In customer service it is important to achieve an equilibrium between automation and the human touch. A robot can be of great value. But there comes a point where humans provide the added value, and are able to prevent frustration by the customer. We talked about bots and customer service with Darren Parker, Director EMEA Enterprise Customer Success at Zendesk. The key is the classic ‘know your customer’.
“Our customers face several challenges,” says Parker. “Firstly, it’s just meeting the expectations of their customers. Consumers are getting more and more demanding in what they look for from a potential supplier, and much less accepting of failure if you’re lacking customer service. From a survey we did,we learned that 31% of customers will change supplier after one bad experience. And actually, Benelux is pretty forgiving.
“In France this goes up to 68%. And then, after two bad experiences in the Benelux region, 77% of customers will leave and don’t come back. This means we need to have a good definition of what good customer service actually looks like. And basically, it’s getting an answer as quickly as humanly possible and making it as easy as possible.
“The second challenge for our customers has to do with unpredictability. Look at what we’ve just gone through in the last year. So being able to scale up and scale down, really quickly change channels, add new channels in, and be able to do that quickly without having to go and hire a whole load of developers and spend six months implementing is important for all of our customers. So democratization of customer experience actually isn’t just something for a large enterprise. Of course they have the biggest budgets. But that doesn’t mean that this technology isn’t available to smaller companies.”
Challenge of bots
“When you have a knowledge base as a company, you want to be able to use that in the most appropriate way to your consumer. So when a consumer asks a question using certain key words, technology can help to find the content within your own library, and serve that to the customer. If that does not solve the question we might go through some structured questioning.
“And ultimately, you’re going to get to a point where either the customer has resolved their question, and you’ve given them the right data, or it’s clear that they haven’t. Now they don’t want to be given another load of articles. So now you need to be able to offer a different channel. I think this is really the challenge of bots. Automation is really knowing at what point to draw that line. You don’t want to upset your customers and frustrate them.”
He emphasizes one of the keys is helping customers to understand what’s working and what is not. “People always talk about the importance of data and analytics being correct when you set up automation. But to know your customer, knowing where those queries are coming from, is equally important. Know what’s appropriate to solve from an automated point of view, and what needs to be pushed very quickly somewhere else. And of course, you need to continually review that.”
Machine learning
At the moment most of what is happening with chatbots is scripted, but artificial intelligence (AI) and machine learning (ML) are on the rise. “This can help in better understanding the links between key things and key expressions that are used. That way we can focus on a larger pool, rather than just on your own knowledge base. Algorithms can drive better match rates for the customer types in the response.“
Customers are getting more and more used to interaction with bots, and become a lot better at phrasing things in a way that’s more likely to help them. “As the matching gets more sophisticated, we’ll be able to suggest more answers to the questions that come in. And be more accurate with that. But for me, there’s always going to be a point where the fastest and most efficient way to resolve this could be another channel. That doesn’t necessarily have to be a live interaction with a human, it can be synchronous or asynchronous, it can be chats, and it can be email, it can be a telephone call. The key thing is to really understand your customers, and to meet them where they want to be to met, in the channel they want to talk to you, with the information that they need.
“And please don’t offer them advice you’ve already given them through the chatbot. You’ve got to know what they’ve already seen. So you don’t ask them the same questions and give them the same options again. This means the whole thing feels like a more natural experience for your customer. At the end of the day, they’re just coming to you as a company for support, and then it becomes seamless to them until they get a resolution.”
