two to three week gap between the time an agent spoke with a customer and a quality assessment manager identified opportunities to improve the interaction. Furthermore, only about four calls or so were being measured each month.

"When you consider that agents take over 1,000 calls a month, we had very little insight," Richwagen noted. Last year, New York Life worked with Verint to build a pilot program with 15 agents to test the value of analyzing agent performance with speech analytics. The company took the following steps to implement the program, which Richwagen recommended for introducing any new solution:

  1. Identify potential categories by which agents will be rated.
  2. Calibrate and test those categories.
  3. Identify gaps in the automated speech analytics solution.
  4. Create a scorecard. The company designed a scorecard that rated agents on whether they did the following during the call: showed compliance/ownership of the brand, active listening and confidence, summarized actions, and ended the call appropriately.
  5. Execute a communication strategy for introducing the KPIs to agents and explaining how it benefits them.

It's important to "test and calibrate" as much as possible, Richwagen added. "If you're going to tell someone, 'I'm going to rate you on this category,' you need to be confident it's accurate and so we kept re-calibrating the categories until we had a 99 percent threshold of accuracy," he said.

The results: the use of speech analytics freed up 40 percent of the quality analysts' time and the team was able to mine and score 100 percent of the calls in real time. The next step is to roll out the speech analytics across greater portions of the company, Richwagen said.

It's unquestionable that call volumes are dropping but telephony and human support remain important parts of the customer experience and companies that take a data-driven approach to optimizing the services have a key advantage over their competitors.