Simple questions are the bane of the lives of many customers and call centre agents. Having to wait in a call queue to find out a company’s returns address or your contract expiration date, for example, is a frustrating experience. And for call centre agents, these interactions can be repetitive and dull – a job more suited to a machine. Or rather, a chatbot.
Also known as virtual agents, chatbots have become a popular tool for solving common customer requests. People no longer need to trawl through FAQ sections for basic information when a chatbot can do it for them in the blink of an eye. And if it turns out their inquiry is rather more complex than it first appeared, intelligent chatbots can transfer them to a human being ready to help.
Also, unlike humans, chatbots can be on hand 24/7. “If you need an answer at 2am, is it better to have a 40% chance of getting the right answer from a chatbot or a 0% chance of getting any answer from contact centre agents because they aren’t working?” asks Ian Jacobs, vice-president, research director at Forrester and co-author of the report, How AI and automation drive better customer service experiences.
In the past, chatbots may have struggled to provide useful answers, but today’s modern examples, such as ServiceNow’s Virtual Agent, use natural language understanding (NLU) to resolve common requests and can integrate with almost any chat or messaging service.
NLU programs ask customers to identify what’s wrong, then they process that query rather than asking the caller to choose from options that don’t accurately capture their problem. For instance, if a student wants to pay accommodation fees to their university, a virtual agent equipped with NLU should be able to pick up on that fact regardless of whether they say “pay accommodation”, “housing fees” or “pay rent”, and share the relevant knowledge article that tells the caller everything they need to know. This saves the organisation the cost of having an agent online to take the call and deliver a routine answer.
“With natural language understanding, your question or your problem may be able to be solved directly by the [virtual] agent,” says Ian Ashby, principal strategist for customer service at ServiceNow. “It’s a really good example of how combining virtual agent technology with search and natural language understanding can be a core component of providing a better service for the customer, benefits for the organisation, and frankly [making] a happier agent.”
Freed from the need to answer basic questions, agents can focus on more high-value “human” tasks, such as providing a better service to customers with complex problems or requests.
However, it’s important to note that today’s chatbots are still open to improvement. Even with NLU, speech-to-text technology sometimes struggles to recognise accents, slang and sarcasm. And while chatbots are perfect for dealing with straightforward questions, more complex or emotionally charged situations are better served by a human – something a chatbot may well suggest itself. “We’re at the point in time where the machine can go: ‘Are we getting anywhere here? Would you like to speak to someone?’” says Ashby.
Chat history and related context can be seamlessly transferred to a human agent. This allows them to jump into the conversation with a full understanding of the situation and quickly resolve the request. In fact, chatbots can even help agents by pulling information from the company’s systems, which can be used to add detail to an agent’s response to the customer.
“These suggestions can be pieces of knowledge; they can be media such as a YouTube video that shows the customer how to do whatever they need to do; they can be process maps that walk the agent through all the steps needed to solve the customer’s issue,” says Jacobs.
In future, customers could even be passed back and forth between an AI and a human agent in a way that best meets their needs, he explains, giving an example of what that interaction might look like. “I called my telecommunications company because my internet service is down. I talk to a human agent who thinks the issue may be my router. She tells me that she’s connecting me with the router diagnostic chatbot, which then takes over the conversation.
“It tells me it will ping my router and it needs me to tell it what lights are flashing. I type in the requested information and the chatbot says that the diagnostic indicates that I need a firmware update. It will bring the human agent back on the line to help walk through that process. Instead of automation to human, we now have a process that is human to automation to human.”
Of course, it’s also important to cater to customers who don’t want to interact with chatbots. “There’s obviously a demographic issue, how comfortable you are [with this technology],” says Ashby. “But if you can deliver a service to a subset of customers who are happy to use it for a subset of their issues, it makes a huge difference to the customer experience and the organisation as well.”
The way we work is changing. Find out more at servicenow.co.uk
