In this article...

• Cognitive assistants automate routine customer care, freeing live agents to focus on high-value interactions.
• Cognitive insights help contact centers improve operational efficiency.
• Cognitive solutions streamline multi-channel support, including social media.

Effective customer care is premised on efficient first point-of-contact, but touch-tone and interactive voice response (IVR) systems often provide no value, and are often seen as irritating barriers between the customer and a live agent.

This experience is probably familiar. Seventy-five percent of customers say it takes too long to reach a live agent. The average contact center routes 32 percent of calls through at least four menu levels. Ninety-one percent of customers have called customer service multiple times for the same reason, 90 percent have been put on hold for too long and 89 percent have been forced to repeat information to multiple agents.

The damage done by poor customer service is huge. U.S. organizations spend $112 billion on contact center labor and technology each year, yet half of the 270 billion incoming customer service calls go unresolved. This failure results in $41 billion in losses.

In addition, customers are more demanding than ever. They expect immediate, personalized and convenient service. Customers often defect because of bad customer service, sometimes after one negative experience, and many will never come back. Poor customer experiences are shared publically twice as often as good service.

It’s no surprise then that the top priority for contact center executives is to improve the customer experience. Contact centers have turned to cognitive technology and artificial intelligence to meet these pressing challenges. In fact, IBM research shows that early adopters of cognitive technology report improved customer service as a top outcome.

Cognitive assistants automate customer care

Unlike touch-tone and IVR systems, cognitive assistants simulate live agents. They mimic how the human brain works by mining data, recognizing patterns and learning from experience. They offer personalized interaction through natural language conversation. They are emotionally intelligent. They detect sentiment and respond appropriately to tone-of-voice. They learn from experience and improve over time.

Cognitive assistants are also fast, accurate and versatile. They are trained to become experts and to scale the expertise of live agents. Almost any customer service channel can integrate with these systems. They can be used either to support live agents during calls or to interact with customers directly. They tailor interactions based on call history, purchase history and other data to establish patterns of care.

However, cognitive assistants don’t eliminate the need for customer care employees. Routine care is best delegated to these systems, but complex and unusual requests require intervention from live agents. For example, wealth advisors tend to spend most of their time consulting high net-worth clients. Cognitive assistants bring this expertise to more customers without placing demand on a wealth advisor’s time.

Volume Global, a technology provider that develops custom applications for large global clients, recently focused on enhancing human performance through smart machines and infusing cognitive technology into customer care. Volume CEO Chris Sykes concluded that when routine care is automated by cognitive assistants, live agents focus on complex and valuable interactions. “It’s not replacing,” Sykes said. “It’s optimizing. It’s taking away the most mundane tasks and putting our brains to better use. Many contact center agents are paralyzed into a robotic state. They read scripts and deal with dissatisfaction day in and day out. Many agents handle 20-30 calls from frustrated customers every day. Cognitive assistants take that away, which makes for a much nicer job environment.”

Volume also developed a cognitive assistant for BP Castrol to support a customer service line for technical requests. “We found that technical solution experts are asked the same questions 60-70 percent of the time,” Sykes explained, “and the questions are not very technical at all. They are things you could find easily on the website. The cognitive assistant gives customers 24/7 access to resolve these issues. More importantly, it allows the technical solution expert to focus on the 30-40 percent of questions they have to really think about, not the 60-70 percent they have already been asked 80 times that week. That’s not very fulfilling for them.”

Cognitive insights enhance operational efficiency

Mitsui Sumitomo Insurance (MSI) makes customer interactions more efficient with insights uncovered by cognitive technology. MSI traditionally relied on temporary contract labor during high-volume call seasons. While this approach successfully addressed high call volume, labor costs rose and customer satisfaction metrics remained flat. Instead of adding seasonal staff, MSI wanted to identify patterns and optimize resources.

The MSI cognitive solution analyzes the content of inbound calls, which represents an enormous volume of unstructured data. The solution recognizes trends and sentiment, which allows MSI to respond by publishing more insightful front-end customer messaging on frequently asked questions (FAQs) webpages or front-end voice recordings. These response measures, the company has found, greatly reduce the volume of inbound calls.

MSI trains its agents to respond effectively to common inquiries based on cognitive insights, and arms its agents with timely information based on customer behavior patterns. For example, the cognitive tool discovered that policy cancelations occur more frequently after customers are insured for one full year. However, the same customers who cancel at the one-year mark are also more likely to purchase additional insurance products, which makes them good upsell prospects. Identifying these customers allows agents to take proactive steps to lower attrition and increase sales.

Overall, thanks to cognitive solution, MSI improved its contact center response rate by 11 percent, improved its customer satisfaction rating from two-star to three-star and substantially decreased it seasonal contract labor costs. “Cognitive technology has given us a whole new perspective on the voice of the customer,” according to an MSI executive, “and has given us insights that have improved our operational effectiveness.”

Cognitive technology streamlines multi-channel support

Although most customer care interactions take place on the phone, the customer service experience is increasingly multi-channel. Digital will overtake phone interactions within two years. Social media is quickly becoming the first choice for digital customers, but 60 percent of contact centers don’t support channels like Twitter, Facebook and Yelp. The public nature of these websites as customer service forums presents both the opportunity to strengthen the brand and the danger of a public relations firestorm. Many major consumer brands have integrated social media into their customer service strategies. However, they often lack the resources needed to manage overwhelming volumes of social interactions.

Cognitive technology can streamline triage through its ability to understand natural language. HelpSocial, Inc. develops cognitive tools that sift through and prioritize this unstructured, often chaotic stream of data. These tools apply natural language processing to social content to understand not only literal meaning, but also sentiment and intent. They identify sensitive topics and strong emotional expressions in order to route these interactions to appropriately skilled agents.

For example, an airline customer may post a sarcastic tweet regarding lost baggage. Using sentiment and emotion analysis, the cognitive solution can detect how upset the customer is, and whether they are feeling anger, disgust, sadness or disappointment.

The tool can help determine whether or not a response is urgent and if the customer should be assigned to an agent with expertise in handling frustrated customers. Armed with these insights, an informed decision can be made to determine, for example, if a response would be more suitable from the marketing or customer service department.

This automated prioritization filters out the noise and reduces the burden on live agents, allowing them to focus on legitimate requests. This new model reduces cost, improves response time and boosts customer satisfaction. “Using cognitive computing can automate the handling of growing data streams from social channels,” a HelpSocial executive said. “This allows agents to do more strategic work in the contact center.”