Return to site

Understanding the type of experience your customer wants

How A.I. powered chatbots make a better customer experience

· ai,wechat,chatbots,miniprogram,cx

Engaging in customer support is an experience that can be unpleasant and overwhelming for both customers and brands. Customers get annoyed when they have to wait for simple answers, while customer support people feel overwhelmed by the flood of people who have questions.

The obvious, but not so simple solution to the problem is to hire more people. If you have a larger staff, each customer support person has fewer customers to take care of, and thus they can answer questions faster. But as manager will tell you, this won’t necessarily solve the problem.

In a world with finite resources, the key to making the customer support experience better for both the brand and the user is to understand the type of conversation that the user wants to have.

In general, when a customer is looking for support, they are often looking for one of two types of interactions.

1. Transactional responses (I need an answer right now!)

  1. I want to cancel my service?
  2. Where can I return my hardware?

2. Emotional conversations (I need to talk to someone)

  1. I’m upset with my service and I need to talk to someone
  2. I need help to figure out how to solve my problem

Chatbots to the rescue

Businesses have quickly adopted chatbots powered by Artificial Intelligence to help to solve this customer support problem. Since chatbots never sleep, never eat, and never go to the bathroom they’re cheap, and because they only need electrons, they are theoretically infinitely scalable.

Not so fast…

But while chatbots are cheap and scalable, they cannot replace a customer support team. Humans are great at providing emotional responses to users, but can struggle, or become frustrated providing a stream of transactional responses to users. Chatbots on the other hand are great at providing transactional responses to users, but lack the EQ to pacify an angry customer.

Over the past two years we’ve worked with companies both big and small to give transactional responses when the user wants it, and emotional conversations when the user needs it.

Blending chatbots with customer support agents

We’ve found that designing a customer experience that blends chatbots powered by artificial intelligence with live people allows companies to...

1. Give transactional users the quick information that they need
2. Give emotional users the attention that they deserve

How can I help you?

The key then is to quickly identify the type of response the user is looking for.

“How can I help you?”

While this is a simple question, the users response to this question goes a long way to determining if they are looking for a transaction or emotional response.

Fortunately, the chatbot is great at identifying what type of response the user is looking for.

How do we know how to respond?

By blending the chatbot with live people we can turn a chatbot bug (chatbots can’t understand everything 😪) into a chatbot feature (chatbots can’t understand everything! 😁).

We do this by training the chatbot to first analyze all the incoming user questions. If the chatbot can identify a transactional request and then it will respond with a relevant answer. If, however, the chatbot doesn’t understand how to respond to a user, we can assume that the user is looking for a more emotional interaction. At that point, the chatbot alerts a customer support person to come and help the user.

While chatbots still have a long long way to go before they replace humans when it comes to providing an emotionally satisfying conversation, they have demonstrated a very natural ability to help customers who just want a quick, transactional response. By blending chatbots with live actual humans, companies can create a better customer experience by giving customers that type of interaction they want in a timely, and efficient manner.

All Posts
×

Almost done…

We just sent you an email. Please click the link in the email to confirm your subscription!

OKSubscriptions powered by Strikingly