Practical applications of AI in the contact center.
The future of work and the impact that artificial intelligence will have on the workforce has been a hot topic lately. The publicity has been a bit unsettling, though, given the tendency to couple the acronym AI with the word “replace”—as in “AI Will Replace Half of All Jobs in the Next Decade” or “Could AI Replace Teachers, Lawyers, Drivers, Doctors, [fill in your job]?”
These types of headlines might lead you to believe that call center agents will soon become unicorns, along with other support-type positions that include a lot of repetitive tasks and activities that can be automated. But exchange the word “replace” with “enhance,” “augment” or “transform” and that may provide a better sense of how customer-centric brands are leveraging AI to support the human element in their service operations.
Certainly, there are still companies that view the contact center as a cost center, and for these, the thought of replacing hundreds of frontline workers with robots might seem like a magical solution to control the costs associated with staffing and managing human agents. But the reality is that, despite the influx of digital channels in recent years, customers prefer to speak with a human when it comes to service issues (confirmed by recent studies from Verint and Accenture).
Separate the Hype from Practical Applications
When considering how AI can be applied in a customer service function, it’s important to separate fantasy from the practical applications, says Mikhail Naumov, Co-founder & CSO at DigitalGenius, an AI customer service solutions provider. “There definitely has been a lot of hype in this space,” he says. “Hollywood’s version of AI may or may not come to pass some time in the future, but let’s focus on the application that is real and practical today—deploying algorithms that save time and help customer service employees to be more confident in their responses. We all want the same things—happier customers and engaged employees.”
DigitalGenius’ Human+AI™ Customer Service Platform is trained on historical customer service logs and provides AI-powered macro suggestions, automation of ticket tagging, auto-triaging, as well as automation of responses. “These are the types of tasks that AI is good at,” he says, adding that the combination of human and machine intelligence can cut average handling time by up to 32% per message.
What is the ideal model for a customer service operation? Naumov recommends a blend where AI suggests content and humans act as the administrators of the system. “A lot of the work is done for them by the machine but then, ultimately, when it comes to sending the final message or solving a complex query, the human is still there to supervise the machine to make sure that it’s doing the best job possible,” he explains.
Freeing CSRs from repetitive manual tasks allows them to tap into their critical thinking and problem-solving skills, and to focus on transactions that require an emotional connection. Take, for instance, TravelBird’s approach to AI. The package holiday provider uses the technology to empower their customer service representatives, deliver a better customer experience and engage staff in an evolving, higher-skilled role.
Eliminating Repetitive Tasks Frees Agents to Connect
Maintaining a human interface with customers is a central element of TravelBird’s service mission. “People travel with their hearts and souls,” says Head of Customer Service & Care Fiona Vanderbroeck. “But travel sometimes can be unpredictable. Things occur that create questions, doubt or stress for travelers. Our core job is to manage our customers’ emotions by responding quickly and making a personal connection.”
TravelBird’s customer service operation handles approximately 900,000 inquiries per year, which are answered in 11 countries by email, phone, social media and SMS. The company’s rapid growth has led to a considerable increase in message volume. Rather than replacing its customer service staff with bots or outsourcing its customer care, the company partnered with DigitalGenius to support frontline associates with machine intelligence.
The AI assistant provided by DigitalGenius relieves TravelBird’s frontline advisors of repetitive administrative tasks like categorizing and tagging the types of questions that customers ask via email. AI also saves advisors time when replying to frequent requests. “A travel advisor doesn’t have to write the same response over and over,” Vanderbroeck explains. “The AI presents the best available replies, which the advisor can select and tailor. Technology can perform those types of tasks better than a human, and therefore, it enables our humans to do what they do best—make that interaction a bit more special and personal.”
In addition to the customer experience and efficiency benefits, Vanderbroeck says that feedback from the advisors has been very positive. Eliminating repetitive manual tasks allows them to focus on the part of the job that they enjoy—helping customers.
Self-Learning System Provides Fast, Personalized Resolution
Increasing complexity in products, devices, applications and services has been making it almost impossible for companies to keep up with content changes or the support collateral that they might need, says JC Ramey, CEO of DeviceBits.
“Consumers are relying more on the Internet to find their solutions, which can lead to either positive or negative outcomes—but more importantly, companies have no way of knowing whether it was a positive or negative outcome,” he notes, adding that, in many cases, CSRs also are turning to Google to find the answers that they need.
How can companies regain control of their self-service while also managing increasing handle times as CSRs attempt to resolve more complex issues? Ramey suggests a hybrid approach that empowers both the consumer and the CSR with information that the business can curate and validate as accurate.
DeviceBits’ Care Assist product is an agent portal that allows agents to converse with customers via SMS, chat, web and social messaging. The software is powered by a self-learning AI engine that pulls content from the brand’s knowledge base, DeviceBits’ curated knowledge base, and from public sources, like Google and Bing, to try to find solutions to continually improve the agent’s experience. Over time, Ramey says, the AI “understands how end users are using the product, how agents are using the product and maps those user journeys through successful outcomes.”
Improve Content Consistency Across Channels
eGain’s SVP Worldwide Marketing Anand Subramaniam believes that delivering a consistent content experience is key to driving value for both customers and agents. According to a recent survey of 500 contact center agents by the customer engagement solutions provider, agents reported the top two obstacles to providing excellent customer service were finding the right answer to customer questions (26%), and different systems and information sources provide different answers to the same questions (25%).
Those responses closely corresponded to an earlier eGain consumer survey. In it, consumers were asked about the worst aspects of getting help from contact centers. The top three responses: (1) Different customer service agents give different answers, 41%; (2) customer service agents don’t know the answer, 34%; and (3) can’t find the answer on website, 31%.
The findings were not surprising, says Subramaniam. “We have these experiences fairly often as consumers. When you use web self-service, you often get the dreaded ‘no results found’ or 100 search hits that are only remotely related to your issue. Human-assisted service is equally challenging. Most recently, when I had to get something fixed in my house through my home warranty company, I had to contact them multiple times. Each time, I received a different answer. To complicate matters even more, I received different answers through the call center, digital and field touchpoints. This problem is commonplace, in our experience.”
What can contact centers do to address content-related consumer and agent pain points?
“Looking at the root causes from our recent surveys, it is clear that the biggest hurdles for great consumer experience and agent experience can be addressed with a robust, omnichannel knowledge management system infused with AI. This will provide consumers and agents with fast, accurate and consistent answers and intelligent process guidance, regardless of touchpoint,” Subramaniam says.
“Moreover, to develop empathy with customers and agents, center leaders should eat their own ‘gourmet food.’” he says. “They should conduct mystery-shopping as a consumer and ‘mystery-serving’ as an agent in their contact center to experience things for themselves. Empathy helps drive action!”
*Can AI Mimic the Brand’s Personality?*
Some businesses have invested considerable time, effort and expense into developing a brand personality. It is infused into marketing materials, logos, website design, policies and company culture—and even contact center scripts. Can AI pick up on and exhibit the characteristics and qualities associated with that personality?
Absolutely, says DigitalGenius’ Mikhail Naumov. It’s simply a matter of training, he points out.
First, historical data—which includes email transcripts, chat logs, email conversations, etc.—is fed into the deep neural network. “The expectation is that the brand’s tone of voice already exists organically within the historical log,” he says.
Next, the training continues when the algorithm goes live. During the initial deployment stage, the system provides suggested answers to agents for approval. Agents can approve, personalize or modify the answer to personalize it for a particular customer. “The algorithm trains on both—the agent’s actions, as well as any changes or personalization they make—to convey the messaging of the brand,” Naumov explains.
*AI Is Not New*
A common misconception about artificial intelligence is that it is a brand-new concept. That is far from reality, says Anand Subramaniam. “eGain has been a pioneer in AI for contact center customer service with many blue-chip clients using it for years, some for well over a decade. He offers the following examples of how its clients are using AI today:
- To converse with and answer customer questions using a virtual assistant: Virtual assistants can be useful in answering questions of low to medium complexity while providing an engaging experience.
- To guide service and sales interactions with AI reasoning: Where queries are more complex, AI reasoning can help guide the end-customer (in the case of self-service) and the agent (in the case of human-assisted service) through the next best steps to complete the interaction successfully. Guidance can also help to enforce compliance with regulations and/or organizational best practices. Moreover, it can be bootstrapped with context, leveraging information such as past customer transactions, current interaction context, IoT data, etc.
- To provide decision support with AI reasoning: Reasoning can also be used to help make decisions, where the decisions can be the output, or they can support a larger service process.