Artificial intelligence (AI), once the province of scary science fiction stories, is quickly becoming a standard business tool. In fact, people are interacting with AI more than they realize as businesses seek the scaling and speed that AI can bring to their brand messaging — in fact, a recent Gartner report predicted that by 2020 at least one-fourth of all customer service interactions will involve AI to some degree — a startling increase from just 2 percent in 2017.
But growing pains are showing, and there’s mounting evidence that while AI’s benefits are very real, the ideal scenario isn’t handing over your branding and messaging solely to thinking machines, but including some degree of human oversight.
More than chatbots
“AI is more than just chatbots,” says Manlio Carrelli, executive vice president of enterprise business at LivePerson, a leading messaging platform for brands. “AI can be designed to handle almost any type of inquiry a customer might have.”
Carrelli is blunt about what separates LivePerson from competitors. “We’re the only one that works,” he says. “One crucial difference is that we have figured out that the crucial thing to getting these technologies to work is having them supervised by human beings, and the best supervisors are the agents in a big business’ contact center. Agents know how to engage with customers, they’re trained on the brand’s voice, and they intuitively understand how to best solve a customer’s inquiry — even complicated ones.”
The human touch
One thing that human oversight helps AI overcome is the unpredictable nature of customer service interactions. “Humans are weird and unpredictable in the ways that they talk to technology and use technology,’ he notes. “If you’re an airline, a lot of your inbound volume is related to things that are happening with the weather. If you are a retailer, a lot of your inbound inquiries have to do with everyday product promotions, that sort of thing. So having that human in the loop is really critical to ensuring that the customer experience is awesome.”
“With LivePerson technology, agents are able to monitor the conversations between customers and bots,” Carrelli notes. “If the bot is struggling or the customer is getting frustrated, the agent can intervene and immediately take over the conversation. For example, our recently-launched Conversation Builder allows agents to update a bot to improve its performance. Even non-technical staff — like agents but also designers or business managers — are able to use it to create and update bots.”
From Carrelli’s perspective, the problem isn’t automation — it’s an over-reliance on automation. “I don’t think just spitting out some automation works — that’s a bill of goods that people have tried to sell for years. People who just think this is an engineering problem are wrong; it’s about humans working with technology. It’s a design and engineering problem — not only how do we teach computers to be able to have conversations with people, but what are the design principles that are necessary to make those conversations effective for the consumer and effective for the business? One of the key ways of solving that as we develop applications for the human and AI to work together.”
It’s not just about technology — it’s about the bottom line. “The companies we work with often see customer satisfaction scores increase 20 percent or more,” Carrelli points out. “Agent efficiency increases 2-3 times because our technology uses AI to help agents manage up to 40 conversations at once, and turnover among call center jobs dramatically decreases with conversational commerce — up to 50 percent or higher — because agents also enjoy messaging more.”
When it comes to brand messaging, the machine revolution is already happening — but the disruption is transforming, rather than eliminating, the human presence.